Investing with Data.

9 Charts: The rotation is starting.

Sean D. Emory — Fri, 26 Jun 2026 19:35:52 GMT

Big Picture

💡 This week we got a tiny glimpse of what we think happens over time. The rotation seems to be starting.

This week we saw two back-to-back days where the Nasdaq was down 3% and then 2%, and our book was up over 1% on each of them. Even today as I click send, we are up 2.7% roughly as we head into closing bell, and market down. It is truly striking market behavior but indications of what we have expressed.

Over the last 24 months with political and geopolitical events like tariffs and war the market has shunned large pockets of the market. Bill Ackman this week said it best in an interview I heard…

Bill Ackman this week: “businesses with highly predictable future cash flows… are ironically where investors are withdrawing capital, making them the most attractively priced they have been in a long time “

We agree with that and we saw emerging signals that ~~investors~~ I mean TRADERS are on edge of the AI trade as Microsoft pushes towards open source models, Open AI delays their IPO, and Uber is gating AI token usage. This is good news.

Now it would also help to finally be done with political ping pongs so that inflation and growth concerns can normalize. Given our positioning today, that is the catalyst.

The way we generate returns is a company takes a significant portion of the cash they generate and reinvests it into the business. The investor’s ultimate outcome depends on the long-term residual value of the company. Which means it is the cash flow over the next 3-5-7 years that matters which eventually drives that stock price. That has been the lifeblood of markets for a hundred years.

This week we continued to see oil prices drop. Good news. Let’s see if that holds, seems more likely than not.

Everything about Uber and Microsoft and OpenAI chips away at the AI or nothing narrative. What you get instead is a rotation into durable, sometimes economically sensitive names that in many cases have better underlying economic stories than the AI leaders, and they are finally getting their day in the sun. It is still 6am in that day tho…

We think this keeps building as war gets pushed to the background, inflation continues heading back in the proper direction over the next several months as it should, and we work through more earnings from the companies we own and watch. AI is very real. We are looking for companies we would be happy to own for the next ten years if we were forced to hold. Most of the AI darlings today do not fit that test. Some do and we own them. If the market had to stay closed for the next 10 years would you really feel comfortable owning Micron at over $1T valuation? Not so sure about that. I feel comfortable about Block, Meta, Amazon, Omnicell and others though.

We also got more confirmation on consumer this week, which has been a constructive view of ours for a while now. The savings rate got revised up again. Q1 GDP was also revised up from 1.6% to 2.1% annualized. Same pattern we keep seeing. Government data shows one thing on release, then quietly revises in the direction of a healthier consumer and a better economy. This is now 18 months of the market fearing a consumer that, in our work, keeps showing up in better shape than the headlines suggest. Another piece of evidence behind why we keep leaning into the durable, economically sensitive names.

This week we also share a little on memory over-earning by close to a factor of 10, how Cerebras shows there is another path that does not require HBM for inference, and the fact that three of our companies use Cerebras. Then a few notes on how Meta is benefiting from AI directly. That is the tone this week. Patient, second-order, and getting a small preview of where we think this market is heading.

Here’s Latest Podcast on Meta: “6 Questions on Meta: Burden Today, Moat Tomorrow” Listen on YouTube, Spotify, Apple.

[1] It has been an AI or nothing trade.

The S&P has added over $6T in market cap this year but AI is doing almost all of the lifting (green). Energy is a small slug but that is evaporating as I write this as energy names are now being hammered as the war trade is degrossing. Two months ago it felt like you should own energy, now it’s the opposite with oil collapsing 30%. Is that trading or investing? The real take here is that the rest of the index is below zero. There’s lots of non-AI opportunities. That is the market we are in. We’ve been beating this drum for a while.

[2] You think Micron margins can stay like this? I don’t.

Micron is running 85% gross margin. That is above every Mag 7 name. Hardware has never sustained margins like this for long. Apple is closer to 30-40%. You really think Micron can hold 85% or even half that? I suspect not over time. Nvidia has averaged 60% over cycles. We think the memory complex is over-earning by close to 10x. Yes, 10x. Good for Micron, but man that is going to be such an ugly pop at some point. Revisit in a year. The cycle always shows up. Not a name we would want to own for ten years, and I will share more in a second.

[3] Is that sustainable investing, or market timing? Like flipping homes in 2007. Don’t be the bag holder.

Here’s some numbers just to see. Bernstein models Samsung, SK Hynix, and Micron EPS peaking in 2027 and cutting roughly in half by 2028. So basically going from $500 in earnings to $20,000, then down to $1,000 over time. So the stock would essentially be up 10x on earnings that 10x’d, that looks normal. But then what happens when that drops back 70% as shown? If you are buying these here, you are timing a cycle. That is fine. Just call it what it is. We would rather compound into a business than try to nail the top tick.

[4] Cerebras runs inference without touching HBM. Nobody is talking about this.

This one is so key. And frankly not enough people are talking about it.

The memory trade is built entirely on HBM scarcity. It is scarce no doubt, Micron revenue and margins show this.

Cerebras CS-3 is wafer-scale and uses none. You heard that. None. No DRAM/HBM. Which means inference is possible without it. Hmmmm. Also 15x faster inference at the same time. If inference really is the order-of-magnitude opportunity, that bottleneck cracks open. Another reason we’d take the cash flow from the memory names today rather than the multiple tomorrow.

[5] Cerebras counts AWS, Meta, and Block as customers. We own all three.

Three of our companies seem to be playing this right. They were named as customers of Cerebras for best in class inference and no HBM bottleneck.

[6] Meta is best in class on basically everything.

Now to AI and Meta. We have said they are the best way to play AI and there’s lots of reasons to believe this. I find it hard that most don’t understand this. Again, maybe because the beneficiaries are being left behind. Bespoke actually has an AI applications basket and it’s down 17%, these are supposed to be the AI beneficiaries, but still the AI builders are seeing the interest. Logically doesn’t make sense, but it may just be the short-termness of this market. For Meta: engagement, ad load, revenue per user. Meta keeps showing up at the top of every chart that matters. AI is making each of these levers better, not replacing them. This is the kind of name we are happy to hold for the next ten years.

[7] 45% of all genAI ad spend is going to Meta.

Every AI company that wants users has to advertise to find them. They are all spending on Meta. Roughly 45% of genAI ad dollars are landing there. Meta is quietly selling the shovels and the market is not pricing it that way. This is the second-order AI exposure we want.

[8] Meta AI surfaces are scaling fast.

Meta AI is now showing up across the family of apps at scale. They are their own biggest AI customer. The distribution advantage is enormous and it is compounding quietly while everyone watches the chip names.

[9] meta.ai traffic is up roughly 4x since January.

Standalone meta.ai web traffic is up around 4x since the start of the year. Nobody talks about Meta as an AI consumer product. The chart says they probably should. Another example of how Meta is winning at AI while the headlines look elsewhere.

Looking Ahead

Rotation. Watching whether the down-Nasdaq-up-us days from this week become a pattern as oil settles and inflation cooperates.
Memory cycle. When the HBM print starts to mean-revert, that is the tell on whether the cycle names get repriced.
Meta print. Next earnings will test the AI ad spend story and the meta.ai user numbers.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

www.avory.xyz

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn’t constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

12 Charts: We said it for years. Finally the Fed listened.

Sean D. Emory — Sat, 20 Jun 2026 19:05:01 GMT

Big Picture

💡 Warsh seems to be re-writing the playbook. Real-time data, real rates, real wages. Music to our ears. If you follow our work then you know this!

This was the most important Fed meeting in three and a half years, full stop. Kevin Warsh delivered his first statement at 130 words. Powell’s last one ran 341. Yellen’s peak was 895. Warsh dropped forward guidance, skipped his own dot, and stood up five task forces on communication, balance sheet, real-time data, productivity, and the inflation framework. That is the Fed moving toward real-time data. We have been beating this drum for years.

We have been on this for 3.5 years now. The lagged-data Fed has been our single biggest critique of this cycle and the source of most of the policy-mistake fear that has whipsawed markets. The way I see it, that fear just got removed. We are not there yet. This was commentary and action is on the way. But when you are a firm focused on the data, the data says things are ok, and markets are on edge due to legacy data, the mental outcome is one of confidence in the future.

Underneath the headlines the data is doing the same thing it has been doing. Real wages are still positive using the real-time inflation data. Retail sales printed +0.9% M/M in May with the control group +0.7%. 5Y breakevens are at 2.31%, below the 10Y at 2.26%, which tells us the market is more worried about a near-term policy error than long-run inflation. Real rates are +1.89%. Restrictive. Anchored.

On the AI side, we are watching open source close in on the frontier in real time. GLM-5.2 hits 74% on agentic coding at max effort against Claude Opus 4.8 at 78%. Token economics at the model layer are not sustainable. Value moves to orchestration. How can enterprises and consumers actually use AI with ease. That is where we are spending time. This has been our view since day one on AI and we think this thesis is starting to show its colors.

And today’s headline confirms a thesis we wrote about in March and April when we made our Amazon investment as it fell sharply. Amazon is opening Trainium to outside data centers. Amazon is the king of taking margin out of high-margin, high-capex businesses. Retail, logistics, cloud, and now frontier silicon. That’s my view.

[1] Warsh’s first statement ran 130 words. Powell’s last one ran 341.

This is what a Fed pivot looks like in literal word count ha. Greenspan’s January 2002 statement was 128 words. Yellen peaked at 895 in September 2014. Powell averaged 400 to 500. Warsh came in at 130. He dropped forward guidance entirely. He skipped his own dot. He stood up five task forces. The Committee moves to real-time data, at least that is what he said during a 2 min stretch during his press conference. We have said this for quite some time here at Avory and this is very very very welcomed news.

Here’s our post from 2022, just to show how long we have said this! Fed was using legacy data tools, therefore market pricing in policy mistakes

[2] Retail continues to defy economists. Matches our view though if you read this weekly!

May retail sales came in at +0.9% M/M versus +0.6% expected. Control group +0.7%. Ex-autos +0.6%. Restaurants and bars +0.8%. The control group feeds directly into GDP and it just printed the strongest read in months. The consumer can keep spending because the consumer is still ahead. Same story, different week. You heard it here first and why we remain confident in our view that you want to own consumer and macro oriented names despite them lagging…

[3] Real wages remain positive on the real-time cut. They have been all along.

Not the first time we show this, but critical. Nominal wages +3.5% Y/Y. CPI-adjusted real wages -0.7% Y/Y because energy is doing damage at the pump. But on real-time basis, our preferred real-time inflation read, real wages are +1.3% Y/Y. The consumer is roughly running behind prices on the headline read and clearly ahead on the real-time cleaner read. This is why spending holds. This is why retail sales were positive, this is why the FED needs to focus on real-time data, and this is why we are happy post FED meeting.

[4] Real-time inflation today once again ahead of CPI. Real-time data telling a much different story.

Not to beat the same idea but real-time inflation is once again deviating from CPI. That is the gap Warsh just acknowledged. Not the exact number, the idea. The Fed needs real-time data, not lagged surveys. The 30-day real-time read keeps running below CPI which says the official series is still overstating the moment. Funny thing is it has ticked down mostly all year long despite energy surges. This is core reading.

[5] 5Y breakeven 2.31%. 10Y 2.26%. Market more worried about a Fed mistake than long-term inflation.

More on inflation as the market can provide signals. 5Y breakeven sits at 2.31%. 10Y at 2.26%. This is what market is pricing in as inflation today. The 5Y above the 10Y tells us the market is pricing more risk in the near-term policy path than the long-run inflation outcome. Anchored regime. The 5Y5Y forward at 2.21% is consistent with the Fed’s 2% target. The bond market is saying inflation expectations are not the problem. That’s the data. Yes a moving target but everything else I just wrote also says the same…

[6] Real rates are +1.89%. The Fed is still restrictive.

Then we have 2Y yield 4.20%. 5Y breakeven 2.31%. This means real rate +1.89%. That is restrictive. The Fed is still leaning against the cycle and the cycle is absorbing it. This is exactly the setup we have been writing about. Trust the data, not the dots. But it means over time we should get some reprieve if the FED actually leans to real-time.

[7] Yields followed oil. Oil is rolling. Yields should follow.

Now this is pretty important. USO, which is a proxy for oil pricing in markets, is at $111.87 and the UST 10Y is at 4.45%. YTD they have moved in near lockstep. Oil leads, yields follow. With Iran tensions easing and the energy curve in backwardation, oil is rolling. The yield should follow. Music to our ears for the long-duration names and a catalyst to what we saw the first month of the year. An economy running nicely, consumer names working, restaurant stocks working, yet the war threw a nice wrench into that. Oil down should equal yields lower, which should be a catalyst.

[8] June dot plot raised the 2026 median to 3.75% from 3.375%. Warsh skipped his own dot.

18 of 19 dots submitted. Warsh abstained. Median 2026 fed funds 3.75%, up from 3.375% in March. 2027 median 3.63%. 2028 median 3.38%. Longer-run 3.13%. Core PCE 2026 revised up to 3.3% from 2.7%. GDP 2026 cut to 2.2% from 2.4%. It is the first meeting and the dots are a forecast not a commitment. Yes hawkish at first glance, but the press conference for us was a little less hawkish and opened the door for flexibility and the ability to change course given Warsh is putting some of the burden on these committees and task forces.

[9] Zoom keeps executing. Top-right of the Frost Radar again.

One day this will get back to more fundamentals over macro. But the FED has historically liked to make comments almost weekly, everyone has become a FED watcher, but how about the companies. This week we got the Frost & Sullivan report. They put Zoom at the top-right corner of the 2026 UCaaS (Unified Communication as a Service) Frost Radar. Highest on both Growth and Innovation. Past Microsoft, RingCentral, Cisco on both axes. They keep executing. This is one we own. More to come in another piece.

[10] Amazon opens Trainium to outside buyers. Confirms our thesis around our investment made several months back during SaaS Apocalypse.

This is the headline of the week if you are in AI land. Amazon is in talks to sell Trainium AI chips directly to outside data centers (Bloomberg).

Peter DeSantis, AWS SVP of AI, confirmed the talks. AMZN +2.55%. MRVL, the primary Trainium design partner, +7.27%. The scale is real and the ecosystem should be worried. Over $225B in Trainium contracted commitments. Custom silicon at $20B annual run rate, $50B if it stood alone. Trainium 2 sold out, Trainium 3 nearly fully subscribed, Trainium 4 already heavily reserved. We wrote about this in our SaaS Apocalypse pieces in March and April. Amazon is the king of taking margin out of high-margin, high-capex businesses. Retail, logistics, cloud, and now frontier silicon. This is why we own it.

[11] Open source is catching the frontier. Fast.

GLM-5.2 hits 74% on agentic coding at max effort. Claude Opus 4.8 sits at 78%. The gap is tiny and the cost curve is collapsing.

We are going from expensive closed models to cheap open ones and the token economics at the model layer are not sustainable if you draw this out over the coming decade. At least that’s the view right now.

The value moves to orchestration. How enterprises and consumers use AI with ease. That is where the moat is. Reinforces our long-standing thesis that frontier will lead and open source is a fast follow.

[12] Real estate portals like Zillow see little AI search impact. Most traffic is still direct.

GenAI web traffic share. This is good news for real estate portals. REA 0.48%. Z 0.33%. HEM 0.30%. Some categories are getting reorganized by AI search. Real estate is not one of them yet. Most traffic still arrives direct. Direct means you still control brand, economics, and can drive AI-based outcomes on platform. Worth flagging because the narrative says AI search disrupts everything. The data says it depends and as of right now, not so much.

Net Net

Ok so that was a lot but Warsh is the story and the data is the support here. The Fed just moved from a backward-looking, forward-guidance-heavy regime to a real-time, data-driven one in a single 130-word statement and a presser that was clear to us. We think this removes the policy-mistake fear that has been the biggest risk every cycle, all while underneath the headlines the data confirms our view. Real wages positive, retail sales strong, real-time inflation slower than CPI, breakevens anchored, real rates restrictive at +1.89%, oil rolling, yields should follow. On the AI side the frontier is no longer alone. Open source is right there (again), the model layer is important but fragile to say the least, and the value is likely in the orchestration. And today’s Amazon news helps solidify the loop on a call we made in March and April. Amazon is the king of taking margin out of high-margin, high-capex businesses, and Trainium is the next chapter. We’re staying with the data.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

www.avory.xyz

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

10 Charts: AI is killing software? Really?

Sean D. Emory — Fri, 12 Jun 2026 13:30:56 GMT

Big Picture

💡 Is this an inflationary cycle or just energy spike? Frustrated GARP investors.

Happy Friday! So this week CPI came in at +4.2% Y/Y in May, the hottest print in over a year. But energy alone added 1.50pp of that. More than double its normal run rate. Strip energy and core was +0.2% M/M, +2.9% Y/Y, the softest core read in three quarters. So the surface number looks bad and the underlying number is fine.

Here’s the way we think about it. It started with tariffs a year ago or so. The narrative has been the conversation ever since and once that got pushed to the back, then came war. Almost the same exact reaction, which has impacted some of our names sensitive to economic factors despite a strong economy. Something we have gotten right over and over if you read this. Regardless market likes to the ride the waves of narratives versus data. So with Iran war making noise, then the faster that resolves, the faster the “war equals inflation” framing falls off. The longer it drags, the more it embeds. Either way, real-time inflation has actually moved lower over the last two weeks and real wages on the Truflation cut are still positive at +1.3% Y/Y. The consumer can keep spending because the consumer is still ahead. That’s why we’re not changing the underlying view.

The other story this week is software. Anthropic launched Fable, a fork of Mythos, and the SaaS apocalypse trade ran again. We continue to look for the evidence in the actual financials. We’re four years in from ChatGPT, the build out is massive, and the names are reaccelerating. Adobe reaccelerated. The top 50 cloud software names just put up the best aggregate net new ARR in five years. Meta keeps taking ad budget. meta.ai traffic +260% Y/Y. The disruption narrative and the tape are not telling the same story. Sucks, but market likes narratives, and its going against those areas. Small side note, apparently Fable is the most deceptive model yet, producing false info more than others, so tread lightly.

One more worth flagging. Cash App launched Cash App Mobile this week. $40 a month for unlimited 5G on AT&T. They are turning the phone bill into a Cash App product. Something we are excited about. It creates a recurring non financial product embedded…

[1] Like it or not, real wages remain nicely positive. They have been all along.

So nominal wages are +3.5% Y/Y. On the CPI print, which is what most economists anchor to, real wages are -0.7% Y/Y. Energy is doing real damage at the pump and that’s pulling the CPI-adjusted line negative. But on Truflation, our preferred real-time inflation read, real wages are +1.3% Y/Y.

Either way the consumer is roughly running in line with prices, and on the cleaner read the consumer is still ahead. That’s why we keep saying the spending side can hold, and it has.

[2] CPI +4.2% Y/Y. Energy is the print.

Look at the green. Lots of it.

Headline accelerated from +3.7% to +4.2% Y/Y in May. Energy contributed 1.50pp of that, up from roughly 0.4pp at the start of the year. Core services held around 2.1pp. Core goods barely moved at 0.21pp. This isn’t broad-based reinflation. This is gasoline at $4 and a diesel curve that’s been ripping since Iran. Strip energy and the print is +2.7% Y/Y. Basically where were year ago.

[3] Energy futures suggest the headline has peaked.

This one is little much but important.

Here you can see the energy contribution to headline CPI plus what NYMEX futures are pricing for the next two years.

The current curve takes the energy contribution from +1.50pp today back down to roughly +0.10pp by mid-2028. If the curve is right, the headline inflation reading mechanically rolls back into the low 2s.

Obviously this assumes no further escalation. But it’s the market’s best guess and worth respecting.

[4] Food prices are cooling, not embedding.

Big for a name we hold First Watch. Go get your BEC from them if you can. Now… the Bloomberg Bacon-Egg-Cheese plus Coffee index. $2.58 per serving in May 2026, down from a peak of $3.30 in October 2024. Bacon is the biggest line at $1.01, coffee $0.48, cheese $0.46, eggs $0.36, bread $0.27.

Worth flagging that the food side of the inflation story has been quietly resolving for almost two years now, even with all the tariff and trade noise.

Nobody is talking about this on food side.

[5] Cloud software net new ARR re-accelerating.

So you are telling me AI is killing software? Then why does this look liek this…

The aggregate top 50 cloud software net new ARR grew +127% Y/Y in Q1 FY26. Best print in five years. YEAH YOU READ THAT RIGHT. The chart was below zero for most of 2022 and 2023, hovering through 2024, and then this. So again where’s the evidence that AI is taking down SaaS in the financials? We don’t see it. We see the opposite. The big names are reaccelerating, not rolling over.

Ugh frustrating

[6] Adobe is re-accelerating into the AI cycle.

Adobe reported Q2 FY26 revenue at $6.62B, +13% Y/Y. Up from +12% last quarter and +10.8% the quarter before. Y/Y has accelerated three quarters in a row, not decelerated. AI-first ARR, the metric they’re using to track Firefly and Express monetization, tripled Y/Y and now exceeds $500M.

So one of the names that should be most exposed to AI disruption, the creative software incumbent, is putting up the best growth rate it’s seen since 2022.

Again frustrating and we are AI pilled over here.

[7] Meta now 54.6% of measured ad budget.

Now looking at ad spending. For the week of May 31 to June 7, Meta is now 54.6% of all measured ad budget, up +1.32pp W/W. Google 18.82%, down -0.26pp. TikTok 4.82%. YouTube 3.14%. And here’s the part. CAC on Meta fell -2.4% W/W and CTR rose +3.7%. They aren’t just taking share, they’re getting cheaper and more engaging at the same time. Something we’ve been explicit about in terms of our thesis on who wins the AI ad cycle.

[8] meta.ai traffic +260% Y/Y. Scaling fast.

Then we got meta.ai pulled in 27M visits in May 2026, up from roughly 7M a year ago.

That’s +260% Y/Y in a category that nobody is paying attention to relative to ChatGPT and Gemini. Meta isn’t just integrating AI into the ad stack. The consumer surface is scaling too. They aren’t the only ones running this playbook, but they are running it well.

[9] AI spend per employee. 680x gap from top to median.

Looking at some Ramp data, the top 1% of companies are now spending $7,500 per employee per month on AI. Top 10% at $611. Median at $11.38. That’s a 680x gap from the top cohort to the median.

Two reads.

First, the AI spend story is real but it’s heavily concentrated in a small group of frontier adopters. When you hear “every company is spending on AI” be careful about who you mean. Second, the median number is still rising, which means diffusion is happening. Just slowly.

[10] Cash App embeds deeper into daily life.

Cash App launched Cash App Mobile this week. $40 per month for unlimited 5G on AT&T, taxes and fees included, no contracts. Powered by Gigs on the backend, integrated directly into the Cash App ecosystem.

The setup here is to turn the phone bill, one of the most predictable monthly bills the consumer has, into a Cash App product. Embed deeper into Modern Earner habits. Block is an investment of ours and this is exactly the consumer-finance operating system thesis we’ve been talking about. Worth flagging.

Net Net

So zooming out. Energy did the work on CPI, not a inflation cycle yet. Real wages still positive which means that only read that matters. Software is reaccelerating, not rolling over, even four years into the AI build. Meta keeps compounding ad share and consumer surface at the same time. Block is quietly turning the phone bill into a Cash App product (different). And the food side of inflation, that nobody is talking about, is back to where it was almost two years ago. Market wants to ride the narratives again, it is frustrating for fundamental investors looking for compounders and not riding waves. Overall though the data keeps telling us something positive. We’re staying with the data.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

www.avory.xyz

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

9 Charts: Here come the IPOs.

Sean D. Emory — Fri, 05 Jun 2026 12:35:10 GMT

Big Picture

💡 Here come the IPOs. History says be patient.

Anthropic just filed for IPO, and this is just after raising a series H (moving down the alphabet). A $65B round at a $965B valuation, OpenAI is on deck, and the AI IPO calendar is loading up fast as they see the window.

Side note, our strategy when you do look through exposure to OpenAi + Anthropic, its roughly 5-10% weight (depends on dilution at IPO) into those two combined.

Overall we are constructive on the cycle beyond AI, but the data from the last IPO cycle is hard to ignore, specifically for the companies that IPO.

Now Berkshire just deployed roughly $20B into acquiring housing company Taylor Morrison, and putting $10B into Google. So hard to get too worried when one of the more prudent allocators is putting significant amounts of capital to work at this time. Keep in mind they have underperformed by about 30% over the last 12-18 months similar to other GARPy investors.

Back to IPOs, we saw a chart floating around showing 29 IPOs from the 2019-2021, what’s crazy is that wave averaged a -55% max drawdown in Year 1. Median was -54%. Only 43% of them were positive at the 12-month mark. So the AI names eventually coming public will need to be priced into a market that already remembers that math. Maybe it’s a known known so market does opposite. But these will have $1T + starting points.

Now the macro underlying data remain stable. We are seeing software become less binary. Late last week IGV, the software index printed the most overbought reading we've seen since early in the year and historically that setup has been bullish, not bearish.

Also covered this week is how Meta is taking share inside the genAI ad economy at a pace nobody else can match. ChatGPT hit 1B MAUs in three years which is the fastest any consumer app has ever scaled. The Revelio labor tracker bounced for the first time in six months. we got a lot to go over.

[1] IPO basket averaged a -55% drawdown in Year 1. History rhymes.

29 names tracked from the 2019-2021 IPO window.

Robinhood was -90% in Year 1. Rivian -88%. Lyft -79%. Even names that recovered (Airbnb, Snowflake) carved out -32% and -45% intra-year before they did.

Average max drawdown -55%, median -54%, and only 43% of the basket was positive at 12 months.

So when Anthropic files at a $1T or more valuation and OpenAI lines up behind them, you can be constructive on the underlying businesses and still believe the printed prices need time.

We've also got Alphabet running an $80B equity raise around this same window, which says a lot about how much capital this AI cycle wants.

[2] IGV 99th percentile overbought. Historically that's bullish.

IGV +5.3% last Friday was great to see. First close above the 200-DMA since January 7. RSI above 80, more than 3 standard deviations above the 50-DMA. On the surface that reads too far too fast. Likely true but the historical setup after IGV closes above the 200-DMA with RSI >80 following a multi-month drawdown averages a +23.5% forward 12-month return versus +10% all-period baseline.

Software is one of those moves where the overbought reading after a long drawdown has been the entry, not the exit.

[3] Waymo downloads +94% Y/Y in 2025. 13% of the rideshare cohort.

We continue to track autonomous vehicles as it’s the future.

Waymo download growth was +569% in 2023, +192% in 2024, and +94% Y/Y in 2025.

The base is bigger now so the rate moderates, but the absolute share is interesting. 13% of the rideshare cohort downloaded Waymo in 3Q25. Uber downloads +6% Y/Y, Lyft averaged -3%. Feels like autonomy went from a science project to a measurable slice of the rideshare cohort in 24 months. Worth tracking against the autonomy thesis we've been carrying.

[4] Meta now 45% of all genAI ad spend. Up from 28%.

Meta's share of genAI ad spend went from 28% in 1Q24 to 45% in 1Q26. Google held roughly flat at 25%. Mobile ad networks lost 13 percentage points. Total genAI ad spend +230% Y/Y, averaging +50% Q/Q across the eight-quarter window.

We're seeing the same thing in the traditional advertiser data: CAC compression on Meta, ROAS expansion on Meta, share consolidation on Meta.

The franchise quality on a platform like Meta is showing up in the AI economy in real time. We remain confident that Meta is big AI winner.

[5] ChatGPT hit 1B MAUs in three years. Fastest ever.

Different times as things get faster but Google Maps took 5 years. Chrome took 6. YouTube took 7. Facebook and WhatsApp took 8. Google Photos took 11.

ChatGPT did it in 3.

And ChatGPT now has more users than Gemini, Doubao, DeepSeek, Meta AI, Claude, Grok, Perplexity, and Copilot combined. That's a crazy adoption curve for the “number 2” in AI right now.

Whatever you think the right narrative is around AI, the user count is growing.

[6] AI app downloads +36% Y/Y, time +40% Y/Y, spend +145% Y/Y in 2025.

The 3 numbers tell a pretty big story. ~1.9B downloads in 2025, ~12B hours spent, ~$3B in consumer spend. Vs 2023 the multiples are +630%, 50x, and 20x respectively.

Spend is key tho. Consumers will pay for AI apps and they're doing it at a rate nothing in consumer tech has ever matched in a 24-month stretch.

We have also shared in the past that retention in AI is pretty bad early on, but cohorts as they mature get better.

[7] Labor tracker +123.7K in May. First clean bounce in six months.

Revelio is the alt-data labor tracker we follow ahead of NFP. It combines private payroll signals with ADP-style inputs. The last six months printed flat-to-negative reads (-12K, +8K, -22K, +14K, -5K, +24K) which was part of why people were worried about the labor cycle cracking. May printed +123.7K. NFP itself drops / dropped today depending on when you read this. (Actually just dropped before I press post, and 172k jobs for NFP, and private payrolls 120k right where our data said)

[8] ISM Manufacturing PMI 55.1. Cycle high.

PMI revised from a 55.3 flash to 55.1 final but still the cycle high. Two-year arc went from sub-50 contraction through last year to a clean expansion regime now.

We've been tracking this against the soft-landing thesis and the manufacturing side keeps confirming it.

Growth is holding up where you'd expect to see weakness first if things were getting worse.

[9] U Mich consumer sentiment 44.8. 48-year divergence vs SPX.

All detail is like to share. SPX near all-time highs. U Mich at 44.8. The gap between sentiment and the markets is the widest reading in 48 years. We've spoken about this in our last two editions and it keeps getting wider. We think this is more of a survey problem than an economic problem. A trend we have come to know across many of the key datasets people historically track.

Net Net

We remain constructive. The cycle data on labor, manufacturing, and consumer spend is stable enough. The corporate side is healthy, Meta is taking share inside the AI economy at a pace that's hard to ignore if you are looking at the right metrics, on top of their 30% growth, IGV's overbought setup is historically a buy not a sell, and ChatGPT just printed the fastest adoption curve in consumer tech. The asterisk is the IPO supply lining up behind Anthropic and OpenAI. History says those baskets average -55% drawdowns in Year 1. So we keep the thesis intact, watch the IPO names get repriced post-listing instead of chasing them at the open, and stay long the franchises that keep executing.

Looking Ahead

NFP today (June 5). Revelio bounced. Now we see what the official print says. 172k jobs added.
AI IPO calendar. Anthropic filed, OpenAI on deck. Pricing and float will matter.
Industry guides coming. AI infra, data centers, semis, energy. Get ready.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

www.avory.xyz

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

7 Charts: Value Is Flowing Up the Stack.

Sean D. Emory — Fri, 29 May 2026 15:00:51 GMT

Big Picture

💡 Pay attention to software, we are starting to see the inflections.

AI value creation is starting to flow up the stack. It’s moving from the infrastructure layer, the semis, the data centers, the networking buildout, into the software and application layer. Two companies reported earnings for us this week and Okta mainly was a standout. We think both are early indications of that transition.

Here’s the way we think about it. Okta is a big beneficiary of what’s to come in AI, specifically around identity and access management and workflow orchestration. Every agent out there is going to require access to the tools that enterprises already use. Those enterprises are going to partner with organizations that help them orchestrate and govern the AI that’s happening inside the organization. That’s exactly where Okta and a company like UiPath sit in the stack.

An important distinction we’ve been thinking about. Today’s AI models are probabilistic. They create probabilities on what’s next. Deterministic workflows are not AI at all. They’re process automation. The combination is powerful tho and can also help reduce costs. When you route tasks through deterministic logic, you increase the success rate of those tasks and you reduce compute and token usage because outcomes aren’t dependent on probabilities running in a loop. They run through a logic tree. That’s the unlock. Okta governs who and what can access the enterprise, a big deal in a world where AI breaks down trust. We think this is one of the more interesting setups in software right now.

[1] The mobile internet playbook says software is next.

A16Z put out a great chart that maps the history of value creation during a major technology transformation, specifically the mobile internet cycle from 2010 to 2016.

The pattern was semis first, then infrastructure and devices, then software.

By the end of that cycle, software had outperformed semis by roughly 1.85x to 1.00x. We think this chart is the right structural lens for what’s happening in AI right now.

The infrastructure buildout, semiconductors, data centers, power, networking, has been the dominant story for the last two years. The software layer is starting to catch up. You can see that on the next few charts.

[2] Okta crosses $1.9B ARR. NRR ticks higher for the first time in 11 quarters.

Okta reported Q1 FY27 results after the close yesterday. This is a new position for us over last two months.

Subscription revenue came in at $750 million, up +11.3% Y/Y.

That’s twelve quarters of grinding growth with Y/Y holding in the low double digits.

ARR also crossed $1.9 billion, up +12.3% Y/Y.

Okta customers with annual contract value above $100K now sit at 5,180. The number we were really watching was net revenue retention, which ticked back to 107%.

That’s the first uptick in 11 consecutive quarters.

To us that’s the tell as it relates to AI.

Identity and access management is the gating layer for every agent that runs inside an enterprise. You don’t let an agent touch your systems without knowing who or what it is and what it’s authorized to do.

They also launched Auth0 for AI Agents and Okta for AI Agents this quarter. Two SKU families, one identity fabric. Builders and IT now both have an Okta-native path.

Still early days on this, but the net revenue retention inflection paired with the new agent SKUs is worth watching closely.

[3] UiPath ARR also crosses $1.9B. Deterministic plus probabilistic is the combination.

UiPath ARR crossed $1.9 billion this quarter as well, hitting $1,901M and growing +12.3% Y/Y.

That’s almost exactly the same scale as Okta.

The reason deterministic automation and probabilistic AI are a powerful combination is straightforward. When you route a task through logic trees instead of asking a model to re-infer the answer every time, you get higher task completion rates and you use fewer tokens. The model handles the ambiguous judgment calls. The automation handles the repeatable steps.

So UiPath is the layer that orchestrates that routing inside the enterprise. They’ve been building this out for years and the AI cycle is now the tailwind. Something we’ve been explicit about in terms of our thesis.

[4] Stripe Atlas is incorporating a company every 12 minutes.

We have shown other data points on prior newsletters, but this is same takeaway. Stripe Atlas is now incorporating a company every 12 minutes. Think about that.

In 2017 that number was 295 minutes per company, so we’ve seen a 25x compression in the time it takes to spin up a new legal entity over the last nine years. We think this is another indication of how infrastructure is enabling a new wave of software creation. As new businesses form faster, they immediately start requiring software to run those businesses. Data management, identity, orchestration, productivity.

Value starts to accrue to the software application layer the moment those companies are alive.

[5] The S&P 500 is near highs yet consumer sentiment is at a 48-year low.

The S&P 500 is near all time high while the final University of Michigan Consumer Sentiment reading for May came in at 44.8. That’s the widest divergence between equity prices and consumer confidence in 48 years of data.

I think this is exactly the deviation you’re also seeing inside the market, where consumer-related names aren’t participating as aggressively and software names are in a similar position. But I think both will come to fruition. Businesses are doing fine. The macro economy is doing fine. Valuations are very cheap in some of those areas.

[6] Used car prices are up +7.52% in 90 days. Tesla is up +14.83%.

This one is interesting.

CarGurus Used Vehicle Price Index is up +7.52% over the last 90 days, with the average used car now at $29,034 and Y/Y growth at +2.88%. Tesla is leading the move at +14.83% in 90 days and +12.5% Y/Y. The move is broad-based across body styles and makes, which is what makes this an important indicator rather than an anomaly.

As you know we are tracking real-time data on used vehicles closely as it relates to inflation.

This isn’t 2021. We’re not seeing demand completely overwhelm supply. But rising used car prices feed into the services inflation print, and that’s a data point we’re watching closely.

[7] Hedge fund shorts are at a fresh 10-year high.

This data point has been floating around.

Goldman Sachs Prime data shows hedge fund short exposure at 12.8% of US gross, a fresh 10-year high that has now passed the 2021 peak.

Historically that doesn’t specifically lead to tops in markets. There are certain instances where it does, but as a rule it’s not a reliable sell signal.

What it is an indication of is how markets can be at all-time highs while professional capital is positioned for the opposite. Generally if you go back multiple decades you see the reverse configuration at market tops. We’ve been saying for quite some time here at Avory that the defensive positioning across institutional capital is a feature for patient long-side investors, not a warning. This data point reinforces that read.

Net Net

Value is starting to flow up the stack. Okta and UiPath both crossed $1.9B in ARR with growth holding in the low double digits, and Okta's net revenue retention ticked higher for the first time in 11 quarters. The mobile internet playbook from 2010 to 2016 is the right structural lens in our view. Semis first, then infrastructure, then software, and we're very early in that third leg. Meanwhile the S&P pushing higher, consumer sentiment is at a 48-year low, hedge fund shorts are at a fresh 10-year high, and used car prices are rising but nothing like 2021. Professional capital is defensive, sentiment is washed out, the consumer is steady, and the fundamentals are doing fine.

Looking Ahead

NFP (June 5). Initial claims are at 209K and the ADP pulse held at 35,750 jobs per week through early May. The payrolls print will tell us whether the labor market is holding its footing or starting to slow. We want a number that keeps the Fed patient without reopening the wage conversation.
US-Iran deal progress. Let’s see here.
Software earnings follow-through. Okta and UiPath set the frame this week. The question heading into June is whether the rest of the enterprise software layer starts to show the same NRR inflection. That’s the signal we’re watching for the AI value-up-the-stack thesis.

That’s all for this week.

Know another investor who'd find this useful? Forward it along or share on your page!

Investing with Data:

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Ticker Tags:

Where are we in the hyperscaler capex cycle?

Sean D. Emory — Fri, 22 May 2026 10:20:53 GMT

Here is our take at Avory. If you want to read the actual piece in PDF format, see it here… Avory’s Analysis on on CAPEX CYCLE.

So lets start… This is an empirical analysis of the AI infrastructure buildout, updated with NVIDIA’s Q1 FY27 results released May 20, 2026. We look to test the prevailing market assumption / confusion that 2026’s roughly $700B in hyperscaler capex represents a new steady-state baseline. The data points to peak buildout still ahead, with implications for NVIDIA, the Big 4 hyperscalers, and Meta’s free cash flow trajectory as a sole example. We are investors in Meta for disclosure purposes, and anything in this research piece is not investment advice.

Abstract

Summary of findings

What the data shows.

This working paper addresses where the hyperscaler accelerated compute buildout stands within its overall cycle. We have witnessed many market participants treating 2026’s roughly $700 billion in Big 4 hyperscaler capex (Amazon, Google, Meta, Microsoft) as a permanent annual baseline. The empirical evidence suggests this is a misreading. Yes, capex is massive, but we think the market needs more clarity on what this spend represents. Here is our version of clarity.

Using verified SEC filings, the latest NVIDIA Q1 FY27 results (released May 20, 2026), and a system-level replacement multiple methodology, we estimate the Big 4 are approximately 45 to 55 percent through the accelerated compute cycle on a dollar basis as of year-end 2026. The cycle is more advanced than chip-only models suggest because hyperscaler spending captures both silicon and supporting infrastructure (power, cooling, networking, buildings).

Peak Big 4 capex is most likely in 2027 to 2028 at $700 to $800 billion, with normalization to a steady-state run rate of $380 to $800 billion annually depending on refresh-rate assumptions. Even the most conservative scenario implies capex never returns to pre-AI levels. For Meta specifically, all three scenarios produce a 2032 EV/FCF multiple between 9.9x and 12.6x at the current $1.54T enterprise value, well below the S&P 500 average of ~25x.

KEY TAKEAWAY

The 2026 capex figure represents an acceleration of likely future capex, not a permanent baseline. The hyperscalers are replacing legacy general-purpose infrastructure with accelerated compute. The $700B isn’t the future; it is the cost of getting to the future.

1.The framing problem

The mistake we think people are making.

The same misreading keeps appearing in analyst notes and X threads. Analysts are taking the roughly $700 billion in 2026 hyperscaler capex and assuming that number persists annually into the future. Underwriting to that as if $700 billion is the new normal forever. The framing we think is wrong, and it leads to bad conclusions on both directions of the trade. The right way to think about this: capex spending right now is an acceleration of likely future capex, because the hyperscalers are replacing all of their current general-purpose infrastructure with accelerated compute. Their entire stacks need to run on accelerated compute eventually. The $700 billion isn’t the future, it is the cost of getting to the future.

METHODOLOGY

3 structural components of cycle need

We tried to build a model grounded in logic, knowing the future is unknowable. But history gives us clues. The goal is not to be perfectly accurate, but to think probabilistically, stay disciplined, and maintain a margin of safety. This is where we landed…

A / Replacement of pre-AI installed base

Cumulative Big 4 capex from 2015 to 2022 totaled approximately $475 billion, of which roughly 70 percent went to data center infrastructure. That puts the pre-AI installed base at around $330 billion in original cost basis. Replacement at 4-6x system multiple equals $1.3 to $2.0T.

B / Net new inference and training capacity

Token consumption is genuinely additive demand with no historical analog. Estimated 30 to 40 percent of current capex is net new capacity. This bucket scales with usage and has no natural terminal point.

C / Refresh of accelerated compute itself

Hidden compounding factor. With Amazon back to 5-year schedules and NVIDIA shipping new architectures every 18-24 months, a meaningful share of even recent GPU capex will need refresh inside this decade.

2. New data, released May 20, 2026

We have been tracking this thesis over the last year and we wanted to update and publish it. The latest report was May 20^th. Nvidia revenue $81.6B (+85% YoY) vs $78.8B consensus. Data center revenue $75.2B (+92% YoY, +21% QoQ) vs $73B consensus. Q2 FY27 guide: $91B +/- 2% (vs $86.8B consensus), with no China DC revenue assumed. Hyperscale customers remain ~50% of data center revenue. Free cash flow $48.6B for the quarter. Gross margin held at 75%. This sets the tone...

WHAT THIS CHANGES.

The Q1 print and Q2 guide reinforced the structural thesis. Hyperscale customers continued to represent ~50% of data center revenue, meaning the buildout is being absorbed by the Big 4 plus Oracle, not redistributed to enterprise or sovereign customers (yet). Networking revenue of $14.8B in Q1 (up 199% YoY) suggests system-level capex is scaling proportionally with chip capex, validating the 2x silicon-to-system multiplier. So net net, we are still on track in terms of where we think we are relative to various points of views out there.

3. Where we stand in the cycle.

So starting, the big 4 cumulative system-level accelerated capex through end of 2026 is approximately $960B. A 4-year stacked total (2023 + 2024 + 2025 + 2026E), not a single-year figure. Cycle progress depends on what we believe total need is. The replacement multiple is the most sensitive input.

KEY OBSERVATION HERE.

If we are 48% through with $960B deployed, the remaining ~$1.04T needs to be spent over 2027-2030. That means annual accelerated capex must remain at $250-300B per year through 2030 to absorb cycle need.

4. Annual capex projections.

What annual spend actually looks like across scenarios

PEAK CAPEX BY SCENARIO

5. Meta deep dive (EXAMPLE)

Base case · 14% revenue CAGR · 5-year refresh · $1.54T EV anchor

Now for many reasons we wanted to put this into practice so one can see how CAPEX flows through PnL. $META is the cleanest test case for the refresh-rate thesis because it has no external cloud business to monetize the GPUs. Every dollar of accelerated capex must be justified by internal product improvement or new product revenue. We actually like that dynamic given their distribution and product velocity.

BASE CASE TAKEAWAY

If Meta trades flat at $1.54T EV through 2032, the multiple compresses to 10.8x EV/FCF. The S&P 500 trades at ~25x. Quality compounders trade 18-25x. The implication: either the stock re-rates higher as the path becomes clear, or FCF generation runs ahead of expectations.

Meta - Conservative Case

11% revenue CAGR · 6-year refresh · slower capex normalization

This assumes Meta’s AI investments hit diminishing returns by 2027-2028. Revenue growth slows to ad-market-plus modest share gains. Meta is forced (or chooses) to slow refresh cadence. Capex normalizes faster, but at the cost of competitive positioning.

COUNTERINTUITIVE INSIGHT

The conservative case for AI capex is paradoxically the best near-term outcome for Meta’s FCF. FCF crosses back to positive by 2028 and reaches $77B by 2030, exceeding the 2024 baseline. The risk: if AI revenue underperforms, the EV may compress before the multiple does.

5. Synthesis of Meta Example

EV/FCF multiple paths converge by 2032

VIEW OF THE TABLE ABOVE

The terminal multiple range (9.9-12.6x) is remarkably tight despite very different scenario assumptions. The long-term math is more robust than the near-term volatility implies. For the sake of your time, we did not show the aggressive case example drawn out.

POSITIONING IMPLICATIONS

1.NVIDIA base case has accelerated capex growing through 2027, then declining in 2028, the timing window where the multiple is most at risk.

2.Cloud businesses on top of infrastructure (AWS, Azure, GCP) are the durability trade. Their economics depend on installed base utilization and pricing, not capex growth rates. But implication is they will normalize around 2x that of prior in aggregate.

3. Meta, as an example is uniquely positioned because it has no external cloud business. Some have seen as a negative, but makes no sense, as they are currently outgrowing their hyperscaler peers. Every dollar of capex must be justified by internal product improvement or new product revenue, the cleanest refresh-rate test case.

4. Power infrastructure, data center REITs, and electrical equipment makers benefit today and the steady-state floor has lifted, however, there will be a MASSIVE air pocket at the end of this as their refresh cycles are likely much longer than Nvidia.

Watch points & conclusion.

Variables that move the answer.

NEAR-TERM CATALYSTS (NEXT 12 MONTHS)

2027 hyperscaler capex guides (Jan-Feb 2027).

The most important data point. Aggregate Big 4 above $750B keeps base case alive. Below $700B activates bear case. Above $850B validates bull thesis.

NVIDIA Q2 FY27 results (August 2026).

Confirmation of the $91B guide vs actual. Sustained 70%+ YoY DC growth into Q2 would push the cycle math further toward bull.

Meta 2027 expense guide.

Expense growth driven by infrastructure costs, depreciation, infrastructure operating expenses. The 2027 guide implies the refresh cadence Meta is committing to. Also some color of flow through economics.

Custom silicon ramp at AMZN, GOOG, MSFT, META.

Trainium 3, TPU v6, MAIA, MTIA. Faster ramp = lower NVIDIA capture rate = different math on cycle progress allocation. This is Nvidia story, less so a accelerated compute story.

CONCLUSION.

The cycle is not what it appears.

The prevailing market narrative that 2026’s $700B in hyperscaler capex represents a permanent annual baseline is empirically wrong. Some understand this, but many do not, they see revenue and assume recurring. The data points to 2026 being near, but not at, peak buildout. Peak is most likely 2027 to 2028 at $700 to $800B, followed by normalization to a steady-state run rate of $380 to $800B annually. If you are an investor(as we are) then this should help you underwrite the next 3-5-7 years. If you are momentum trading then this is not about whether the cycle ends. It is about the slope of the descent and where the new floor settles. Position accordingly.

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Sources & methodology

Working paper appendix

PRIMARY SOURCES

Bloomberg Data for earnings, revenue, consensus

NVIDIA Q1 FY27 earnings release and CFO commentary (sec.gov 8-K, May 20, 2026)

NVIDIA Q4 FY26 / Full Year FY26 earnings release (February 25, 2026)

Meta Q4 2024, Q4 2025, and Q1 2026 earnings releases

Platformonomics CAPEX Retrospective 2025 (Charles Fitzgerald, February 2026)

Epoch AI hyperscaler capex tracker (SEC EDGAR, Q1 2022 to Q4 2025)

Bloomberg projections

Methodology

All hyperscaler capex figures include finance leases where disclosed. Pre-AI baseline excludes Amazon retail capex using AWS-specific breakouts where available. Replacement multiple range (3x to 7x) captures the difference between like-for-like server swap and full system replacement (HBM, NVLink/InfiniBand, liquid cooling, power delivery). Hyperscaler share of NVIDIA DC revenue blended at 47% based on CSP commentary. The Q1 FY27 CFO commentary stated hyperscale customers remained ~50% of data center revenue. Custom silicon estimate (TPU, Trainium, MAIA, MTIA) carries the widest margin of error. Silicon-to-system multiplier of ~2x reflects that full hyperscaler accelerated capex includes the chip plus the buildings, power, cooling, and networking around it.

DISCLAIMER · Avory & Co. internal research working paper. Not investment advice. Distribution to clients and prospects of Avory & Co. only. Past performance not indicative of future results. Forward-looking projections subject to material revision as new data becomes available.

10 Charts: Markets On Edge. Data Says Stable.

Sean D. Emory — Fri, 15 May 2026 09:34:33 GMT

Big Picture

💡 Markets remain on edge for consumer and inflation, yet both still suggest things remain stable.

I have seen a lot of charts and commentary on inflation and growth this week, and the story still depends heavily on which data set you are looking at. Of all things this week, just look at chart 3 below.

Official inflation data suggests inflation is running around 3.8% Y/Y, while consumer sentiment data says consumers are about as negative as ever. But, and this may sound familiar to anyone who has read our Avory newsletter since 2022, we continue to believe the real-time read on inflation matters more. Today, that figure is closer to 1.95%, creating roughly a 185 basis-point gap versus the official print. It is not rocket science to figure out that this has been a better predictor, as the economy has been stable enough over the last 3 years despite the seesaw of the rhetoric.

Now, the market seems to be treating the cyclical AI buildout as one of the few places to hide. AI is clearly a structural theme, but the current spending cycle is very much happening now. It has been relatively insulated from whatever macro narrative the market wants to run with on any given day. More importantly, the underlying spend keeps showing up in the data, including Ramp activity, CapEx plans, and the share-take patterns we are tracking.

Underneath all of this, the key question is simple: is wage growth running ahead of the inflation rate consumers are actually experiencing?

Using real-time inflation, real wages are running around +1.8%. Yes positive despite the charts flying around this week. Using the official CPI basis, they are roughly -0.1%. Those are two very different realities.

The other point worth highlighting is the disconnect between consumer sentiment and actual consumer behavior. Sentiment keeps printing near all-time lows, while spending continues to print near all-time highs. Credit card data though shows consumer spending up +3.2% year over year, with strength still showing up in travel, restaurants, and leisure.

We already knew energy was going to perk up in CPI, but shelter continues to do some of the heavy lifting in the official inflation print, but market rents tell a different story. Apartment List data shows rents running -1.7% year over year. The issue is the well-known 12 to 18 month lag in how shelter flows through CPI, something we have been discussing for years. Net net, real wages are positive, so consumer will spend, and they are spending, and check my article this past week for more color on it!

[1] Real-time inflation is already below 2%.

So lets start here… Truflation’s real-time read is sitting at 1.95% with CPI Y/Y at 3.8%.The real-time price data pulls across millions of items, CPI is largely survey-driven with lagged shelter. Obviously this suggests inflation in the real economy has already done what the Fed has been waiting on for two years. That’s our read atleast. If it was not for energy spike, I actually think we would be talking deflation right now, which is much more concerning…

[2] Shelter is the largest CPI component and it lags by 12-18 months.

Apartment List has rents running -1.7% Y/Y, well off the +17.8% peak from Jan ‘22. The problem is BLS still counts shelter heavy in the basket, and the methodology lags real-time rents by 12-18 months.

Something we’ve been talking about for years now. If you strip out the lag and use what’s actually happening in housing today, the official inflation print drops meaningfully closer to the real-time number.

[3] Real wages depend on which inflation print you trust.

This is the most important chart here… maybe.

Nominal average hourly earnings +3.7% Y/Y. Minus CPI you get -0.1%, which is what the government data shows the consumer is feeling. Using real-time you get +1.8%, which is what we think they’re actually experiencing.

That’s the entire macro debate in one chart. We’ve been in the camp that real-time data tells the truer story, and the consumer behavior on the next few charts backs that up.

Subscribe now

[4] Consumer spend +3.2% Y/Y, sitting near the highs.

Credit card data has consumer spend running +3.2% Y/Y as of May ‘26, near the top of the post-pandemic range.

Sentiment surveys say the consumer is miserable. Their actual spending says otherwise.

Either the macro view has been wrong, or something else, but clearly the data we’re tracking shows a consumer that can stand whatever is taking place. Their spending figures keep supporting that view.

[5] The spend is showing up in travel, leisure, and restaurants.

Bank of America internal data has travel & tourism contributing +1.6pp to Y/Y services growth in April, the highest of the cycle. Restaurants and leisure are flowing positive alongside it. If the consumer were actually cracking you’d see discretionary go first. Instead, discretionary is leading. Same story, different week. Strange right?

[6] First Watch keeps outrunning the restaurant peer set.

Now something in the portfolio. A company executing well as shown above. But hard to run away from consumer spending fears in the short term.

Here’s the facts though.

First Watch posted +2.8% same-store sales in 1Q26 against a peer average of +0.6%. The brand keeps taking share quarter after quarter, and the recent trajectory tracks the broader breakfast and brunch tailwind. Small sample, but worth watching. This is the kind of name where the consumer thesis we just walked through actually shows up in earnings but not in price.

[7] Zillow has outrun the housing industry for 10 straight quarters.

Same story to above. Rates push higher on inflation fears, which sends housing related assets lower.

Yet! Zillow revenue growth has outpaced the broader housing industry by an average of +14.3% over the last 10 quarters.

Industry volume was negative through most of ‘23, Zillow was positive every quarter. Q2’25 had Zillow at +15% versus industry at +1.5%. This is what share-takers look like in a soft category.

Reinforces our thesis on owning the dominant platform when the macro setup is choppy.

[8] Anthropic just overtook OpenAI in business adoption.

This one is crazy and not something you probably thought a year ago. Using Ramp’s AI Index, which tracks paid AI subscriptions across 50,000+ US businesses, it shows Anthropic at 34.4% adoption versus OpenAI at 32.3% as of April. That’s the first crossover.

Overall AI adoption is now at 50.6% of tracked businesses, up from basically nothing two years ago. The cyclical buildout we’ve been talking about is real, and it’s happening at the business level, not just the consumer app level.

[9] AI app uninstalls are running at a record pace.

Now this supports the crossover from above. Web traffic data has consumer AI apps churning hard on the install/uninstall front.

Gemini +169% Y/Y uninstall rate

Copilot +138%

ChatGPT +132%

Claude +90%

The business adoption from the previous chart is the durable part of AI. The consumer app layer is far more volatile and people are trying everything. Two very different stories under the same umbrella.

[10] Stripe Atlas startup formation just hit a record.

Now another cool data point showing what is taking place. Stripe Atlas hit 6,250 incorporations in Q1’26, a new high and roughly 2.3x the post-pandemic 2024 trough. Whatever else is going on in the macro, more people are starting companies right now than at any point in the platform’s history. That’s a tell on real economy formation that doesn’t show up in any sentiment survey. That’s the data.

Net Net

We continue to focus on the main data here that matters. We have a piece out next week showing where we think we are in the AI buildout, when we think it peaks, and how to think about it. It will only show on Substack day 1, so make sure you are subscribed.

Investing with Data:

That’s all for this week.

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

The Market Isn’t Wrong About Hormuz. The Bears Are.

Sean D. Emory — Tue, 05 May 2026 16:15:48 GMT

The market keeps shrugging at the Strait of Hormuz, and the bears keep calling it irrational. The numbers say it isn’t. Roughly $255B in fresh dollars is on its way into the US economy this year, and most of the desk research hasn’t priced it in.

There is a recurring question I keep getting from investors or our advisor clients right now. How is the market still holding in when oil ripped from $61 to $118, when shipping through Hormuz was de facto closed for weeks, and when the geopolitical headlines read like 2022 all over again? On the surface, that is a fair question. Underneath it though the answer is just math… let’s jump into it.

Three structural offsets, two policy refunds, and one shifted producer base mean the net dollar flow into the US economy in 2026 is positive, sizable, and almost entirely separate from anything the Fed has to do.
Sean Emory from Avory & Co on Bloomberg TV Global.

Let me walk through it in the order I worked it out, and qualify each piece with a real number from a real source.

Before that, here’s what prediction markets think about WHEN traffic at the strait returns to normal. 16% chance in May… ie the market is already priced for closure.

Now lets get into it…

The oil shock is real. The wallet impact is half what it used to be.

Why a $120 Brent print doesn’t hit the consumer the way it would have in 1979.

In 2024, the average US household spent $2,411 a year on gasoline, which works out to 3% of total household expenditures. That is straight from the BLS Consumer Expenditure Survey. Compare that to the 1970s, when energy products were 7-10% of real consumer spending. So roughly half the wallet share.

The Kansas City Fed put numbers on this directly. Their analysis concludes that gasoline expenditure shares are about 1% point lower across all income quintiles than they were in 1984, and that the effect of oil price changes on consumption has become much more muted, even among lower-income households.

The pass through math reinforces the point. Academic research shows oil price pass-through into broad consumer prices is about 4.2% in the long run. We know this continues to move lower. So a 50% oil shock translates to roughly 2 percent of headline CPI pressure, not 5 or 6% we have seen in prior decades.

Real oil is elevated. It is not extreme.

The level matters less than the comparison. Compare correctly people

WTI, in inflation-adjusted terms, suggests the average YTD run rate for oil prices ($83) is only about 13% above the inflation-adjusted average going back to 1980. Important because the 1980s oil peak would be equivalent to something in the mid-$100s today, or roughly 30% to 50% higher than current levels. And even then, as we just discussed, oil had a much larger impact on the consumer wallet than it does today. So yes, oil prices are elevated relative to history, but they are nowhere near prior inflation-adjusted shock levels, and their contribution to consumer spending pressure is less meaningful than it was in past cycles.

There is one important caveat the bulls should not gloss over though. The EIA itself flagged Q1 2026 as the largest inflation-adjusted Brent move since 1988. So while the level is not a record, the velocity of the move is genuinely historic. That is the part of the bear case that is real. But then there are other offsets that happen to hit at the same time. Here are some of those..

The US is now a net exporter.

In the 1970s, an oil spike was pure tax. Today, half of it is revenue.

This is the structural piece that most analysts trained in a different era still get wrong. Per the EIA, the United States exported 55% of its petroleum production in 2024, and total petroleum net imports were the lowest on record. During the Q1 2026 disruption, US crude exports surged to record levels as global buyers sought alternatives to Middle East supply.

Also this chart tells that story very clearly.

Translation to above: when crude rips, a meaningful share of the dollars consumers spend at the pump cycle back through US producer revenue, capex, hiring, and royalty payments to landowners.

The Kansas City Fed calls this the “oil producer channel” and notes it is materially more important to the US economy than it was during the 2014 to 2016 slump, let alone the 1970s.

The net drag on the broader economy is a fraction of the gross drag on consumer wallets when isolated by itself.

The OBBBA tax cut is already in checking accounts.

Refunds are running $43B ahead of last year, and we are not done filing.

Through April 17 of the 2026 filing season, total refunds were running $43.0B ahead of the 2025 pace, with cumulative refunds of $296.1B versus $253.1B at the same point last year. That is a +17.0% year over year pace, and the average refund is up 11.3% to $3,275. The IRS Filing Season Statistics confirm the pattern week by week.

This is the Tax Act from last year feeding through. Something we discussed early last year. It includes deductions for tips, overtime, auto loan interest, and senior citizens, and the math shows household after-tax income just got a real lift.

Annualized, with the rest of the season still to come, the full year incremental refund delta is plausibly $55-65B higher than 2025. Real cash, hitting real households, in real time, at same time as oil shock. That’s what I meant when stating that the oil shock occurring at the best time it could have.

The tariff refund is the other input nobody is talking about (still).

The Supreme Court invalidated IEEPA tariffs.

On February 20, 2026, the Supreme Court ruled 6-3 in Learning Resources, Inc. v. Trump that IEEPA does not authorize the President to impose tariffs. The Court of International Trade subsequently ordered Customs and Border Protection to refund the duties collected. As of the filings that prompted the CAPE refund portal launching on April 20, the government owes an estimated $166 billion to roughly 330,000 importers, with the outstanding balance accruing interest at about $650 million per month.

Even Penn Wharton’s independent estimate puts the refund total at up to $175 billion when interest is included. Either number is enormous.

For context, the entire 2024 tax refund total from the IRS was roughly $329 billion. We are talking about half a year of the entire IRS refund machine, landing on importer balance sheets in the back half of 2026.

There is a second-order effect that is just as important.

IEEPA tariffs represented roughly half of all customs duties collected by the end of 2025, running at roughly $18 to $20 billion per month in collections at year-end.

Going forward, those collections stop. Importers retain the cash that would have been remitted. On an annualized basis, that is a $200 billion plus run-rate boost that just structurally disappears from the tariff column.

Section 232 expansions on aluminum, steel, copper, and derivatives took effect April 6 and will claw back some of this, probably $30 to $60 billion per year by our estimate. But the gap between IEEPA cessation and Section 232 replacement is a real net positive in 2026.

The math, end to end.

What it actually nets to in dollars flowing through the system.

Here is the bridge, all in. Every line ties to our calculations above. I would view the number as directionally accurate given some of the unknowns in timing.

Here’s another visual.

For scale: US nominal GDP printed $31.86 trillion in Q1 2026 per the BEA. That puts this net flow at roughly 75 to 90 basis points of GDP, on a one-year view, before any multiplier effect.

Tax refunds to households have a higher marginal propensity to consume than tariff refunds to importers, so blended through a 0.5 to 1.0 fiscal multiplier, the GDP impulse meaningfully expands.

And note the BEA’s own footnote on the Q1 2026 release: the IEEPA refunds are being treated as a capital transfer in the National Accounts, which means they are not yet reflected in the GDP print. That is upside still to come in the data, not behind us.

Why the market is doing what it is doing.

Good reminder to do the work don’t just read the terrifying headlines…

If you put all of this together, the market’s behavior stops looking irrational and starts looking like efficient pricing of a fiscal and regulatory regime change that the bear narrative has missed.

The oil drag is real but contained, the consumer is structurally less sensitive than 1970s playbooks assume, and the offsets are large and arriving in 2026.


Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube”: Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

SOURCES

Bureau of Labor Statistics. *Consumer Expenditures 2024.*Released December 19, 2025. Average annual gasoline spending of $2,411, or 3.1% of household expenditures. bls.gov/cex
RBC Economics. Oil price shock: Higher US inflation could weigh on consumers. 1970s energy spending share of 6 to 7 percent of real consumer spending. rbc.com/en/economics
Federal Reserve Bank of Kansas City. Çakır Melek & Vigfusson. The Evolving Link between Oil Prices and U.S. Consumer Spending. Economic Review, 2021. kansascityfed.org
ScienceDirect. Oil price pass-through into consumer prices: Evidence from U.S. weekly data. Long-run pass-through of approximately 4.2% to consumer prices.
U.S. Energy Information Administration. Crude oil and petroleum product prices increased sharply in the first quarter of 2026. Largest inflation-adjusted Brent quarterly move since 1988. eia.gov/todayinenergy
GuruFocus / EIA WTI series. Inflation-adjusted WTI of $99.89 as of April 27, 2026; long-run median $63.47.
Britannica Money. Highest Crude Oil Price. 2008 real peak ~$188; 2011 real peak ~$165. britannica.com/money
U.S. Energy Information Administration. U.S. energy facts: imports and exports. United States exported 55% of its petroleum production in 2024; net total energy exporter since 2019. eia.gov/energyexplained
Internal Revenue Service. Filing Season Statistics, weekly cumulative through April 17, 2026. Cumulative refunds of $296.1B vs $253.1B in 2025; average refund of $3,275, up 11.3% YoY. irs.gov/newsroom
Avory & Co. analysis. Tax refund pace and OBBBA pass-through tracking. +$43.0B cumulative delta vs 2025 same week; +17.0% YoY pace.
U.S. Supreme Court. *Learning Resources, Inc. v. Trump,*No. 24-1287, decided February 20, 2026.
Penn Wharton Budget Model. Supreme Court Tariff Ruling: IEEPA Revenue and Potential Refunds. Up to $175B in refunds; IEEPA at ~50% of customs duties; ~$20B/month in collections at year-end 2025. budgetmodel.wharton.upenn.edu
U.S. Customs and Border Protection. CAPE refund portal; ~330,000 importers, ~$166B principal, ~$650M/month interest accrual. Refund processing began April 20, 2026.
Norton Rose Fulbright. Potential refunds: US Supreme Court overturns IEEPA tariffs. Legal mechanics of refund and re-liquidation process.
BDO Global. Supreme Court Reins in IEEPA Tariff Authority: What Happens Now. Court of International Trade order of March 4, 2026; Section 232 expansions effective April 6, 2026.
U.S. Bureau of Economic Analysis. GDP Advance Estimate, 1st Quarter 2026. Released April 30, 2026. Nominal GDP of $31.856 trillion; IEEPA refunds treated as capital transfer. bea.gov

13 Charts: In one year: $416B to +$700B.

Sean D. Emory — Fri, 01 May 2026 15:22:59 GMT

Big Picture

💡 Big Tech earnings just confirmed it. Capex is exploding, AI is converting, and the consumer is fine.

Ok so this was the week.

Meta, Microsoft, Google, and Amazon all reported, with Starbucks and Omnicell on the side (holdings). And the takeaway for me is pretty clear: the macro still looks healthier than the CNBC guest circuit suggests.

Across the hyperscalers, demand commentary was strong, AI revenue is converting at scale, and capex commitments into 2026 and 2027 remain massive. Meta raised CY26 capex to $125B-$145B. Alphabet moved to $180B-$190B including Intersect. Microsoft is around $190B. Amazon is still $200B+. That is $700B+ of combined hyperscaler capex this year, off roughly $416B in CY25. +71% Y/Y.

The signal we keep coming back to is simple: these companies are not spending because they hope AI works. They are spending because the demand they see internally requires it.

Susan Li said it well on Meta’s call: “we have continued to underestimate our compute needs.”

I’d say that’s the read from our desk too. Especially platforms like Meta with so much distribution.

And the consumer side still lines up with what we have been discussing. Amazon retail units +15% Y/Y, strongest since COVID. Google Search +19%, with retail and finance strength called out. Not perfect. Not booming everywhere. But still a “good enough” economy.

On AI infrastructure, vertical integration keeps showing up. Amazon with Trainium. Google TPUs generating external hardware revenue. Microsoft Maia 200 live in Iowa and Arizona. Meta rolling out custom silicon with Broadcom alongside AMD and Nvidia.

So yes, we remain constructive. AI is real, the winners are emerging with plenty of time left. Let’s get into it.

[1] Hyperscaler capex is exploding. ~$710B combined CY26 guide.

Meta raised CY26 capex guide to $125B-$145B (from $115B-$135B). Alphabet bumped to $180B-$190B (from $175B-$185B, now including Intersect). Microsoft committed roughly $190B for CY26 on the call. Amazon held at $200B+. That is ~$710B at the midpoint of guides, all on a capex.

[2] Amazon AWS +28% Y/Y. Backlog +$120B in one quarter to $364B.

iwd_aws

AWS revenue $37.6B, up +28% Y/Y, the fastest growth in 15 quarters. Operating margin 37.7%. Backlog hit $364B excluding the Anthropic deal, up ~+$120B in a single quarter. AI run-rate now >$15B annualized. Bedrock spend +170% Q/Q.

Amazon is on the frontier again.

[3] Amazon retail units +15% Y/Y. Consumer is fine, full stop.

Worldwide retail units grew +15% Y/Y. Per Jassy, that is “the highest since the tail end of covid lockdowns.” North America COMMERCE segment revenue was $104.1B. Advertising revenue $17.2B (+24% Y/Y), TTM ad revenue >$70B.

We care about Amazon retail units specifically because it is the cleanest proxy for actual consumer goods demand. +15% Y/Y is not a weak consumer signal.

[4] Amazon’s chip business. $20B run-rate, $225B+ in commitments.

Amazon disclosed its custom silicon business (chips) runs at >$20B, growing triple digits. >$225B in Trainium revenue commitments. Trainium 2 is sold out, 30% better price-performance vs Nvidia GPUs. Trainium 3shipping in early 2026, ~30-40% better than T2.

Big tech wants to control the silicon stack and Amazon is moving aggressively. More signal.

[5] Microsoft Azure +40% Y/Y. RPO doubled to $627B.

Microsoft crushed. Azure grew +40% Y/Y in FY26 Q3. Total RPO surged +99% Y/Y to $627B. AI revenue now at a $37B+ run-rate, up +123% Y/Y. M365 Copilot has 20M+ paid seats with +250% Y/Y seat adds.

Not sure how I feel about those copilot seats. As a product I think it’s third tier. So Microsoft Azure is doing great, but I still think there is some risk to their productivity suite do you do AI encroachment.

[6] LinkedIn +12% Y/Y. Hiring momentum showing up in ads too.

LinkedIn revenue grew +12% Y/Y (+9% constant currency) with growth across all lines of business. Q4 guide is ~+10%.

I always liked to look at this as LinkedIn is a useful real-time read on enterprise hiring and B2B advertiser demand. +12% growth is consistent with the ADP weekly pulse picking up that we flagged in our last edition.

[7] Google Search +19% Y/Y. AI Overviews are not killing Search.

Google Search revenue $60.4B, up +19% Y/Y. Pichai called out retail and finance verticals driving strength. Search queries hit an all-time high.

This is the cleanest read on advertiser demand and downstream consumer demand. +19% Y/Y growth in Search ads is a strong economy signal.

[8] Google Cloud +63% Y/Y. Backlog nearly doubled to $462B.

Google also winning in compute needs. Google Cloud segment revenue $20.0B, up +63% Y/Y, with GCP itself growing materially faster than the segment. Operating income $6.6B, tripled Y/Y. Margin jumped to 32.9% from 17.8%. Backlog hit $462B, nearly doubled sequentially, with just over 50% expected to convert within 24 months.

Google is now a real third platform alongside AWS and Azure.

[9] Meta ad impressions +19%, price +12%. AI WORKING.

Honestly outside of them trying to spend more on things that are working, this was great quarter.

Meta revenue +33% Y/Y to $56.3B. Ad impressions +19% Y/Y, average price per ad +12%. Both metrics accelerated vs Q4 2025. Reels time spent +10% on Instagram from AI ranking improvements. Facebook video time +8% Q/Q, the largest quarterly gain in 4 years.

They are only one of biggest 100% of CAPEX is tied to direct usage. Their new Adaptive Ranking Model added +1.6% conversion rate across major Facebook and Instagram surfaces. Value Optimization ARR is now >$20B, more than doubling Y/Y. More than 8M advertisers are now using at least one of Meta’s gen AI ad creative tools, up from 4M when first disclosed at Q4 2024 earnings.

This is the AI play.

[10] Meta business AI. 1M to 10M+ weekly conversations in one quarter.

iwd_meta_business_ai

Meta’s business AI weekly conversations grew from ~1M at start of 2026 to more than 10M by end of March. That’s 10x in one quarter. WhatsApp ads and Status are scaling, with hundreds of millions of people now viewing them daily.

Zuckerberg framed the strategy directly: “We’re building a personal agent focused on helping people achieve the diverse goals in their lives. We’re also building a business agent focused on helping entrepreneurs and businesses across the world use our tools and others to grow their efforts, reach new customers, and serve existing customers better. These agents will work together to form an ecosystem.”

[11] Atlassian: AI is a tailwind, not a headwind, for the right software platforms.

We’ve said for a while that software should not be treated as one bucket. AI is essentially software, and the platforms that integrate it well are going to benefit, not get disrupted. Atlassian’s earnings is one of the cleanest data sets we’ve seen on that thesis.

A few of the standouts. Customers using Rovo (their AI) are growing ARR at ~2x the rate of non-AI customers. AI credit usage is compounding more than +20% month over month. The Service Collection, their AI-heavy offering, is now >$1B ARR growing +30%+ Y/Y. There are other datapoints but clearly software can win.

[12] Starbucks comps +6.2% global, +7.1% US. Niccol is delivering.

Global comp +6.2%, US comp +7.1%, North America comp +7.1%, China +0.5%.

The strongest in three years and the third consecutive positive print after 7-8 quarters of declines.

Long-time followers know we’ve been on Brian Niccol for a while. We owned Chipotle when he took over and that worked out. We were patient with Starbucks because it is a different animal but the early returns are paying off.

[13] Omnicell beat across the board. Titan XT cycle is the next leg.

Omnicell Q1 revenue $309.9M (+15% Y/Y), beat by ~$5M. Non-GAAP EPS $0.55 vs $0.33 consensus, a +66% beat. Non-GAAP EBITDA $44.7M, +42% above estimates. Product revenue +20% Y/Y.

More importantly for our thesis, Titan XT first orders already booked in Q1 2026, with hardware shipments beginning H2 2026 and OmniSphere software rollout in H1 2027. Total replacement opportunity is >$2.5B over the cycle, with bookings ramping through 2026 and revenue scaling in 2027.

Looking Ahead

Jobs data Friday. April NFP and unemployment rate. We want to see continued labor strength but not so hot it changes the inflation read.
Iran. Any news on ceasefire.
Earnings: more earnings next week.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube": Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

10 Charts: Macro data keeps confirming our view.

Sean D. Emory — Sun, 26 Apr 2026 15:33:54 GMT

Big Picture

💡 Macro keeps firming up across labor, consumer, and inflation. Earnings say the same thing.

Ok so busy week and busy month really with market gyrating daily on headlines. Funny though with the noise, the macro continues to hold up and improve. Another week of supportive data reinforces our view that peak uncertainty is behind us and the longer-term backdrop remains healthy. We’ve been beating this drum for a while. Sometimes doesn’t feel like it but it is the right move to stay pat.

Retail sales this past week came in around +1.9% ex-autos, beating estimates and extending a run of resilient prints. The ADP weekly pulse 4-week moving average took another leg higher to +55K, now 5 straight weeks of improvement. Initial claims held in the low 210s. The S&P Global US Composite PMI rebounded to 52.0, a 3-month high. Different data sources, same read for the most part, ie the economy is on solid footing, something we have not backed away from.

Earnings season kept reinforcing it. Tesla printed a mixed quarter but margins expanded nicely to 21.1%, which we liked. ServiceNow had really healthy numbers. Subscription revenue up 22% Y/Y with cRPO +21% constant currency, and the AI target for the year got raised from $1B to $1.5B. The market sold it off on the margin guide (maybe or just SAAS jitters). Our belief continues to be that the large software platforms central to mission-critical operations, with technical ability at the top, execute well in this environment.

We also said goodbye to Tim Cook. Incredible career, roughly 2000% total return under his watch. Worth remembering what it took to earn that return though, on avg Apple stock was down 24% a year at some point with Cook at the helm. Patience and stomach are part of the equation, forsure.

We remain constructive. Consumer is in good shape, jobs market is building momentum, inflation dynamics remain absorbable (rent disinflation keeps grinding lower), and guidance is good enough right now. WTI futures currently price mid-to-high 70s by year-end. Moving target, but directionally consistent with conflict easing over time. That’s the read from our desk.

[1] Retail sales +1.7%, +1.9% ex-autos. Hard to say weak consumer.

Lets start with the March retail sales data that came in at +1.7% M/M headline vs. +1.4% expected. Ex-autos hit +1.9% vs. +1.3% expected. 12 of 13 categories were up, the most in months. That follows an upwardly revised +0.7% in February.

The consumer keeps spending because the consumer has money. Tax refunds are tracking ~10-15% Y/Y and hiring has picked up. This is what “good enough” looks like. That’ll do.

[2] Hiring momentum is building. 5 straight weeks of improvement. Oops more strength…

The ADP weekly pulse 4-week moving average hit +55K private sector jobs. That’s the 5th consecutive week of acceleration.

To put that in context, this number was sitting near zero through most of the fall. It has inflected meaningfully in a short window. The labor market is firming here????

[3] Initial claims in the low 210s. No crack in the labor market.

Initial jobless claims printed 214K for the week ending April 18. The prior week was 208K. We are sitting in a range that historically correlates with a healthy labor market.

Combine this with ADP accelerating and you get a consistent message. Hiring is picking up and firing is not picking up. Not bad

[4] S&P Global US Manufacturing PMI Highest in a LONG TIME.

The flash Composite Output Index went from 50.3 in March (a 2.5-year low) to 52.0 in April. Manufacturing PMI hit 54.0, the highest since May 2022. Services moved back to 51.3.

Activity growth rebounded as we kicked off Q2. Another data point pointing the same direction as retail sales and ADP. The economy is on a solid foundation. Checks the box for us.

[5] Rent disinflation continues. +1.7% Y/Y.

Zillow’s Rent Index grew just +1.7% Y/Y as of March. That compares to the +15%+ peak in early 2022. This is well below the pre-COVID average of ~3-4%.

Shelter is the single largest weight in CPI. The BLS methodology lags private market data by 6-12 months. So what shows up in today’s Zillow print is likely to flow into official inflation data over the next several quarters. This is disinflationary fuel that has not fully shown up yet.

[6] Tesla margins expanded. 21.1% total gross margin.

Tesla reported Q1. Revenue +16% Y/Y to $22.4B, EPS of $0.41 vs. $0.36 expected. The headline was the margin line: total gross margin 21.1%, up from 16.3% in Q1 2025. Auto gross margin ex-credits improved sequentially to 19.2%.

Mixed quarter overall. Deliveries missed and some of the margin lift came from one-time benefits. But in the grand scheme of things, constructive. We care about the direction of margins and that direction was up. Can’t complain about that.

[7] ServiceNow had a really strong quarter. Market didn’t love the guide apparently.

Subscription revenue $3.67B, up +22% Y/Y (+19% constant currency). cRPO $12.64B, +21% constant currency. Beat guide by 100bps. Non-GAAP operating margin 32.0%, beat by 50bps. FCF margin hit 44%.

Biggest tell for us: the internal 2026 AI target moved from $1.0B to $1.5B, same measurement methodology. Customers with >$1M on Now Assist grew +130% Y/Y. Deals with 3+ Now Assist products grew ~70%.

Stock sold off ~12% after hours on a softer margin guide that includes Armis dilution. That is the debate. Our read: demand side of the business is accelerating, AI is converting. Mission-critical software platforms with strong leadership keep executing in this environment. This is one we may want to own.

[8] Tim Cook retires. ~2000% return during his tenure BUT Avg 22% drawdown.

Tim Cook announced he will step down Sept 1. John Ternus takes over. Under Cook, Apple market cap went from ~$350B to $4T. Annual revenue roughly quadrupled from $108B in 2011 to $416B in 2025.

What is interesting though is what it took to earn that return. The chart we built this week shows the average and max drawdowns along the way. The run wasn’t straight up. Patience and stomach are always part of the equation. Great operator, great run.

[9] WTI futures curve: mid-to-high 70s by year-end.

The WTI futures strip is currently pricing the mid-to-high 70s by year-end. Moving target, but the curve continues to reflect the market’s expectation that conflict pressure eases over time.

The daily tell for us remains the front-month. Every week without escalation slowly repricing what has been a big source of uncertainty since February. Watching closely.

Looking Ahead

Earnings acceleration. Big Tech next week. We want to see whether AI features are converting to software revenue in the same way ServiceNow showed us this week despite getting hit.
Conflict resolution. Every week without escalation changes the probability table. Oil futures remain the daily tell.
Breadth. Still thin. We want to see more stocks participate in the ATH.
Fed speak. Watch for commentary on the Composite PMI rebound and whether it shifts the tone. Lets see…

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing Forward.

Speak to us:

Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube": Channel

🎙️ Avory: Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Rise of Prediction Markets Is Just The Start of What COULD Come…

Sean D. Emory — Wed, 22 Apr 2026 12:32:10 GMT

Regulation can change everything but sports betting apps, insurance companies, and traditional financial rails must pay attention now! Things are coming together…

11% of US sports bettors now open Kalshi.

Why? Because this is already happening.

The easiest way to see it is not through theory. It is through behavior. We can speculate all we want, but the data is already moving: roughly 11% of U.S. sports betting users now also open Kalshi, up from roughly 1.5% to 2% in early 2025. That is a very large shift in a short period of time.

More importantly, prediction markets are no longer sitting outside the sports betting ecosystem. They are beginning to merge with it. What started as “betting” or “predicting” elections is evolving into something much broader and, over time, likely much more meaningful for society.

I think some people believed this could happen. But this may be the first real sign that the category boundary is starting to give way.

Kalshi is no longer niche.

Kalshi reached 21% MAU share within the combined sports betting cohort in QTD 1Q26, up from just 3% in 1Q25.

That is not niche anymore. That is a real part of the market mix.

And it did not happen in a vacuum. FanDuel, DraftKings, BetMGM, Caesars, theScore Bet, and Fanatics all lost share over that same stretch. So this does not look like simple category expansion. It looks more like encroachment.

An Uber analogy is interesting here.

It is not a perfect comparison because that market led to direct category expansion, but I’d say that the lesson still matters. Legacy structures can remain important until a new model starts routing around them. Sportsbooks still operate inside a more constrained, state by state framework. Prediction markets may be building a parallel access layer that reaches users faster and more broadly.

That does not mean incumbents disappear. But it can absolutely change who captures growth at the margin.

And think back to Uber. It was criticized. Cities pushed back. Regulators tried to block it. But users loved the product, and that mattered. Over time, Uber did not just enter the market. It exposed how broken the legacy system was. The old structure had become more focused on protecting its position than building a better product for users.

Prediction markets feel similar, but potentially much bigger. And here’s why:

The distribution advantage

Part of this rise is structural, and part of it is being paid for.

Structurally, prediction market platforms have had more room to operate across geographies where traditional sports betting remains restricted. I think this matters because habit businesses are often won early. The first product a user installs, opens, and trusts during a live event has a real edge.

But Kalshi is not only benefiting from broader access. It is also spending aggressively. More than half of its downloads in QTD 1Q26 came from paid channels, well above the sports betting cohort average.

So yes, some of this adoption is being bought. But that does not make it less important. It may actually be the point here.

Prediction markets are no longer spreading only through niche internet communities. They are now being pushed into mainstream consumer awareness.

Sports are the wedge, not the endpoint.

This is where I think the story gets much bigger than Kalshi versus DraftKings.

Sports may just be the easiest wedge.

What prediction markets are really doing is turning outcomes into tradable exposure. At first that looks like betting. Over time it starts to look like something broader: speculation, yes, but also hedging, protection, and insurance.

That is the part I do not think people fully appreciate yet.

Because once people get comfortable using markets to express a view on a sports outcome, it is not a huge leap to imagine them doing the same thing around weather, inflation, travel disruption, political events, business conditions, crop yields, local economic shocks, or even highly personal risks. At that point prediction markets stop looking like a gambling adjacency and start looking more like a flexible risk transfer system.

That is a much bigger story.

From speculation to insurance

The most interesting version of this may not be betting at all.

It may be insurance. Not insurance in the traditional sense of heavy underwriting, deep intermediation, and slow moving claims processes. Insurance in the practical sense: protecting yourself against an event that would hurt you financially or personally.

That could be macro.

A company hedging a rate move.

A household hedging inflation.

A traveler hedging disruption.

Or it could get much more micro.

A hurricane.

A delayed wedding.

A heat wave.

A regional crop issue.

A missed local event.

If these markets become liquid enough, simple enough, and trusted enough, you can start to imagine a world where people do not just speculate on outcomes. They insure themselves against them.

That is where prediction markets start to become much more important than sports.

Just look at the size of the insurance market.

Why tokenization matters

If prediction markets become broader risk-transfer systems, then tokenization starts to matter a lot.

Because tokenization can turn fragmented exposures into programmable, tradable, and composable units. It’s going to happen. And that matters if the future of this category is not just “who wins the game?” but “how do I hedge a specific real world outcome that affects me?”

The more granular the exposure, the more important the infrastructure becomes.

You could argue tokenization provides that substrate. A way to represent risk, transfer it, settle it, and eventually bundle it across many forms of real world exposure. Decentralized, non decentralize, it doesn’t matter, that’s a side convo.

Truth is we are still early but logic and gravity are going this way. But that is part of why this category feels much bigger than most people think. Throw in AI agents also I can already see it.

but look sports may be where the consumer behavior shows up first.

But the longer term opportunity may be building the rails for a much broader market in OUTCOMES.

Conclusion

JUST TO PUT A BOW ON IT…

Prediction markets are already here, and they are moving faster than many expected. As I mentioned sports may be the first mainstream consumer entry point, but it is unlikely to be the endpoint.

If this model keeps gaining trust and liquidity, the category could expand far beyond gambling and into hedging, protection, and insurance like use cases.

And if that happens, tokenization becomes more than an adjacent theme. It becomes part of the infrastructure layer that helps the whole system work.

About Avory & Co.

Avory & Co. is an investment firm focused on investing forward.

or contact us at team@avoryco.com

To learn more, visit www.avory.xyz

To explore more of our research, analysis, and conversations:

Podcast: Around the Desk

Website: www.avory.xyz

Email: team@avoryco.com

LinkedIn: Avory & Co.

X / Twitter: @_seandavid

YouTube: @AvoryCo

Substack:

We publish ongoing research, market commentary, earnings analysis, and long-form discussions with founders and operators shaping the next decade.

Disclaimer

This article reflects the personal views of the author and is provided for informational and educational purposes only. It should not be construed as investment advice, a recommendation, or an offer to buy or sell any securities. The opinions expressed are subject to change without notice and are based on publicly available information believed to be reliable, though no representation is made as to its accuracy or completeness. Any forward-looking statements involve risks and uncertainties, and actual results may differ materially. Readers should conduct their own research and consult with a qualified financial advisor before making any investment decisions.

$HOOD $XYZ $COIN $SPY $QQQ $ETH $BMNR $IWM

Another Example of Traders vs Investors.

Sean D. Emory — Sun, 19 Apr 2026 21:45:17 GMT

Big Picture

I think it’s clear at this stage that our view that peak uncertainty was likely behind us is playing out. This past week added further confirmation...

We remain optimistic, while recognizing that sentiment can still shift quickly on every headline. That is part of markets. Staying the course is often the right approach, especially when the backdrop feels most uncomfortable.

Markets moved higher just as sentiment felt close to its worst. Again that is a good reminder. The VIX is now at 17. Just six weeks ago, it was nearing 31. This was one of the six largest single-day VIX declines on record. That is the market repricing a meaningful risk premium that had been building since February.

What makes this even more interesting is positioning. Hedge funds remain defensive, with the long/short ratio sitting at 1.96x, a cycle low. In other words, a lot of capital was still positioned for further downside, and the market moved the other way. That kind of disconnect can create real opportunity for patient capital.

Some may ask whether markets now look expensive. That is harder to argue when all-time highs are being accompanied by only 12 companies making new 52-week highs. To us, that likely points to meaningful dispersion beneath the surface.

Simply put, this does not look like a market where everything is expensive and moving together. It looks like a market with selective leadership and plenty of opportunities still out there for investors.

1. The VIX collapsed. One of the six largest single day declines on record.

Fear peaked at 30.75 on March 27. By April 8 it dropped 19.15% in a single session. That is one of six such moves in the VIX’s history. Shows again how fast things can change and not to overreact to single headlines.

Today it sits at 17.84. Normal range. The gyrations are not over but the direction seems clearer to us.

2. Hiring is healing. Four straight weeks of improvement.

The ADP NER Pulse 4 week moving average hit +39,250 private sector jobs added for the week ending March 28. That is the fourth consecutive week of acceleration.

To put this in context, in January that same average was sitting at 4,250. It has nearly 10x’d in 10 weeks. The labor market is quietly getting stronger.

3. Core PPI came in at +0.1%. Expected was +0.6%.

This was a very constructive reading this week. Headline PPI was elevated, obviously energy impacts. But strip that out and core came in at a fraction of the consensus estimate. .5% below what was expected and essentially flat.

This is exactly what we have been saying. The inflation story is an energy story right now, not a demand story outstripping supply like 2021. The underlying machinery of the economy is not overheating. That should keep the Fed on the sidelines assuming escalation continues.

4. The consumer is still spending. Redbook at +7% Y/Y.

Redbook same-store retail sales came in at +7.0% Y/Y for the week ending April 11. That number has been above 6% every single week this year. People keep waiting for the consumer to crack. It is not happening.

Part of why? Tax refunds. The average refund this season is $3,521 tracking around +14% from last year. About 62.9 million returns processed. That is roughly $200B flowing back into consumer pockets in a compressed window.

5. Tax refunds up 14% Y/Y.

As just mentioned. 14% above 2025. This is good, not bad.

It is not complicated. $200 billion in refunds hits checking accounts right as people were supposed to be pulling back. The consumer keeps spending because the consumer has money.

6. Tech valuations have reset.

This one is pretty cool.

Tech (XLK) is trading at forward of 25-26x P/E today. The S&P 500 broad market is at 20x. That gap is near the smallest it has been in years. It has compressed meaningfully relative to its own history, below on a 5 year and above on a 10 year.

This is not a call to buy tech blindly. It is a signal that the premium you used to have to pay is less.

7. Market breadth is thin. And that matters.

The S&P hit an all-time high. 12 stocks made new 52-week highs that day. Not normal.

The historical average on an ATH day is 50+. This sub 12 companies has only happened 6 times since 1999. Thin breadth on a new high is not a sell signal, but it is a signal. For us this means more opportunities not less. Does that make sense??

8. Hedge funds are at a cycle low in exposure.

Funny. Coming int week hedge funds were really short. L/S ratio of 1.96x. That is the lowest it has been in years. Funds de-risked hard through the conflict uncertainty and the index ran away from them. Patient long-term capital owning quality companies through the noise is the structural advantage here. We spoke about this in our last two editions and it keeps playing out.

9. Bank earnings kicked off. The read is constructive.

JPMorgan, Wells Fargo, Citi, and Bank of America all beat this week. All four.

The consistent thread: consumer spending is up, delinquencies are in line, corporate debt levels are healthy, credit quality is stable. JPM was constructive on both sides of the ledger. BofA called out accelerating consumer spend specifically.

Looking Ahead

Earnings season continues. Some of the real signals over the next 3 weeks is whether AI features are actually converting to software revenue.
Conflict resolution. Every week without escalation changes the probability table. Oil futures are the daily tell for us. Following the story has not been a staright line by any stretch.
Breadth. We want to see more stocks participate in the ATH.

That’s all for this week!

Know another investor who'd find this useful? Forward it along or share on your page!

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Ceasefire talks a good start, earnings coming soon.

Sean D. Emory — Fri, 10 Apr 2026 14:23:43 GMT

Two weeks ago we said peak uncertainty was likely behind us. Last week the data supported that view. This week it helped to confirm it. The gyrations continue. Oil just posted its second largest non-COVID single-day decline in 30 years on ceasefire talks. Only September 24, 2001 was bigger. That is the market pulling out the risk premium it had been charging since February 28. And the reaction across equities was also pretty well noted. Energy gave back gains while everything else pushed higher. Communication Services, Consumer Discretionary, Tech, and Financials were all up, reversing the prior months move. Key though is that the rotation tells you a lot. Investors are binary and short term. The sectors hit hardest during the war were the first to snap back.

That is why missing the best days is so costly. Trying to time markets is tough, data doesn’t support it.

Anyways, underneath the war headlines, the fundamentals have held. The consumer is still spending. Jobs remain stable enough. Real-time inflation is running at 2.3% and looks pretty anchored. March CPI tomorrow (this is Thursday night) may look hot because of gasoline, but Truflation continues to suggest the underlying disinflation trend is intact but perking.

Either way we remain constructive. The path still looks like resolution, and the data continues to support positioning for what comes next.

This Week’s Data

1. Oil posted its 2nd largest non COVID single day drop in 30 years.

Only Sep 24, 2001 was bigger. The Iran ceasefire talks triggered a -15.3% single day move in oil.

Then you have…

Iraq War begins: -14.1%

Saudi Financial Crisis: -12.8%

Russia Ukraine Spike: -12.0%

But overall, this was the market repricing the war risk premium very quickly and also just shows how binary the market has become. We have spoken about this numerous times.

2. The consumer is still spending. Strip out gas and the picture is healthy.

Consumer spending index for March is out, and the Y/Y picture still looks pretty solid. Overall spending rose +3.1%, with online only spending up +8.4% and accelerating, nonstore retailers up +9.0%, and air transportation up +6.5%. Gasoline stations were up +17.1%, which is really the war showing up in the data.

What is most interesting to me is what is not collapsing. Most categories are still flat to positive, which suggests the consumer is absorbing the gas price shock without broadly pulling back. That continues to support the idea of a healthy consumer balance sheet.

3. Real time inflation remains anchored at 2.3%.

Truflation today: 2.30% Y/Y as of April 9.

Housing rents: -1.7% Y/Y per Apartment List.

March CPI prints tomorrow. It will look hot because of gasoline. That is mechanical and expected. But Truflation has historically led the official CPI by roughly 6-12 months and it is sitting at 2.3%. Strip out energy and the underlying disinflationary trend is intact. Housing is the biggest component of CPI and it is moving in the right direction.

4. Tech layoffs nothing like 2022

I keep hearing about fears of AI and jobs yet 2022 was really the year of the tech layoff story. We will track close and update accordingly.

5. Cash App is winning the U.S. download race.

Had to throw this in. We own Block. Cash App: +27% M/M in downloads.

PayPal: declining Y/Y.

Venmo: flat.

Cash App momentum continues here.

6. Claude product announcements but 97% uptime…

I am being a little tongue in cheek here, but Claude keeps launching new products while API and site uptime sits around 97%, which is not where you want it if you expect important workflows to run through those pipes.

7. The AI Paradox.

97% of companies say AI is a priority.

23% have actually deployed anything at scale.

Looking Ahead

Tomorrow April 10: March CPI: First reading to fully capture the oil shock. Let’s see. Expected to look hot. Watch core ex shelter to see if the underlying trend holds. Truflation says it does.
Q1 Earnings Season: Kicks off next week. Guidance matters more than results probably. Companies that signal resilience through the conflict will be rewarded. FactSet is projecting 10%+ earnings and revenue growth for the broad market.
Iran Conflict: Ceasefire talks are moving. Every week without escalation changes the probability here. We are watching the oil futures curve daily as you may have thought.

Have a good weekend!

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

The Case for Optimism is Building. Here Are 14 Data Points.

Sean D. Emory — Fri, 03 Apr 2026 15:46:06 GMT

Last week we wrote that we believed peak uncertainty was likely behind us. This week’s data reinforces that view.

The most telling signal came on Tuesday, when markets got a whiff of ceasefire progress. In a single session, the growth and economically sensitive parts of the market surged 3 to 5%. Communication Services, Consumer Discretionary, Tech, and Financials all ripped higher. Meanwhile, Energy gave back gains. That rotation pattern is not random. It is the market telling you what it expects to happen when the conflict eases.

This is exactly the kind of day that makes timing the market so costly. The data on missing the best days in markets is clear: those best days almost always come right after the worst days. They cluster in the most uncomfortable moments. If you are sitting on the sidelines waiting for clarity, you miss the move. And missing just a handful of those days destroys long term returns. The probabilities favor the optimist who stays the course, not the pessimist who tries to trade every headline. That’s the stats.

On the inflation front, real time data continues to paint a more constructive picture. Yes, energy prices will push inflation readings higher over the next several months. That is mechanical and expected. But the real core of inflation is housing, and shelter is flat to falling. CPI shelter is at 3.0% and decelerating. Real time rent data from Zillow and Apartment List is flat to negative. The biggest component of the inflation basket is moving in the right direction. Meanwhile, Truflation’s real time CPI shows that the entire recent jump was driven by gasoline. Strip that out and the underlying trend remains disinflationary. Google Trends data on gas price searches tells an interesting story too. You will see that below.

The labor market is bending but not breaking. ADP beat expectations at 62K. Challenger layoffs dropped 78% year over year. Job openings crossed above their 3 month moving average. The consumer is still spending. Small businesses posted gains in March.

And the forward curve is doing the work for us. The oil futures market is pricing WTI back to $72 by December, near pre war levels. The market is not pricing a permanent disruption. It is pricing a shock that fades. Could be longer for higher, sure but not now.

When you put it all together, the case for staying invested is stronger than the case for selling into fear. The fundamental economy is healthy. The inflation impulse is energy driven and temporary. The technical setup is coiled for a move higher. And the sectors with the strongest earnings growth are the ones most discounted by the war.

We continue to believe the direction is toward resolution, and that the data supports being positioned for what comes next.

This Week’s Data

1. Missing the best days destroys returns.

Over a 20 year period, the S&P returned +7.8% annualized. Miss the 10 best days and it drops to +4.1%. Miss the top 40 days and you are negative at -2.3%. The best days tend to cluster around the worst moments. Exactly the kind of moments we are living through right now. Sitting on the sidelines has a measurable cost unless you are Oracle from the Matrix.

2. The best time to invest is when it feels the worst.

This data was fascinating. Every major crisis low in the past century produced extraordinary 5-year forward returns.

Great Depression: +367%

Severe Recession (1982) +267%

Aggressive Fed Tightening (1994) +251%.

Great Recession (2009): +178%

Just more evidence that market turbulence is just that, turbulence that’s all.

3. When ceasefire hopes emerged, growth sectors ripped.

In just four trading days (March 27 to April 1),

Communication Services gained +3.9%

Consumer Discretionary +3.9%, and

Tech +3.8%.

Meanwhile, Energy collapsed -5.7%.

This is what the rotation looks like when the market begins to price in de-escalation. Financials, Industrials, Materials… everything cyclical moved. The sectors that got punished most during the war are the ones that snap back fastest. And it happens quickly. If you wait for confirmation, you miss it. Just a reminder.

4. The futures market is pricing oil back to $72 by December.

More data. The WTI forward curve tells a simple story. May at $111, June at $97, and a steady decline to $72 by December, which is near pre-war levels of $68. This is steep backwardation. The market does not expect a permanent disruption. It expects the shock to fade, consistent with the incentive structure we outlined above. By Q4, the oil premium is largely gone in futures pricing.

Yes this moves and we are tracking it. So will update.

5. Housing is flat. That is the most important inflation signal.

Everyone is focused on oil-driven inflation. But shelter, the single largest component of CPI at ~40% is running at 3.0% and falling. Real-time rent data from Zillow shows 2% growth. Apartment List shows -1.4%. Truflation’s real-time CPI jumped to 1.47% but was driven entirely by gasoline. Strip out energy and the underlying disinflationary trend is intact. Housing is the medium-term signal. Energy is the noise for now.

6. Gas price anxiety is nowhere near 2022 levels.

This one is pretty cool. Google search interest for “gas prices” peaked at 36 in early March and is already declining to 21. During the Russia-Ukraine crisis in March 2022, the same metric hit 100. Consumer anxiety about fuel costs is roughly one-third of what it was four years ago. People are paying more at the pump, but they are not panicking. Yet…

7. Small-cap short interest is the highest in 12 years.

Russell 2000 short interest as a percentage of float has climbed to 7.2%, the highest since 2014. The 12-year median is 5.2%. This is extreme positioning. When the catalysts arrive, ceasefire progress, a strong jobs report, better-than-expected earnings, forced covering in small caps will create a violent rally. The shorts are already in. The incremental seller is gone. This is a coiled spring. Let’s see.

8. April is historically the second-best month for stocks.

New month who dis? Over 51 years of data, April has produced the second-highest cumulative return of any calendar month, behind only November. The S&P 500 is currently -4.9% YTD versus a historical average of +2.8% through March. Seasonality alone does not override an active war. But it adds to the weight of evidence. Entering the year’s strongest seasonal window with this much pessimism already priced in is not a bearish setup.

9. Small businesses posted gains in March.

The Fiserv Small Business Index showed both month over month and year over year sales gains in March. ADP confirmed: private payrolls added 62K jobs, beating the 40K estimate, driven by small companies for the second straight month. Redbook retail is at 6.7% YoY and accelerating. The consumer engine that powers 44% of economic activity is still running.

10. The hardest-hit sectors have the strongest earnings growth.

Q1 2026 FactSet revenue estimates: Technology at 27.4%. Communication Services at 12.7%. Financials at 10.0%. These are the three most punished sectors during the war and they have the three strongest fundamental growth profiles. Consumer Discretionary is also strong at roughly 8.5%. Energy is last at 1.6% but has seen strong upward revisions for obvious reasons. The S&P 500 as a whole is projected at 9.7% revenue growth. So fundamentally market is pretty healthy.

11. ADP beat. Private employers added 62K jobs.

The ADP National Employment Report showed +62K private sector jobs in March, above the 40K consensus estimate. Education and health services drove the gains (+58K).

Interestingly, small companies (1-19 employees) led for the second consecutive month. The labor market is not collapsing. It is a slow bend, not breaking. And with inflation components outside of energy remaining soft, the policy balance should tilt toward supporting employment. That’s our take atleast.

12. Layoff announcements dropped 78% from last year.

Challenger reported 60,620 announced job cuts in March, well below the 275,240 recorded in March 2025. That is a 78% Y/Y decline. February was similar at -72% Y/Y. March 2025 was the peak of those DOGE-related federal cuts and tariff-driven layoffs. That wave has passed. Something we spoke about in our annual report heading into the year. The underlying trend in layoff activity is normalizing.

13. Earnings growth remains remarkably strong.

I mentioned this at the beginning, but here’s this… S&P 500 Q1 2026 consensus: 13.3% earnings growth and 9.7% revenue growth. Double digit earnings growth despite a war, an oil shock, and rate uncertainty. This is not a recession earnings profile. Tech leads at 27% revenue growth. Energy revisions are sharply higher. But even ex-energy, the earnings picture is constructive.

14. The agent economy is accelerating.

I obviously have to add something around AI.

AI coding agents. Claude Code, Cursor, OpenAI’s Codex, are emerging as the first clear winners in agent powered applications. The Agent Building Activity Index rose 51% M/M in March, the strongest signal of enterprise AI adoption we have tracked. OpenClaw downloads surged amid China’s push for AI-driven business workflows.

Looking Ahead

April 10: March CPI: The first reading fully capturing the oil shock. Markets are braced for a hot number. The question is whether core ex-shelter confirms the disinflationary trend underneath.
Iran Conflict: Both sides have signaled willingness to negotiate. The gap between signaling and execution is where the noise is. But every week that passes without escalation increases the probability of de-escalation.
Q1 Earnings Season: Begins mid-April. Guidance will matter more than results forsure. Companies that signal resilience through the crisis will be rewarded disproportionately given current positioning. Plus AI talk.

That’s all! Peak uncertainty may already be behind us but we are watching. Keep watching the data with us.

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Peak Uncertainty May Be Behind Us. Here is What the Data Says.

Sean D. Emory — Sat, 28 Mar 2026 05:00:32 GMT

Happy Friday. We’re about four weeks into the Iran conflict, and something feels like it’s starting to shift. Even it is on the edge. Our view is we’ve probably hit peak uncertainty around war. That doesn’t mean this ends tomorrow, and it won’t be a straight line. But when you step back, the incentives are starting to lean more toward resolution than escalation.

Start with domestic politics. Republicans heading into midterms with elevated gas prices just doesn’t work. Every extra week of disruption in Hormuz adds pressure. Then look at who’s actually feeling the impact. It’s not just the US. Japan, South Korea, India… all heavily reliant on those oil flows. Even parts of the Middle East are losing revenue daily if they can’t move crude. Meaning the push for resolution is getting stronger, not weaker.

And then there’s Iran itself. Based on what’s publicly observable, this doesn’t look like something they can stretch out for a long time. So net-net, this feels like we’re starting to move in the right direction. Not clean, not linear, but directionally better. That’s my take.

Also what is interesting is when you zoom out from the headlines, the underlying data still looks pretty constructive. Inflation expectations remain anchored. Consumers are still spending. Job openings are starting to pick back up. Tech insiders are buying their own stock at the highest levels in years. Pension funds have roughly $14B to deploy into equities into quarter-end. Short interest is elevated, which can flip quickly if sentiment shifts. And real rates are coming down, which is supportive even without Fed cuts.

That’s the part that stands out to me.

This Week’s Data

Oil is up 40% since strikes began. But it may have peaked.

WTI went from ~$60 pre-war to a peak of $104 on March 18, and has since pulled back to ~$97. As I’m writing this, it’s moved about $5, so take that as you will heading into the weekend.

The move itself was pretty extreme, +40% in four weeks. But what matters more now is the plateau. Feels like the market is starting to price in some form of resolution, not a prolonged disruption. We may get boots on the ground, but my take on that is that would mean a faster resolution.

Long-term inflation expectations remain anchored.

The 5-year, 5-year forward breakeven is sitting around 2.3%. The 30-year breakeven is at 2.4%. Neither is showing any sign of de-anchoring. Markets view the inflation impulse as temporary, tied to the conflict, not structural. Historically, energy shock inflation passes through in 6-9 months. The bond market is saying this and we agree.

Consumers keep spending. Redbook at 6.7% YoY.

The Redbook Index that measure consumer buying for the week ending March 21 came in at +6.7% YoY, up from 6.4% the prior week. Same-store retail sales are growing even as gas prices rise. Consumer balance sheets are healthier than the headlines suggest it seems.

People are still flying. TSA throughput holds despite the war.

TSA checkpoint data through March 26 shows the 7-day moving average at 2.59M passengers, still running +0.6% above the same period in 2025. The consumer is not pulling back on travel.

Job openings cross above the 3-month moving average.

JOLTS job openings at 6,946K, pushing above the 3-month moving average for the first time since October 2025. The labor market is stabilizing and potentially inflecting. Lets see.

Tech insiders are buying at the highest rate in years.

Corporate insider buys in tech at 25 transactions , highest since 2010. When insiders buy, they are betting their own money. This is skin in the game.

Pension funds need to buy $14 billion in equities.

Goldman Sachs estimates $14B in quarter-end rebalancing. 75th percentile absolute, 97th percentile net. This is mechanical, forced buying heading into Q2. Not a bad setup here.

Real rates are falling. That is stimulative.

While nominal rates are expected higher, real rates are falling, making financial conditions more stimulative even without Fed cuts. The inflation impulse is paradoxically making the backdrop more accommodative. Strange how that works.

Short interest is already at cycle highs.

Now how about this. When looking at the Russell 3000 median short interest it sits at 4.4%, the highest since 2011. When everyone is already short, the incremental seller is gone. This is a coiled spring. Any positive catalyst forces covering and so like we have seen in the past, moves come fast, so stay the course.

VC fundraising hit $80 billion in Q1. Biggest since 2021.

Now money continues to pour in areas. Q1 2026 fund closes hit $80B, largest since Q1 2022. To me this shows the demand environment.

Zillow investor day: AI-native real estate is coming.

Turning to one of our names. Zillow had their investor day and we came away more bullish. AI suite demos, 235M monthly uniques, 80% direct traffic. Once mortgage rates settle, this name could re-rate significantly. But we need to get past the war and concerns on macro first. But they continue to do well and the view of a housing super app is becoming more clear.

Looking Ahead

March jobs report (Friday April 3): The next major data point for the recession debate.
April 9 PCE release: First reading capturing oil-shock price pressures.
Iran conflict: Both sides have signaled willingness to talk. Direction is constructive.
Q1 earnings season: Begins mid-April. We get back to fundamentals!

The data says the world is still moving forward. Peak uncertainty may be behind us. The incentives point toward resolution. And the technical setup, short interest, pension flows, insider buying, falling real rates, is quietly building a case for upside.

Have great weekend!

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Oil, Vix Spikes, What do they Mean?

Sean D. Emory — Sat, 14 Mar 2026 01:37:10 GMT

Happy Friday. There is a lot going on right now. Earnings, war headlines, AI developments, and plenty of macro noise. Adding to that, we got a chance this week to hang out with some of the companies that came down to Miami for Future Proof conference.

On the week I wanted to step back and simply share some of the data we have been looking at. The reason is that historically, even during periods of oil spikes and geopolitical conflict, markets have generally been able to absorb the noise. The key variable tends to be the economic backdrop. The bigger issue is whether the economy was already heading toward a recession. When that is the case, returns tend to suffer. When it is not, markets typically look through the shock.

Right now we see a reasonable case that the economy remains resilient. Oil spikes historically tend to be short lived, and that pattern is even more common during midterm election years. At the same time, tax refund data shows we are running roughly 10% higher in dollar terms even though fewer returns have been filed so far. That suggests more refunds are still on the way.

More refunds mean more liquidity flowing back to households, which can help offset some of the pressure from higher oil and gasoline prices.

On the corporate side, AI continues to move incredibly fast. We use it heavily ourselves. But despite all the noise around AI disruption, we continue to see companies executing well, especially in software and software-adjacent businesses. So while the headlines feel intense, the underlying data still points to an economy and corporate sector that remain fairly resilient.

Here is the summary if you want just that:

Oil spike, what does it mean for markets?
Vix spike, what’s it mean?
Insiders buying back shares in financial sector
+10% growth in tax refunds…

As we always suggest, markets rarely move in straight lines. That’s a basic reality of investing. Yet every so often you see stretches of momentum that signal something changing beneath the surface.

Recently the IGV software index posted a 10-day streak of open to close gains, if you follow us consistently then you know it is something we had flagged as a possibility as bottoming signals were starting to emerge in the sector.

What is key is that when IGV experiences these kinds of win streaks, the forward returns tend to skew positive, with median performance improving meaningfully over the following months and even into the next year. That’s our base case

This week’s Iran headlines added another layer of sentiment fear across markets, which can create short-term volatility, but it does not necessarily change the underlying setup.

Over that time frame software has been up 24% 12 months later with 100% positivity for software and the market.

To add to that software comment. We continue to see tech related job openings perk up.

Now people are clearly concerned about oil. And that concern is understandable. So we went ahead and broke down the data to see whether the worry actually makes sense.

Two things stood out.

First, the share of consumer spending going toward gasoline has steadily declined. Today it is roughly 50% less significant than it was in the 1980s.

Now spikes can feel nerve-racking. But when we broke down the data, a spike like the one we’ve seen over the last three months, with oil up roughly 68%, has not historically had a meaningful negative impact on stocks.

If anything, the opposite tends to happen.

Following similar oil spikes, the S&P 500 has often gone on to post strong returns. +33% on average 12 months later.

Then we asked another question. What happens when oil is above $100?

Brent is currently above that level, though WTI is still below it. Looking at history, forward market returns in these environments have generally not been bad. The median return for the market is about +9.4%.

The key caveat is the economic backdrop. When oil is above $100 and the economy is moving toward a recession, returns tend to be much weaker. In those cases the setup skews negatively.

So it really comes down to your base case.

If you believe a recession is coming, the risk profile is less attractive.
If you do not see a recession on the horizon, the market has historically been able to look through the noise.

We remain in camp #2.

Why camp #2.

First, we shared how gas has become less of a factor.

Second, midterms likely incentive the admin to get prices and wars away from the story.

Third, we are seeing +10% growth in refund amounts and still have more refund checks.

That matters because tax refunds act as a seasonal liquidity injection for consumers. Larger refunds can translate into stronger short-term spending across discretionary categories.

Now markets have been skiddish. Are we peaking here in fear? Hard to tell but the 73% spike in VIX is usually met with healthy returns on the other side. Markets are up median 28% 12 months later with 87% positive outcomes.

So obviously we remain at the forefront of what is happening globally and ensuring our companies are doing their part.

One last datapoint worth noting is that we are seeing corporate insiders at financial firms increase their buying quite dramatically. Historically, this has been a bullish signal.

Financial companies sit at the heart of the economy. They see credit demand, deposit flows, capital markets activity, and business conditions in real time. When insiders at these firms begin buying their own stock, it often reflects confidence in the underlying economic backdrop.

So when insider activity in financials starts to rise, it is typically viewed as a very constructive signal for the broader economy and markets.

We will keep you updated. Go check out our videos on Youtube, where we go deep into these topics.

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Buybacks, Identity, 3% of Job Cuts AI?

Sean D. Emory — Fri, 06 Mar 2026 13:16:50 GMT

Happy Friday. The week started with rising tensions in the Middle East and appears to be ending with those concerns still in focus.

For markets, the key variable is oil, particularly the Strait of Hormuz, a major chokepoint for global supply. So far there has been no material impact on prices, and historically these events tend to have limited longer-term market effects.

On the company side, OKTA, one of our newer portfolio additions, reported strong earnings.

We also saw a wave of buyback announcements across software companies, which we expected as earnings reopened buyback windows. Lastly, labor data showed Challenger job cuts fell sharply, reinforcing the view that while the job market is not booming, it remains stable. Let’s get into it.

Here is the summary if you want just that:

+1.2% spending growth
OKTA earnings. Ai agent identity
Job cuts fall
3% of jobs lost from AI?

This week one of the newest companies in our portfolio reported earnings: MEET OKTA

OKTA is an Identity and Access Management (IAM) platform. It verifies identities and controls who or what can access applications, systems, and data.

Historically that meant employees and customers.

But with AI expanding, AI agents and non-human identities are becoming a new category that also needs authentication, permissions, and governance. That makes identity infrastructure increasingly important.

Meanwhile the financials are improving.

Key metrics:

Revenue: $761M in Q4, +11% Y/Y
Free cash flow: $252M (33% margin)
RPO: $4.83B, +15% Y/Y
Customers >$100K ACV: 5,100
Total customers: 20,000+
Rule of 40: 41%

Profitability is improving while AI agent security is emerging as a new growth driver for identity platforms like OKTA.

Over the past month we’ve been discussing software quite a bit. Partly because many of the companies reporting are software businesses, but also because large parts of the internet, digital economy, and even areas like private credit are increasingly tied to the software ecosystem. During that period we also shared our bottom indicator, which was flashing multiple signals.

So far we’ve gotten the bounce we were expecting. Are we past the worst of the software fears, or was this just a short-term relief rally?

Let’s see if it holds.

This week we also got the Challenger Job Cuts report, and it showed a sharp decline in layoffs.

U.S. Challenger job cuts fell 71.9% Y/Y

That suggests companies are not aggressively cutting workers, which aligns with other data showing the labor market remains relatively stable.

There was also a datapoint showing AI only accounts for 3% of job cuts.

Here’s an update on buybacks. We spoke weeks back that we would likely see a large swatch of buyback announcements. Here’s just a handful in the last 4 hours.

Last but not least. Fiserv spend data came out. Spending continued in February.

So look likes steady consumer. Steady jobs.

Have a good weekend!

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

Software check in, oh and hiring improving…

Sean D. Emory — Sat, 28 Feb 2026 02:16:22 GMT

Happy Friday. This week felt important. We heard from major software platforms, and the takeaway was not disruption or decay, i’d say the take was durability. Seat growth is holding, Salesforce called that out directly. Large deals are getting done, with Zoom’s $100K customers growing 9%. You would think large orgs would delay deals no?

At the same time, macro data continues to cooperate. Real-time inflation is running near 1.4% Y/Y. Super Core PPI Y/Y ticked lower. That gives the Fed room and lowers the odds that policy becomes the source of instability. That is constructive and something we have been calling out.

Then layer in labor data. White-collar job postings are not collapsing, like not at all, if anything perking up. They are growing in areas people assumed were first in line for replacement. Customer support is growing. That seems pretty AI-able to me, yet the postings are rising.

We know AI will reshape work. But extrapolating today’s capabilities in a straight line and assuming immediate displacement feels too simple.

And finally, a reality check. AI models are powerful, but they are still probabilistic systems with meaningful error rates. The data I shared shows that clearly.

As always, we focus on durability, positioning, and where the world is headed, not just where the headlines are today. The software narrative feels too linear, and we are selectively buying names that have been hit indiscriminately.

Here is the summary if you want just that:

Block accelerates to +24% growth
Clear Secure +2 million non health users
Software developer job opening +11% Y/Y
Zoom top 10 deals all has paid AI

Let’s kick it off with earnings from a few of our companies that reported this week. Beyond just being portfolio positions, I see these as real-time signals for broader industry trends.

Starting with Clear. Most people know them from the airport, but what’s more interesting is the non-airport segment, which is smaller but growing quickly. They’re expanding their identity verification platform into enterprise use cases like hospitals, Uber, and LinkedIn verification. The story is shifting from airport convenience to becoming a broader identity layer across both the digital and physical world.

Clear’s non-airport enrollments are growing 57%, with roughly 2M quarterly adds outside the airport channel. They now have over 7M paid subscribers in the core business and landed their largest enterprise deal ever. Stock was up +30% on earnings. If our view is correct that AI increases identity complexity, verification becomes more valuable. Clear is positioned at that intersection.

Block also reported. I am a big Block bull, we have owned it twice, this is the second go around, and we are starting to see real execution. So I am happy. You can see in the chart, Cash App gross profit growing 33%, overall gross profit growing 24%, and Cash App inflows growing 16%.

Key is they added 1M Cash App banking actives, deepening wallet engagement. Making them more and more central to users lives.

The second story was they announced a much needed workforce reduction. Not ideal for those involved, but it looks more like post-COVID normalization, they want to reduce layers and allow for them to drive further top line by moving faster and net net more margins.

Guidance calls for continued acceleration. So happy here also.

Here is quick blurb on how Block also posted record net new seller volume and 2019 low churn in Food & Beverage, which grew 16% Y/Y.

Now onto Zoom. They continue to execute well, and it’s becoming clearer that their competitive position is strengthening. Their AI playbook looks like it’s working.

Customers spending over $100K grew 9% to 4,468 and now represent roughly one-third of total revenue. The contact center business grew high double digits Y/Y. So this is not just a “meetings” story anymore. They’re winning with larger customers and expanding into newer segments.

They also hold roughly $1.5B in strategic investments, likely largely in Anthropic. With valuation changes, that stake could be meaningfully higher today. On top of that, there’s a $3.5B buyback in place, which adds another layer to the capital allocation story.

All of their top 10 deals included at least one paid AI seat, and seven of those deals were taken from competitors. That tells you AI is not just incremental revenue, it’s becoming a wedge in competitive displacement. Eric talked about this on the call.

Workday reported this week. We do not own it but interested in how they are functioning.

What we learned was Anthropic, Google, and OpenAI all run Workday. Workday also expanded with Anthropic during the quarter. If AI labs were replacing enterprise software internally, you would not see that. Instead, they are scaling on top of existing platforms.

Next up in software land was Salesforce. They reported steady results and highlighted seat growth. Yes seat growth. They mentioned how Anthropic and OpenAI are customers. So same story for large software players.

They authorized a $50B buyback. That is meaningful support and reinforces the idea that software vendors may be net buyers during volatility. Also how is AI translating to sales. AgentForce is now at an $800M run rate growing 200% Y/Y. Early, but real monetization.

There’s a loud narrative right now that AI is replacing white collar work. And to be fair, some roles will disappear. History shows that clearly. We once had over 1 million switchboard operators. Today we have none. Technology removes certain tasks. That part is real.

But what’s interesting is what the data is actually showing today. According to Indeed, software development job postings are +11% Y/Y. Customer service, sales, HR, and even therapy roles are also positive Y/Y. That runs directly against the idea that white collar hiring is collapsing across the board.

Worth thinking about…

Lets get an inflation check. Truflation is running around 1.4% Y/Y, down sharply over the last year. So real-time inflation is good.

We just got PPI. Month over month was elevated for headline but Super Core PPI Y/Y ticked lower. Historically, that feeds into Core PCE. So in reality this was pretty good.

Lower inflation plus resilient earnings gives the Fed flexibility. Below I show the relationship between PCE and Super Core PPI.

Lastly, if you think AGI is here think again. For now.

The top AI models score around 50% on broad omniscience-style testing. Impressive. But not AGI.

AI is probabilistic and improving. But 50% accuracy is not enough to run critical systems without human oversight.

So this week software earnings are holding. AI is being monetized. Identity is strengthening. Inflation is cooling. The story is more balanced than the headlines suggest.

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

AI data that’s not like the others…

Sean D. Emory — Fri, 20 Feb 2026 17:11:56 GMT

Happy Friday everyone. And of course another week deep in the AI conversation. We had the AI summit in India with Sam Altman and Dario Amodei sharing their latest views on where this technology is headed. We did not learn all that much more.

At the same time, earnings season rolled on. And what stood out again was that companies largely finished the year strong, and consumers are still spending. The data does not reflect an economy rolling over.

Next week, we hear from a large group of software companies. That will be important. Software is where AI expectations are arguably the highest or lowest depending on the view. This week’s data I think is important to framing the path of the AI story. Some of the real time indicators suggest the outcome may not be as linear as many assume.

Here is the summary if you want just that:

Consumer spend intentions up
Apple CAPEX vs others.
Software developer job opening up.
Websites +30% Y/Y

Corporate insiders at companies are buying their shares at a pace we have not seen since late 2022. Before that was late 2020, and 2015-2016 was the time before that. I’d say 2015-2016 is closest to today when software was also in question at the time. Look at the cluster of buying today and how that compares to other market conditions. Takeaway = favorable.

Now back to AI. This week I asked AI a simple question: the car wash is 100 meters away, should I walk or drive?

It gave me a perfectly logical breakdown about time efficiency and effort. Walking would take about a minute. Driving might actually be slower once you factor in starting the car and parking. Clean reasoning yet completely missing one small detail. The car.

That’s the joke, but it’s also the point I am trying to make here. AI is exceptional at processing the literal question in front of it. What it still struggles with is unstated context. Humans instantly understand that if you’re washing a car, the car has to move. AI optimizes the math.

Just saying…

Next is such an interesting chart. This CAPEX chart tells a story I don’t have the answer to…

Amazon +42% Y/Y

Microsoft +89%

Alphabet +95%

Meta +48%…

and Apple -19%

While hyperscalers are aggressively building AI infrastructure, seems like Apple appears restrained. Hmm… what do they see?

Another narrative worth challenging: software developers are supposedly being displaced by AI. Yet FRED data shows software development job postings on Indeed slightly above levels from the same time last year, recently ticking up to 69.1 on the index. The popular storyline is that AI eliminates coding jobs. The data, at least for now, points to continued demand. It may be different type of work, but still interesting data.

Now what is moving? App creation.

Since the release of agentic coding tools, iOS apps released each month have accelerated sharply, with Y/Y growth moving from low single digits to +40%, +60%, and recently near +68%. On a trailing twelve month basis, growth has climbed to +31%.

The narrative says AI replaces coders. Maybe but the bigger story should be that data suggests AI is increasing software supply.

Here’s another datapoint. Says same thing as above. Lots of creation happening out there.

Next up. Salesforce data. This is looking at API usage for voice and messaging and signals their AgentForce platform. It is seeing some acceleration here. Maybe it’s not going to zero.

Let’s switch it up for a moment.

That data on AI should at least make you pause and question the simple “AI eats the world” narrative.

If the story were that linear, the data would likely look a lot cleaner. At the same time, let’s be clear: we believe AI is one of the most important technologies we will see in our lifetime.

Here’s spending trends. Softlines, which include apparel and footwear, are starting to show renewed traction. On a 2-year basis, spending intentions are running +3.7% Y/Y. Consumers are still leaning in on discretionary categories.

Now is inflation a factor? 32% of those surveyed say they are not seeing price increases on apparel. Down from +40% couple years ago.

Last but not least, ChatGPT’s business model is starting to look a lot more like search.

Ads are now appearing on the platform, and early indications suggest they are commanding some of the highest CPMs per 1,000 impressions. High CPMs imply high intent. And high intent is what advertisers are willing to pay for.

But here’s what I’m thinking. If ChatGPT leans into an ad-driven model, its incentives shift. For the platform to win, advertisers need to win. They need conversions. They need revenue. They need ROI.

That dynamic can create a locking-in effect. Instead of replacing every company on the internet, the model may increasingly prioritize enabling businesses that perform well within its ecosystem. In other words, success on the platform becomes mutually reinforcing.

That doesn’t mean disruption disappears. But it does complicate the simple narrative that AI will just disintermediate everything in its path. this business model can shape behavior.

See you next week.

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.

AI = Why are these tech names dropping…

Sean D. Emory — Fri, 13 Feb 2026 15:25:52 GMT

Happy Friday everyone. Busy week at Avory as earnings season continues. Ironically we heard thesis confirming results from Zillow, Paycom, Omnicell, yet reactions have not been great. Currently, AI has introduced a level of noise across market segments that feels wildly exaggerated (see the first image below). In many ways, the “AI will replace everything” sentiment mirrors the 2020 belief that the entire global workforce would never step foot in an office again.

I’ve put together some thoughts on the current state of tech and software, specifically addressing:

What can slow down the “everything will be AI’d” narrative.
and why Avory is uniquely positioned to thrive even in the extreme scenario where AI disrupts all of software.

Outside that and important but feels irrelevant this week but jobs came in at 170k+ job adds, which again if you look back at our annual letter, we stated how jobs will accelerate in 2026. Also Bill Ackman made a $2B investment in Meta this week (one of our holdings), stating it is the best AI name, something I have been stating for nearly 2.5 years, and now people believe it. Sometimes narratives take time.

Here is the summary if you want just that:

+~170k+ jobs added
Zillow revenue grows +16% Y/Y (rentals +39%)
AI narrative analysis
Zoom 10x

This image basically summarizes the last 3 weeks. AI controlling the narrative in markets.

This time, it was the announcement of a new AI-powered tax planning tool within the financial services ecosystem, and within hours brokerage stocks sold off sharply. Charles Schwab, LPL Financial, Raymond James, and Stifel all dropped and have continued to fall since.

There was no earnings miss and no guidance cut, just perceived disruption. Over the last three weeks, the pattern has been consistent: a new AI headline hits, the market assumes disintermediation, and stocks sell first while analysis comes later. We don’t own those but they have no business going down. It has very much been a shoot first, ask questions later environment.

At the same time, strong earnings results have not been rewarded across many sectors. Companies can post solid growth, stable margins, and constructive outlooks, yet the stocks struggle to gain traction. I think that’ll change and I share how and why below but unless you own something clearly defensive like McDonald’s or Walmart, the last couple of weeks have felt disconnected from fundamentals.

I agree with Orlando Bravo from Thoma Bravo. Here’s what he just stated.

Now I wanted to write a more complete piece today on AI and how we think the market, particularly in tech, can climb the wall of worry.

First, AI is not a passing theme. We are not going to stop talking about it this year or next year. It is becoming embedded in how companies operate, how software is built, and how time is allocated across teams. This is not a feature cycle. It is a structural shift in productivity and workflow design. We use it a lot and we have talked about this for a long time and we believe we are well positioned for that world.

We think our portfolio is well equipped to thrive in an AI world. Why tho?

For one, our largest holding is Zoom. Zoom has $8B, no debt, is seeing growth as Zoom Phone and Zoom Contact Center perform well. But that is not the story I wanted to bring up.

If we extrapolate the idea that AI eats software, then Zoom could be worth many multiples of where it trades today. 10x more in an extreme scenario. On a 20% holding, that is 200% to the portfolio. Meaning we have a deep out-of-the-money call option on frontier AI concentration.

Here’s some math. Zoom invested $51M in Anthropic, which is the leading frontier AI lab. Here you see they are gaining massive share. They also have investments in Perplexity and Coreweave, but that’s for another day.

That investment may look small in isolation, but today that investment is worth around $3B.

Let’s take the extreme case to understand the asymmetry here.

If AI meaningfully displaces large portions of traditional software and even parts of knowledge work, the value capture at the model layer could be enormous. For simplicity, assume the U.S. opportunity ultimately consolidates into a $20T duopoly. If Anthropic represents half of that structure, that implies a $10T company, not far off from today’s value of the chip company providing GPU’s.

In that outcome, Zoom’s stake becomes highly relevant. For the $51M investment it would be worth $100B in a $10T Anthropic scenario. If ownership were in that range and survived dilution over time, the Anthropic stake alone could be worth roughly 10x Zoom’s current equity value.

From a portfolio construction standpoint, the implications are material. If Zoom represented 20% of a portfolio and that position increased 10x, that component alone would translate into roughly a 200% contribution relative to the original portfolio base.

Of course, this is an intentionally exaggerated scenario. It assumes AI captures trillions in equity value, the market consolidates into two dominant labs, Anthropic secures half that value, and Zoom’s ownership remains meaningful.

That is not a base case.

The broader point is not that this outcome is inevitable. It is for the doomers on AI winning everything, we have an upside option in our portfolio that would significantly outweigh technological shocks.

Now what can slow down the narrative out there and relieve some of the pressure AI is infusing in minds.

Here’s the summary :

CapEx Growth Moderates: Market says no more. ie 2023 year of efficiency.
Valuations Reset: Lower multiples embed downside and create potential support.
Positioning Washed Out: We are already seeing this.
Integration Shows Up in Earnings: Retention, upsell, and AI revenue can reclassify “losers.”
Buybacks Provide Support: Capital returns reinforce confidence at lower valuations.
Security Shock Risk: We will hear of AI related cyber risks.
Macro Stability Helps: A resilient economy reduces urgency around the bearish thesis.

Right now, the dominant tone around parts of software and application-layer businesses is defensive. The story suggests hyperscalers are spending aggressively, infrastructure is winning, software margins and seats can come under pressure, and traditional seat-based models are vulnerable.

So far we have not seen that. If anything we have seen churn and retention improve and these models have been a round since 2022. Yes models are improving, but there is little evidence of weakness in software land…

However, that framing though carries weight because the CapEx numbers are real and the technological change is real. Now narratives can turn when expectations start to recalibrate. So lets dig into how or why they recalibrate.

1. One of the clearest catalysts for a recalibration would be a deceleration in AI-related CapEx growth.

Just a couple of years ago, hyperscalers were criticized heavily for aggressive investment spending. Now we are seeing Oracle and Microsoft stocks get hit hard, and they could be the first to adopt a more measured tone. I am not talking about a collapse in spending or a reversal, simply a moderation in the rate of increase. Again, AI is real, but if hyperscaler CapEx shifts from extraordinary acceleration to a more sustainable pace, the optics change quickly. The narrative moves from an AI arms race to capital discipline. We saw this dynamic in 2022 when investors shifted their focus back to returns and efficiency. Remember? So a normalization in spending growth could actually stabilize sentiment around software.

2. Valuation also quietly reshapes perception.

Multiples compress before fundamentals fully confirm any reality. When software companies trade at levels that already assume slower growth, margin pressure, and competitive risk, the bar lowers. Remember valuation is a proxy for durability. Once the market stops paying for peak outcomes, the downside becomes more limited. That does not guarantee upside, but it changes the psychology from panic to reassessment. Look how earnings in blue are pushing higher, yet multiples fell below the 2022 rate hike fears! Adobe, Salesforce, and many others are trading near or below their all-time low multiples.

3. Positioning adds another layer.

If fund manager exposure to certain software segments is already near historic lows and much of the selling has cleared, incremental downside becomes harder to generate. Markets often stabilize not because news improves, but because sellers exhaust themselves. At that point, even modest improvements in fundamentals or messaging can drive a re-rating. You can see here that they are very much out of software. Momentum traders get out of the way while long term holders remain.

4. Ultimately, fundamentals determine whether the “non-beneficiary” label sticks.

If companies begin reporting improved retention, AI-driven upsell, higher ARPU, or incremental revenue tied directly to AI features, the classification shifts. The market does not require explosive AI revenue overnight. It needs evidence that AI is enhancing workflows rather than displacing them. As earnings transcripts reflect integration instead of disruption, the tone can shift. Look at ServiceNow, seeing improved retention. Need more of this.

5. Buybacks and capital return policies can reinforce that stabilization.

One thing to consider is we still are not passed the bulk of software earnings, and therefore they are not in their buyback windows yet. So we could see stability from that. companies cannot buyback shares roughly 30 days prior to earnings.

6. Macro conditions matter as well: 2026 IT spend growth = highest reading since Jan 2022 In a resilient economic backdrop, enterprise IT budgets are much less likely to get slashed. That matters. Stability in the economy is generally a positive for tech, because it reduces the odds that companies start cutting spend across the board.

Right now, it feels like the market may be mixing macro fears with AI anxiety, and I do think that has been partially responsible for some of the recent weakness. But here’s the interesting part. The latest ETR survey shows tech spending is actually accelerating and is at its highest pace since 2022.

7. There is also a less discussed but very plausible 2026 risk: security.

As AI tools scale across enterprises, the attack surface expands. The vulnerability does not need to originate in the model weights themselves. It could come from prompt injection exploits, data leakage through integrations, orchestration layer weaknesses, or autonomous agents interacting with external systems. If a highly publicized AI-related breach occurs, adoption may not reverse, but it could slow. Enterprises would move from aggressive rollout to controlled pilots. Legal, compliance, and board oversight would intensify. I think we see more of this.

Interesting though, such a security shock could strengthen long-term defensibility of the incumbents. In a more cautious environment, enterprises consolidate around trusted providers with robust compliance frameworks and monitoring systems.

The bearish AI story persists only if fundamentals confirm the worst-case version of events: accelerating CapEx with no revenue realization, structural margin compression without offset, meaningful seat declines, and valuations that still assume peak outcomes. Some of those remain possible. But markets do not require perfection to stabilize. I’d say they require expectations to become more reasonable and that has happened.

So look we are well positioned for the AI era. The AI narrative is not ending. The turn for perceived non-beneficiaries likely comes through decelerating extremes, disciplined capital allocation, valuation resets, clearer earnings evidence, and perhaps even a phase of security-driven issues. Not a single catalyst, but a gradual shift in framing throughout 2026 as the market moves from extrapolating disruption to thinking through it more.

That is all folks. I will leave Zillow earnings here. +16% growth, guided to same growth in 2026, margins to improve by another 2% this year, and rentals growing +39% Y/Y. Thesis is confirmed, stay the course!

Have great week/weekend.

About Avory & Co.

Investing where the world is headed.

Avory specializes in high-conviction equity strategies, emphasizing Secular Growth and Transformation Stories driven by exceptional teams. Data guides decisions. We cater to high net worth investors, family offices, and institutional investors. Note: This information doesn't constitute a recommendation to buy or sell any mentioned securities. Avory is based in Miami, Florida with clients all across the globe.

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms.

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast

Disclaimer: Not a recommendation to purchase or sell any securities mentioned. This is for educational purposes only.