Big Picture

First part today is on yields and opportunities, and the second part is what to make of the OpenAi revenue figures we got this week. Big week for both.

To kick off… the yo-yo of the AI trade is almost like the yo-yo of the inflation thesis.

Our view is simple: all of AI should not trade as one trade as there are emerging differences between the power trade versus the chips trade versus the model trade, etc. Now even within those categories there are very big nuances I think investors need to begin paying close attention to as there is lots at stake there.

We have decided to go more into the cyber / identity space as that was always the “easiest” area to see where demand due to AI would flow. All you had to believe was more AI, not what chip or model or generator was being used. More importantly and economically, those are recurring relationships, as opposed to the non-recurring spend in a good portion of the AI infrastructure build.

We will update that within our upcoming quarterly letter which will go out, along with quarterly video so stay tuned.

Now rates have been rising, creating some concerns on the why… I think the why is pretty clear. Deficits cannot last forever, and in theory they should not last at all when it comes to the economic growth we are seeing. In theory, now is the time to balance it. But the wrench is war, which is a spending line, that war is creating some anxiety around inflation through commodity pricing, and the fact that economic activity is strong while we spend more has pushed up yields. Now real yields are up, so the signal is growth + deficits are the main culprit for yield moves. When real rates rise because the demand for capital is strong, that is a strength signal, not a distress one. HOWEVER, we think this is looking to subside somewhat, and the implication of that is what we shared back in May/June on rotation trade back into what we predominately own, and where we think there is real value in the market. We are investors, not traders. So staying close to the trades, but not reacting to the moves is our DNA unless something fundamentally changes. So we think rates stabilize, cyclical areas do better, non AI trade gets a lift. All things we have pointed to.

Some details here. The 10-year Treasury yield hit 5.31% this month, the highest level of the entire cycle, and above the 2023 peak everyone remembers as the moment the bond market had spoken. You will see below some data on this…

We know a couple of things. We have seen the worst 10-year run for Treasuries in about a century. Connected to a drawdown underneath the index that many were not aware of, and it sets up another rotation as we expressed 1-2 weeks back. Rate-sensitive sectors left trading well below their own average valuations. All sets up well.

Our view here at Avory is that the backup has likely run its course for now. Let’s get into the details and data!

[1] The 10-year yield hit a cycle high. But…

The 10-year hit 5.31% on October 5, the highest of the entire cycle going back to 2021, and above the 4.98% peak from October 2023. The move off the February low of 3.97% is the fastest backup since the hiking cycle itself.

But the Fed did not do this one in our view. Quite the opposite, when they cut rates, rates rose, and when they raised, rates stayed somewhat neutral.

The bigger influence is that the market is pricing a stronger economy and a bigger deficit at the same time. The deficit side of that equation is what the next chart underneath is about

[2] The US rolled its debt into higher rates.

Here you can see the average interest rate on total interest-bearing US debt from Treasury Fiscal Data.

It bottomed around 1.6% in late 2021 and has more than doubled since, sitting at 3.53% as of September. Missed opportunity to lock in rates at these levels. Shame on the admin then.

The rise in rates now becomes a headwind to the budget. Which is why it is a risk, but also a known risk that should be viewed with the knowledge that the Fed and Treasury do not want this to rise too much more. Not because of economic activity or inflation, because of a budget. Because a bad budget can crowd out economic activity along with driving inflationary impulses.

So net interest used about 5 cents of every federal dollar in fiscal 2021, and in fiscal 2025 it took about 14 cents, or roughly $970B of interest on about $7T of outlays per OMB’s historical tables.

Worth noting fiscal 2026 just closed in September and those numbers are not published yet, so this series ends at FY2025.

As the green line moves higher, so does the blue, and both need to be resolved here soon

[3] Bonds just posted their worst 10-year run in a century.

This is critical and doesn’t signal a bottom or anything by itself. But the fact that real rates have risen, i.e. rates versus inflation expectations, means real income after inflation is coming in to buyers of bonds. Meaning likelihood for some buyer demand versus sellers.

Here’s the data.

10-year rolling returns on long Treasuries sit at -2%.

Historically the structural troughs were around 0% in 1959 and about 1-2% in 1981, so this is the worst run in roughly a century. You buy a bond for its yield, and when yields spend the next decade rising from the level you bought at, the total return math eventually goes negative. Again and obviously this does not mean yields have peaked, but it does mean the ‘bonds are uninvestable’ sentiment is likely improving.

[4] Under the surface, most stocks already had their drawdown.

Now switching from bonds to what this did to stocks.

Since June 1,

89% of Russell 3000 stocks had a drawdown of 10% or more

54% had one of 20% or more, and

29% fell more than 30% at their worst point.

Semiconductors took it the hardest, with 83% of the group down more than 30% at the lows.

Banks barely participated, with only 46% of them down more than 10%, and utilities and REITs held up too.

…this is max drawdown, not where prices sit today. Some of these names have already bounced. But the message is that the index looked calm while most of the market had a real correction underneath it. That is actually good news as we sit here today

.

[5] Continued Opportunity… Rate-sensitive sectors are trading below their average valuations.

Now on the opportunity we have expressed here.

The sectors that get hit hardest by yields are very cheap now.

Utilities trade at 15.3x forward earnings versus a 17.8x five and ten-year average

Materials at 16.1x versus about 18x

and real estate at 17.9x versus a 19.4x five-year average.

All three sit well below their own history, and these are the parts of the market where the yield move has already done its max damage.

[6] When expansion drives rates, cyclicals win.

What is interesting is that when rates rise during economically positive periods, the outcome for the cyclicals I mention is positive.

When looking at the 2008-2022 backtest, the industries that correlate positively with rising 10-year yields are the same ones that correlate with the manufacturing cycle: banks, capital goods, transportation, materials, tech hardware. Defensives like pharma, telecom correlate negatively with both.

So if this yield move is being driven by expansion rather than distress, and that is our view, the market has already handed you the playbook…

[7] Middle East oil flows are almost back to normal.

Now we move to the inflationary / economic fear. Which is war and commodities.

Data shows that the Middle East crude exports bottomed at about 42% of January levels in March and are back to 16,972 kbpd, roughly 94% of pre-war levels.

The interesting part is how. Hormuz itself is only back to about half its January volume, and the flows simply rerouted through Gulf of Oman ship to ship transfers, Fujairah, and the Red Sea.

The supply was dynamic it seems. We will keep following the story but that is where it sits right now…

[8] OpenAI’s run rate went from $2B to $50B in three years.

Now switching to Open AI. This is a big one and we tried modeling out some numbers here.

Per the FT, their annualized revenue was approaching $50B at the end of September, up from $21.4B at year end 2025, $6B at the end of 2024, and $2B at the end of 2023.

On a recognized revenue basis, the leaked audited financials had them at $3.7B in 2024 and $13.07B in 2025, with $5.7B in Q1 and $6.7B in Q2 this year per The Information and the WSJ.

They then came out and said $70B by year end, since the market did not like their $50B run rate.

So if we draw revenue to year end to about $65-70B by December, on the $852B March financing mark that works out to about 17x the current pace. One thing to watch: the $70B figure making the rounds is an adjusted cloud-partner comparison, not the reported number, so make sure you know which one you’re quoting. Changes things, and honestly cannot wait for the S-1 for these companies…

[9] So then is this peak OpenAI growth rate?

So we model that trajectory and we get what???

Run rate growth peaked at 256.7% in 2025, and our base case steps it down to 203.7% this year, 92.3% in 2027, and 56% by 2028. We have no foresight here, just trying to regress the numbers that are somewhat known.

On the current trajectory, this is peak rate of change for OpenAI, for now, and on a base that is now enormous. To be clear, that does not mean peak AI for the entire space, and we still have breakaway stories like Muse increasing usage over there. Also none of this means OpenAI stops executing well, it is just the math of the high base, and it is why the leaked 2030 plan of $350B of revenue requires everything to keep going right, and ALSO why the roughly $278B of cumulative burn that came with it is already being disputed inside the company. It is tough but if this is the outcome, then lots to think about and why the market maybe did not like it.

[10] OpenAI’s quarterly revenue is still accelerating.

More numbers on it to tie the bow.

Q1 came in at $5.7B and Q2 at $6.7B, both reported per The Information and the WSJ.

Q3 and Q4 are our model built off the ~$50B September pace, which gets 2026 to about $38B of recognized revenue, a touch above the $36B internal projection that leaked in September.

So even in our base case they’re running ahead of their own plan. The only thing is the quarter to quarter step up is flattening.

Net Net

Our read: the yield backup has done its work for now. Real rates rising on demand is a strength signal, not a distress one, and the deficit fear trade is crowding an economy that keeps refusing to slow. So we think rates stabilize from here, which sets up the rotation we shared back in May and June: cyclicals do better, the non-AI trade gets a lift, and the value areas we predominately own finally get their turn. All things we have pointed to for months, and it has been the correct view for now.

On AI, this week’s numbers are a good reminder that all of AI should not trade as one trade. OpenAI looks like peak rate of change on a base that is now enormous, but that does not mean peak AI for the entire space, and we still have breakaway stories like Muse compounding on their own curve. We are investors, not traders, so we stay close to the trades without reacting to the moves unless something fundamentally changes. Nothing this week changed that, forsure.

That’s all for this week!

About Avory & Co.

Investing Forward.

www.avory.xyz / www.avoryfunds.com

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.

Big Picture

Alright, happy Friday! I think it’s time to bring the AI discussion back to the more technical level. Less about Macro this week. It is getting old. So on AI it is where are we, and how are the models actually making money... or not? Anthropic has already confidentially filed to go public, private AI companies keep getting funded at massive valuations, and the economics are starting to matter a lot more. So I think we need a reset to gauge current conditions in a world where AI is dominating much of the oxygen.

So to start here, most of the AI debate still spends its time on what the models can do. Whose model is better? What can one do that another can’t? The wow moments that drive demand. In many ways using scare tactics (Dario and Sam). “Our model is so good we cannot release it”.

If we look just 9 months ago Anthropic and Claude came out swinging, really pushing usage from chat and search toward more agentic workflows. OpenClaw, Cowork, and other similar-ish offerings...

Since then we have Muse and Instinct, essentially always-on personal agents that can actually go do things for you. The distinction today is that these are mostly single-player products. They work for you, not really with your family or your team. It is actually a pretty big technical, orchestration and governance challenge.

Companies we know that have created tools to assist here are Block with Buzz which is open source, and Salesforce with Slack trying to build the shared surface where humans and agents work together. All of this means more workloads, more inference, more tokens. That is obvious.

But less attention has been paid to the cost side of the equation. That is starting to change. The headlines are slowly moving from “what can the model do?” to “how efficiently can the model do it?” That matters because the labs now have to balance three things at once: growth, frontier intelligence, and some path toward economics that actually work. Kind of reminds me of Uber / Lyft days, or WeWork vs its field. Whoever has more money + execution obviously has the chance to become the dominant player.

And that is where things get interesting, at least for me. A lot of these model companies are stuck between being the intelligence layer and being the application. But in just the last month Muse is one of the clearest examples of the application starting to abstract the model away. You are not choosing a model for every task. You are just trying to get something done at a price you can afford. This is how the world has worked forever.

Think about solar panels on your roof while you’re still connected to the grid. Once it’s set up, you are not constantly deciding whether the house should run on solar or grid power. You just want reliable energy at the best price. AI is starting to feel the same way. The model becomes a backend layer and the product is the experience on top. This happened during internet boom, cloud boom, mobile boom. Infra comes forward then takes back seat to everything else.

So this week we are walking the full cost stack of an AI model. Where the build money goes, what each layer costs to run, what happens to demand when price collapses, and where the value ends up when the model layer gets commoditized. We’ll give you our view at the end but the short version is this: we think cost at the model layer continues to squeeze lower. Models become the internet a smart yet commoditized layer to do things. And the value migrates to whoever owns distribution, proprietary context, orchestration, and inference scale.

First, the three places every model dollar goes. You can see that on the first chart.

[1] Every AI dollar goes to one of three places.

At a high level, there are three places compute goes. And yes some readers here like hedge fund managers, and financial analysts or even some of the technology experts know this. But i do think its important to refresh your mental model and ensure the logic from 9 months ago still holds today. To be frank, it’s becoming ‘does logic from 4 weeks ago hold today’ this is moving that fast… so lets go.

Training is where the base model learns broad patterns from huge datasets.

Post-training is where that raw model gets shaped into something useful, through fine-tuning, reinforcement learning, tool use, and a lot of trial and error.

And inference is the model actually doing the work for the user, one request at a time, so it becomes the recurring variable cost that scales with usage.

Plain English: training is med school. Post-training is residency. Inference is every patient after that. Couldn’t think of a better analogy lol.

The lines are getting blurrier because modern post-training can involve the model generating and scoring enormous numbers of its own attempts. Operationally, that can start to look a lot like inference even though it is still part of training.

[2] The famous training run is a small line item.

Now the number everyone quotes is usually the final training run. But Epoch AI’s work suggests that is only a fraction of total R&D compute spend.

Their estimates put OpenAI at 9.6%, Z.ai at 12.3%, and MiniMax at 22.6%. The rest is everything around the run: experiments, synthetic data, failed attempts, post-training, research, and all the work required to figure out what is worth training in the first place.

Think of it like a movie made in Hollywood, or that may be a bad example when we review this piece in 10 years..but the final shoot is expensive, but it is not the whole production budget. That is key the next time someone quotes one giant training-run number as the total cost of building the model. So again keep that in mind and the image below shows this…

[3] And the build bill keeps growing.

Indexed to 2020, the trend for frontier language-model training cost goes from 1 to roughly 1,838 by 2026.

That is about 3.5x per year.

So yes, efficiency is improving everywhere, but the labs are still spending more to push the frontier. The arms race has not gotten cheaper. Building is getting more expensive. It is almost like CAC, cost of acquiring a customer… but that customer is spending less per unit of value you provide…

Now lets flip to the other side of the ledger here, because this is probably the most important chart in the piece. I think at least.

[4] The cost of fixed intelligence is collapsing.

Take a fixed level of intelligence that cost $100 to use in late 2022.

Now hold the capability constant and track the cheapest model that can deliver that same performance over time. Estimates suggest that same level of intelligence costs roughly $0.007 today. That is a 47% decline in cost per quarter, or roughly 13x cheaper every year.

Yes, you read that right.

This is the crazy tension sitting at the center of AI economics right now. The labs are spending exponentially more to push the frontier forward, while last year’s level of intelligence is becoming dramatically cheaper to use. It is wild really.

So both things are happening at once: the level of intelligence is rising, while the cost of existing intelligence is collapsing. That is why the economics of AI can look completely different depending on which side you are staring at. But we know for sure that we do not pay estimates. We pay invoices for this stuff. So let’s look at what businesses are actually spending…

[5] The proof is in real spend data.

So we turn to some latest data. Ramp AI Index now tracks token volume, token spend, and effective token pricing across a sample of businesses using its Token Spend Management product. It is not the entire Ramp customer base, but it gives us a useful real-world read on the direction. Conceptually and speaking to various public and private companies this direction at least lines up.

Since January, token volume in that sample is up roughly 10x. Token spend is up too, but nowhere near as fast. That is the point. Companies are consuming much more intelligence without the bill scaling one-for-one with usage.

That is deflation in actual TOTAL business spend. The next two charts help explain why I think.

[6] Not all tokens cost the same.

So reminder, a token is a small chunk of text, roughly a word fragment, and it is the unit AI is generally priced in. But a token is not just a token.

GPT-5.6 Sol lists cached input at $0.40 per million tokens, new input at $4.00, and output at $20.00. Same model, same moment, a 50x spread between cached input and output.

Cached input is context the model has already processed and can reuse. This matters a lot for agents because they often revisit the same instructions, files, and conversation history over and over. Caching does not make every agentic workload cheaper, since agents can also generate a lot more output and reasoning tokens, but it is one big reason effective token prices can fall as usage scales.

[7] And you choose how long the model thinks.

Reasoning settings are the second lever. Move GPT-5.6 Sol from LOW to MAX and Artificial Analysis has its intelligence score going from 34 to 47, while cost per task goes from $0.26 to $1.99. That is 7.6x the cost for roughly 38% more intelligence on their index.

You do not need MAX reasoning to answer what time a restaurant closes. The better AI products increasingly route the easy work to cheaper models and settings, then spend more compute only where it actually matters.

That is why routing is increasingly becoming a margin line. Cost discipline is becoming part of the product. Also routing is becoming a business, Stripe acquired OpenRouter, and NVIDIA acquired Hugging Face..

[8] Cheaper intelligence means more compute, not less. But still, where is the real value?

Now this helps the infrastructure debate. If intelligence gets dramatically cheaper, demand does not have to fall. It can explode.

So far estimates around AI are that the total computing power of the installed AI-chip stock is growing about 3.4x per year. Separate token-demand proxies from providers and routing platforms have been growing much faster, often on the order of 7x to 30x, with around 10x used as a rough market-wide midpoint. Those demand estimates are noisy for sure, so I would not pretend the exact gap is known.

But the setup is pretty obvious here: efficiency gains can get absorbed by more usage. That helps explain why better models and lower token prices have not automatically translated into less infrastructure demand.

It also does not mean every GPU or data-center dollar will earn a great return. It just means the bear case cannot stop at “models are getting more efficient, so we will need less compute.”

[9] Where it goes: agents doing the buying.

Now take that cheap intelligence and let it transact. We see estimates that if end-to-end agentic commerce scales, which I think it will, U.S. e-commerce could approach $2 trillion by 2030, roughly 11.5% above its baseline forecast. That does not mean $2 trillion of purchases will be made fully autonomously by agents. It means agents could become a meaningful part of discovery, comparison, and checkout across a very large transaction base. We have a view here on the opportunity for this, but that’s for another piece maybe.

Once intelligence gets cheap enough, it stops being a product you visit and starts showing up inside the transaction itself. That is when the economics move beyond model subscriptions and API calls.

[10] The landscape: who builds the model, who builds the product.

The market is splitting into two camps. Model creators build the brain. Agent platforms build the product on top. Below is the landscape as of today, public and private, across both sides. Some are playing both sides, eventually I think all will NEED to play both sides.

Net Net

Training bills keep growing while the price of fixed intelligence keeps collapsing. That sounds contradictory, but it is really the core of the whole thing here. The frontier gets more expensive to build while intelligence gets cheaper to distribute, and lower prices can create more usage.

So I do think the model layer faces real deflationary pressure, both from competitors and from each company’s own next release. But I would not jump from that to saying the model companies cannot make money. They can still capture value through distribution, product ownership, enterprise integration, proprietary context, and scale. Actually I think they will NEED TO to be around a decade from now. The question is how much of that value stays with the model itself versus the products and workflows built around it.

If this trend keeps going, the model layer starts to look more like infrastructure: essential, everywhere, and under constant price pressure. That would push more of the economics toward the layers that own the customer, the context, the workflow, and the transaction. This is why what’s happening with Muse is so important to pay close attention to.

So it is not just who has the smartest model, but who can turn cheaper intelligence into durable economics.

That’s all for this week!

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.

Big Picture

Happy Friday! As promised, lets talk Meta Muse, ChatGPT, and macro. If you want to stay on top of Ai data and macro, and companies, this is the place… Meta Connect was this week, that was fun given Muse launch…

But first, some market talk. Markets have been gyrating at the surface level but even more so underneath. Wait until you see the drawdown chart. The rotations are vicious, very much a trading mentality going on, little Buffett like owning. Now naturally the view is, markets near highs = bad. Or bad breadth in the market = bad. (Breadth essentially measures whether we’re seeing more stocks rise vs fall.) So two debunking data points on that today.

But quickly, look at the numbers on what actually happens near all-time highs. One-year forward returns after a record close have averaged 9.9% versus 9.4% for every other day. Sounds super counterintuitive to some degree, but strength begets strength.

So far this year we’ve gone from only AI working, to AI getting hit very hard and everything else working, to AI again in the last 2-3 weeks, and now I think we are probably heading for another rotation. The potential catalyst is there. As we get oil up, yields up, and potential Fed hikes, the administration is on its back heels here, and there seems to be a desire to get the Iran conflict out of sight so oil prices can come down. More on that below.

But cheaper oil would be supportive for yields, and both together would be supportive for economically sensitive names. Consumer discretionary struggles every time you get commodity and long-bond gyrations. The housing-related space has struggled with the exact same pressure. Remove that pressure and those are the groups with the most room to work. So the moves will likely be explosive, but it’ll take resolution.

The gap is already wide. Over the past year the S&P 500 is up yet consumer discretionary is down -8% to -25%, and homebuilders and real estate names are down -13% to -40%. That is the rotation setup I express on a single chart, and it pairs with a money market complex that has never been this full of cash while stocks are elevated. Historically, as MacroCharts points out, this may be setting up for an interesting moment.

We will also cover more macro that validates we are likely still in a prolonged cycle, along with AI, Muse, and ChatGPT, so read until the end.

[1] New highs have not been a sell signal.

Some market talk, as mentioned, to kick it off. Since 1950, buying the S&P 500 at a record close has been a perfectly reasonable strategy. Forward returns from record days essentially match, and slightly beat, returns from every other day across one, three, and five year windows. The dataset covers 1,537 record closes across 19,304 trading days. Either way, human emotion naturally says hey, highs, time to sell, versus okay, are there other opportunities that have not yet been realized? That brings us to the second chart.

[2] Most of the index is already in a bear market.

Here’s the drawdown chart I mentioned. The index headline sits just 2% below its highs, but 82% of constituents are more than 10% below theirs, 59% are more than 20% below, and 17% are down more than 50%. Index realized vol is 9.5% while the average constituent is running at 34.7%, nearly 4x higher. So the calm on the surface is hiding a lot of churn underneath, which is why the rotations feel so vicious. You can see where that churn is coming from on the next chart.

[3] The AI trade runs through hardware and industrials.

This looks at how sector’s relative returns have correlated with their 74-stock AI Plays basket this year. Hardware correlation sits at +95% and industrial capital equipment correlation at +63%, while consumer staples, healthcare equipment, and real estate are all around -60%, with financials and consumer cyclicals near -45%. Obviously the groups moving opposite AI are the same economically sensitive names we think benefit from a rotation. So again in some ways this is not about fundamentals, and more about trading and timing.

[4] Then the rotation watch.

There are clearly plenty of opportunities if you look around. AI is all the conversation, but clearly with sectors down 10% to 30% from highs, having exposure to those (assuming valuations and fundamentals make sense) can be the other phase of this bull market. These haven’t gone anywhere for a couple of years. If we can get the oil concerns fully in the background, then we think we can see a big uplift in non-AI and non-Tech. Despite us being Tech-centric, the value outside of direct Tech has been interesting for years. If oil and long yields roll over, this chart is the scoreboard for the rotation.

[5] Then the cash pile.

This one is interesting because it seems very logical. I found this chart from MacroCharts, and it tracks flows into money market and cash-like ETFs against the S&P. The insight is simple and it has repeated for a decade: the big cash surges cluster at market bottoms, when people panic. Late 2018, March 2020, and the 2022 bear market all line up. The latest surges are arriving with stocks at all-time highs instead, which is the part worth watching.

Our own read on the Fed’s Z.1 data backs it up. Money fund assets just hit a record $8.4 trillion, with $960 billion of net inflows over the trailing four quarters, and the three biggest inflow years on record are 2023, 2025, and 2024. That much cash at highs, combined with what we mentioned earlier, is the fuel for the next leg, and it could go into economically sensitive names should we see resolution on yields and oil.

[6] The macro backdrop is boring, in a good way.

Now, productivity is such an important data point. Eight-quarter annualized output per hour is running at 2.16%, right on top of the 2.15% post-war average, and well above the 0.9% trough from 2013 that everyone remembers. So we are seeing productivity. Not the promise of AI making everyone 10x more productive, even though I do think the best users of AI are likely 5-10x more productive, but productivity remains firm, which is healthy.

[7] Volatility of volatility agrees.

Then we have the volatility of volatility, and it agrees. The VIX’s own 200-day vol-of-vol sits at 120, below its 2010s average of 127 and its 2020s average of 129. The options market’s stress gauge is calm and has been falling, so that’s not a fear signal.

[8] Capex is outgrowing GDP.

This one is big. The investment side of the economy is not waiting for anyone. Capex is growing at 8.4% against 2.3% GDP growth, a gap that has only shown up in the big investment booms like the late 1970s and the 1990s. So this is healthy too, and shows we are seeing a true cycle at play.

[9] Consumer AI keeps growing.

Now to AI, because it’s critical to our views. Data for August shows ChatGPT at 53 million combined downloads, ahead of TikTok at 40 million and WhatsApp at 38 million. AI apps now sit permanently alongside social and messaging at the top of the charts, with Gemini at 22 million. We are waiting for Meta’s Muse data to come out, and over the next couple of months I expect it to make some noise on these charts.

[10] Muse is number 1 on both app stores.

So far, here’s the data. Meta’s Muse is the new entrant most are hearing about. It went from 78 on the US App Store in launch week to number 1 in under two weeks, and it now holds number 1 on both the App Store and Google Play. This is a big deal. The best version of consumer AI to date. Alexandr Wang and Zuck nailed it. They spoke about it at the Meta Connect conference this week, which we tuned into. Lots of exciting stuff there. And if you listen to our Avory Around the Desk podcast, you’ve heard how we have been on this Meta AI train for some time.

[11] Enterprise AI is still early.

Ok, turning to enterprise. We tuned into the UiPath investor day, and UiPath’s 2026 adoption survey was released. Only 31% of large enterprises have AI fully embedded in the business, and only 29% have the orchestration layer fully in place. The barriers are boring and hard: data quality at 38%, integration with existing systems at 37%, governance and compliance at 33%. But the payoff shows up where deployment is real. 89% of firms with orchestration embedded met or beat their ROI expectations, and 36% expect agents to play a significant role within a year.

So as we have said, consumer or SMB AI is much easier than enterprise. Having enterprise exposure already means you are in position to help the AI transformation for orgs. Think Salesforce, ServiceNow, even UiPath, which are all entrenched in the enterprise. It’s also why last week we spoke about how monumental it was for OpenAI, Nvidia, and Anthropic to be at the Salesforce Dreamforce event. They know that to get into the enterprise, usually the best way is to already be there.

Net Net

We are back on rotation watch. Oil and long yields are the catalysts, and consumer discretionary and housing-related names are top of mind. A record $8.4 trillion of cash sits in money funds, built at highs rather than lows, which is either dry powder or an insurance policy. But productivity is at trend, realized vol is calm, and capex is showing normal bull-cycle behavior. So the play is likely rotation as the cycle continues, versus end of cycle. The big deal this week and last is consumer AI. We are seeing real use cases here, and also seeing what the sales playbook may be for enterprise AI, as it’s still early. More Meta Muse and AI data next week!

That’s all for this week!

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

Make sure to subscribe to Avory Around the Desk Podcast

Listen Here: Apple or Spotify

Big Picture

💡 AI fears, AI launches, Salesforce, Fed, and more!

Happy Sunday! I am here reviewing some of the data and thoughts we accumulated this week.

Lots going on, as we sit here in that awkward period between earnings season. We have a war that continues, we have midterms, we have AI fear mongering, we got a 25 basis point hike, and Meta launched a very interesting agentic product.

Starting with the Fed. They hiked this week. Pretty much something expected, and priced in for weeks. When futures are pricing a move, the Fed almost always falls in line with futures, so that is exactly what happened.

Should they have hiked? Our view is no, and we are in good company on that. Rick Rieder at BlackRock and Mohamed El-Erian have both made the same call.

Hiking does not fix the actual issue, which is mostly energy driven, so yields are going to be tied to what happens in the energy market more than anything the Fed does with the front end. We showed last week that the rate-sensitive parts of inflation are actually running softer than the non-rate-sensitive parts, so raising here mostly buys the Fed a layer of perceived credibility. I am not sure I buy that framing, but it is the discussion with some folks. Also if you just draw out another 5 months, the energy shock will be lapping in the inflation data, and that will naturally take the government inflation rates much lower. I think Warsh knows this… and so does the committee, which is maybe why they only have about 1 hike embedded in their forecast.

As always, we take the info we know and analyze the data to understand whether we should be concerned or stay on course. Interestingly, in periods where you have both a rate-cutting cycle and elevated-and-rising rates, the market has historically performed fine, averaging 11% annual returns 12 months out. That lines up with a view that the economy is holding up and earnings and revenue are growing alongside.

There is real precedent here, so investors should not be overly concerned that rate hikes have arrived. Unless inflation spikes meaningfully in the next months, October is most likely a hold, and the ceiling I see is 50 more basis points of hikes from here, which does not change the economic picture given we were 50 basis points higher inside the last three years.

Under the hood, we have seen a little rotation back and forth between software and semis, and both areas are still executing at the company level. On AI, the “AI ends humanity” narrative that ran from last weekend through Monday pushed a sharp rally in cybersecurity and added some fear to software. I think people are extrapolating. If AI can do that, it can also do software. Then Salesforce Dreamforce happened, with Sam Altman, Jensen Huang, and Dario Amodei all on stage, and that for me says a lot about how strategically important the larger software players are to the model labs. The labs need the distribution, they need the buy-in, and these are arguably their biggest customers.

The smart software vendors are attacking AI in three ways. They are building a harness so businesses can orchestrate and govern AI through a single interface. They are embedding AI across their existing platform. And they are going headless, exposing their database so other AI tools can interact with it inside a customer’s own harness. Salesforce laid all three out this past week.

Lastly… Meta Muse also launched this week, Meta’s personalized agent app, and I have a podcast on it worth queueing up before our take below.

[1] App Store rankings show AI is now consumer default.

Lets kick it off with the more unexpected moves in AI.

Meta launched Meta Muse and boom. #1. More on that.

Another look here is how AI continues to dominate app rankings. Between social, AI and some prediction markets, we can clearly see what consumers are seeking. Good or bad, that is the reality of behavior today…

[2] Meta Crushed Agentic Peers.

Ok so again I have a podcast on what makes Meta Muse special. (Apple , Spotify , YouTube). But more important is the why…

Below is a platform that tries to break down which AI agentic consumer apps performs better on certain key attributes. The overarching conclusion is that Meta Muse is leading in many ways. From speed, to experience, to costs, to capabilities.

This moment for Meta should not be overlooked. They have spent, they discussed importance of AI, and now a product is delivering for them.

We have expressed our views time and time again that Meta is the #1 AI play. They have what we consider the 3 important C’s. Context, Capital, and Compute.

[3] AI referrals are taking share and improving conversions…

You can see the impact on how people search.

AI-driven referrals for high-intent commercial queries continue to grow their share, and the traditional search interface is losing ground.

Now some will say this is a negative for the legacy SEO platforms. I think that is partially true. However, look at the conversion rate. What this says is, when people search on AI, they are more ready to convert than basic Google search queries.

So yes, companies are getting LESS referral traffic, but when that traffic shows, it is 60% higher conversion. My take is that companies that have mostly organic traffic can possibly lean in on referred traffic more, and those that have mostly referral traffic have to probably lean in now on social and more authentic ways to get in front of audiences. This should benefit YouTube AND Instagram…

[4] OpenAI vs Anthropic battles continues…

Under the hood at the model layer, the leaderboard flipped this month.

OpenAI moved back to majority share of OpenRouter wallet volume for the first time since February 2024, with Astra driving this change. Anthropic is still right there, but the crown is back on OpenAI for now.

Time will tell. Not shown here too is just how open source continues to take even more share from these two.

[5] Clear MAU seeing strong monthly active user growth.

Two things here. We know Clear’s identity products are in the airports, along with their emerging and fast growing (5x last quarter) non airport verification business. So this gives us signal for travel broadly, but also their verification business.

Full disclosure, this is an AVRY ETF holding.

So monthly actives or MAU has continued to build.

We have seen a 3 month acceleration just as people question the consumer. For me this says the opposite. We will share more on this later with OpenTable and retail sales data…

.

[6] OpenTable says the consumer is fine.

Now more consumer data… “the consumer is weak", "the consumer is resilient”, “real wages are negative”.

Nope… nope… and nope.

It has now been 3 years we have heard the same thing. A weak consumer. Yet here we are, with most consumer datapoints continuing to chug along. So feels like then that your avg economist is using bad, stale, old, garbage data.

OpenTable US seated diners are running plus 13 to plus 17 percent year over year through late August and early September.

You would expect to see this retrench if consumer was weak…

[7] Hormuz throughput is rebuilding toward normal.

Now in between consumer info and retail sales data, we look at a couple other items.

On the geopolitical front, Kpler shipping data through the Strait of Hormuz shows a recovery.

The data collapsed from around 22 million barrels per day to 3 to 4 million barrels per day this spring, and we are now back near 12 million barrels per day. Not fully restored, but the throughput is trending in the right direction as most continue to say that it is shut down. I do not know where this is headed short term, but nobody really wins here with it being shut down…

[8] Weekly Economic Index is running above trend.

Now if we aggregate some economic data and turn into a Weekly Economic Index we see that it is still running at plus 3.07, and the eight-week change is plus 0.20.

That is materially above the three-year average of plus 2.31, so the growth impulse is broadening rather than fading. Hmm.

[9] ADP weekly pulse is adding real hiring.

Now jobs data. ADP’s weekly pulse is getting stronger, not weaker. The Fed called out the economy improving, I agree. The eight-week trend on private employment sits at plus 11.7 thousand per week, so the labor market is stabilizing, not deteriorating. This is a good thing. But for a second, remember just 3-6-9 months ago every economist was saying that the employment market is breaking? Or that AI would eat 40-60% of jobs by now?

We follow the data.

[10] Retail sales continues to show a resilient consumer.

NOW…

All that showed up this week in the retail sales report. “ Bad jobs market” “weak consumer” ….

And we get 1.2% month over month against a plus 0.8% forecast, ex-autos plus 1.4% versus plus 0.6%, and July revised from minus 0.6% to minus 0.4%.

Almost every segment for retail sales did well…

[11] The whole curve rallied on the day of the hike, but will it hold?

Ok this is getting long, so let me make it quick.

After the rate hike, rates actually fell, which is counter to what you would normally expect. But again, markets move ahead of the action.

Will rates stay lower? Hard to say. There are a lot of variables at play here and abroad. But the Fed and Treasury both want rates to stabilize, and markets do too.

Volatile rates tend to create more equity volatility, especially as quant funds adjust positioning and sizing based on those moves.

[12] High and rising rates have historically been fine for stocks.

This one is easy to read.

History has actually been kind here. In the high-and-rising rate regime we are in now, using the Cleveland Fed’s modeled real yield above 2% and rising, forward 12-month S&P 500 returns have averaged plus 11% . So the setup of higher rates plus a functioning economy has historically been a fine one for equities…

[13] Contrarian alert, investors are the most bearish all year.

And investors are nervous. 53% bears in latest survey. That is near highs. Only 20% of stocks are above 20 day moving avg, so lots of chop has happened here.

[14] Realized vol is back in the calm zone.

Then we have had little volatility, and we should expect some growing volatility with midterms. But low vol as shown on this chart is normal and does not signal good or bad market enviorments ahead…

.

Net Net

The consumer remains firm, so if corporate earnings are a function of the health of the consumer and economy, then our tone should be constructive. Yes the Fed hiked, and yes that is never a positive, but we do know that historically that hikes in an environment where rates are elevated and the economy is growing is met with healthy returns going forward. As a firm we expect this to continue. Lots going on in AI and stay tuned as we bring Meta Muse retention data here weekly.

Have a good week ahead.

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.

The viral use cases are real. People are sourcing cheaper auto insurance quotes, canceling forgotten subscriptions, and getting hundreds of dollars of value back in the first week. Sean explains why the product actually works, and why Meta's computer use approach, running secure virtual browsers on a per-user VM, outperforms the older connector-only stack.

The core thesis: an agent needs three things to be useful. It needs to know you, it needs to reach you, and it needs to act for you. Meta is the only company at consumer scale that has all three natively. Sean walks through his own scheduled Muse tasks, including a monthly cheapest-per-ounce seltzer run across Instacart and Amazon, and a daily baseball-schedule check for his son that quietly adds new games to his calendar.

He also gets into the piece nobody is talking about, the brand inflection. Consumers just handed Meta access to their bank accounts, calendars, and payments. That is a shift the market has not priced. The ripple hits ChatGPT and Anthropic. ChatGPT is going to start looking a lot more like Muse. Anthropic has a real on-ramp problem if the consumer market gets split between Meta and OpenAI.

Chapters
00:00 Muse goes viral
01:28 What Muse is
02:58 Money-saving use cases
04:11 Why it actually works
07:24 Why Meta wins
11:53 My personal workflows
13:18 Brand trust shift
15:11 Competitive fallout
17:31 How Meta monetizes
19:40 Risks and what we watch
21:23 Closing takeaways

Follow Avory & Co.
Website: https://avoryfunds.com
Newsletter: Investing with Data
YouTube: https://youtube.com/@avoryco

Disclosure
For educational purposes only. Not investment advice. Avory & Co. and Sean Emory may hold positions in the companies discussed.

Make sure to subscribe to Avory Around the Desk Podcast

Listen Here: Apple or Spotify

Big Picture

💡 The Fed can only reach half of CPI, Apple's foldable finally ships, and Altman is calling for an AI freeze.

Happy almost weekend! CPI hit so waited for that…

The viewpoints this week are pretty simple and data backed. Growth is fine. Inflation is stuck in the parts of the economy that rate policy doesn’t actually touch. That last part is kind of important. You can see what I mean below in inflation section. And the consumer keeps funneling more of their marginal dollars into new categories like sports betting. Betting looks like the new lottery as per the cohort data below.

Then two other signals I think worth calling out. Apple finally launched the foldable. I like it and pretty excited as I do quite of bit on my phone and large iPad-like is my preference. What is key is that I flagged this back in 2019 as they issued a patent for something resembling this weeks announcement. That alone is a good reminder that if you pay attention to what companies are doing, saying protecting (via patents), who they are hiring, etc, you can possibly see the roadmap years before the analyst community catches up. Lastly, Sam Altman is now floating the idea of an AI freeze right after launching a frontier model, which is a loaded signal about where the buildout goes from here. Ai buildout bros may not like that.

But, then Oracle also put up a backlog number that is massive.

Let’s walk through it. (we spoke about Oura IPO on latest podcast, check it out)

[1] Growth continues in the economy.

The weekly economic index we track continues to show a healthy economy.

Latest reading is +3.27, well above the 3-year average of +2.30, and it’s been climbing for eight straight weeks.

That’s a healthy impulse. On one side it’s good for the economy. On the other side, that keeps the door open for the Fed to move rates.

[2] GDPNow is tracking Q3 at +4.42%.

On top of that, if you want a longer look, the Atlanta Fed’s GDPNow model is now tracking Q3 real GDP growth at +4.42%.

More than half of it is the consumer at +2.43 points, and inventories are adding another +1.99. Business investment is contributing +1.38. Net exports are the big drag at -1.44.

So those two data points together are congruent.

The one thing to watch is the inventory contribution. If inventories reverse in Q4 you can lose two full points of GDP quickly, so this print is a bit stronger than the underlying run-rate.

But even stripping that out, you’re still looking at growth in the mid-2s. Fine.

[3] Producer prices, rising or cooling?

We got PPI this week, which is prices at the producer level. There are two ways to look at it, so I’ll give you both.

Core PPI on a 1-month annualized basis is +1.96%, and the 3-month annualized is +3.36%. Both are the lowest reads we’ve seen in the last 12 months. That lines up with what Waller said last week, which we plugged into the newsletter, that the direction of travel is still on track.

Now if you take the headline figure, yes it looks elevated. The question is whether you look through headline energy or not for now.

[4] Critical: the Fed can only reach half of CPI.

CPI then dropped this morning, which is why I waited to post. Headline tells a similar story to PPI. August came in at +3.40% Y/Y but core at +2.45% Y/Y.

But the mix is important (VERY IMPORTANT), and it’s the thing we want to make clear. Someone send this to the Fed.

When you split CPI into the pieces the Fed’s rate policy can actually influence versus the pieces it can’t, you get a really clean divide.

The rate-sensitive basket, which is shelter, lodging, vehicles, furnishings, and apparel, is running at +2.56%. Basically at target.

The rate-insensitive basket, which is food, energy, insurance, medical, subscriptions, and wage-driven services, is running at +4.54%.

So cutting rates does nothing for your insurance premium, your streaming bundle, or your grocery bill. But it does relieve the parts of inflation that are already near target.

If you take this for what it is, the honest question becomes why cut rates at all if cutting doesn’t touch the sticky pieces, and why risk damaging the parts of the economy that are working or not working.

[5] Sports betting adoption is going vertical.

Sports season is now upon us with baseball and football both in full swing, so I wanted to look at sports betting. We’ve covered prediction market data in the past. Here’s more.

Total betting user growth is +40% since January, and first-time bettors tripled in June and July. This is becoming a material consumer data point.

[6] Frequency is getting interesting.

57% of bettors are placing bets at least once a week, and a meaningful chunk are betting daily. That's the range where I think this stops being entertainment spending and, right or wrong, becomes a fixed line item in the household budget.

[7] And it’s coming from the bottom of the income stack.

Lower-income households now make up 37% of betting activity, the largest single slice.

It's a real consumer behavior shift happening at the bottom of the income distribution, and the timing plus the growth rate says it's not going away, and becoming habitual.

We were talking about this internally at Avory and the read is that online betting has effectively become the new lottery ticket. I'd guess if you pulled the data on lottery spend by income cohort, the distribution would look very similar.

[8] Apple foldable: from patent to product, seven years later.

Something I’ve been talking about for years is how much you can learn from paying attention to patents, not just to earnings reports.

Here’s a tweet I posted back in June 2019 flagging a specific Apple foldable patent that had just been granted. This is exactly why looking into data and areas that aren’t so obvious can pay dividends over time.

And here is what it actually looks like.

Apple launched the iPhone Duo this week at their September 9 event.

Not cheap. It starts at $1,999 for 256GB and goes up to around $3,000 for the 2TB. 5.4-inch outer display, 7.6-inch inner display which is the largest iPhone screen ever, A20 Pro chip, and a new C2 modem, which is Apple’s first real move off Qualcomm. Preorders October 16, ships October 23.

The point is not the phone. The point is that seven years passed between the patent grant and the launched product. If you were paying attention in 2019, you had a data point about where Apple was heading long before any analyst had a model for it.

On the device itself, if I own Apple, which we have in the past, the question is whether this erodes iPad sales over time. I’d say yes on the margin.

[9] Oracle's backlog is big.

Oracle's remaining performance obligations sit at $664 billion as of the most recent quarter. That's up +923% since August 2023. It's quite the buildup.

The important piece is that customer prepayments are becoming a meaningful part of how Oracle is funding its AI buildout.

The company still had to fund roughly $18 billion of net CapEx in Q1, but customers effectively financed more than $11 billion of that themselves.

And on the latest $30-plus billion of AI contracts, Oracle said it needs no incremental increase to its planned capital raise. So lots to build, lots of forward visibility on contracts. But the next section makes this one more interesting.

[10] Altman floating an AI freeze.

I’ll end with this one, because it’s the signal underneath the signal. Altman floating the idea of a freeze right after launching a frontier model is loaded.

A few reads. First, this cannot be good for the AI buildout crowd. Second, is he trying to complicate the Anthropic IPO backdrop and insulate OpenAI from competition?

Third, is this real fear about where AI is heading and what it can do?

Either way, I think this is a more important signal for thew space. It says we may be closer to bending the growth curve than some believe.

To be clear, that’s a view about deceleration in the rate of growth, not an outright slowdown. The absolute numbers can still be big while the second derivative starts to roll.

Net Net

Growth is fine. The parts of inflation the Fed actually controls are already near target. The consumer is finding new categories to spend in, and the AI infrastructure buildout is being partially financed by the customers using it.

Once again, the narrative that this is a fragile economy waiting to break down isn’t showing up in the actual data.

The one thing to keep an eye on is the AI signal underneath. Altman calling for a freeze right after shipping a frontier model, paired with Oracle’s backlog being customer-funded, is a hint that the shape of the buildout is starting to change. Not slower in absolute terms yet, but the growth curve may be bending. With that make sure to subscribe to Avory Around the Desk Podcast + This Newsletter!

That’s all for this week!

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.

He frames the shift as driven by cheaper and smaller sensors, consumers paying more directly for health through higher deductibles, GLP-1s, and cash-pay services, and AI making health data more usable. Using Oura's S-1 cohort data (72% women, 73% age 45 or younger, 63% earning $100K or more, over half with a chronic condition), he explains why Oura won by targeting finger-based wearables, improving signal quality and wear time, and pairing hardware with subscriptions. He runs Oura lightly through Avory's 6M framework and flags retention and churn as the key unknown until more cohort data is public.

Chapters
00:00 Oura IPO Snapshot
01:10 Channel Intro and Approach
02:03 Why Oura Matters Now
03:56 Big Tech Health Moves
04:51 Three Forces Driving Change
07:04 Who Uses Oura
09:56 Winning the Finger
13:31 Subscription and Engagement Loop
16:06 6M Investor Lens
19:59 What Comes Next

More from Avory & Co.
Newsletter (Investing with Data):

www.investingwithdata.com 

Website: avoryfunds.com

Informational only. Not personal investment advice. Opinions reflect Avory & Co.'s views as of the recording date and may change. Avory & Co. and Sean Emory may hold positions in securities discussed. Do your own research and consult your financial professional before making investment decisions.

Make sure to subscribe to Avory Around the Desk Podcast

Listen Here: Apple or Spotify

Big Picture

💡 Netflix attention battle is getting more intense. Labor? Inflation in 2008 was…

Happy Friday! I had a chance this week to spend some time in Philly connecting with fellow asset managers at a conference. Thanks to Wes Gray and Patrick and team for the invite. It was quite different being that it was a marine themed event. Good conversations throughout.

Now today we're covering a couple of different things. As we exit what feels more and more like an extended earnings season, one company we keep thinking about is Netflix, and what’s taking place in the battle for attention. We found some cool stats for that.

Then we'll walk through what happened in AI this week, both on the model side and on the businesses actually benefiting.

We got earnings from Snowflake, a name we don't own, and UiPath, a name we do. Both lifted on AI. From there we can zoom out to the macro as jobs hit this am, and we finish on inflation of course, because the underlying data is a lot more muted when you deconstruct it, so more data there.

Lets get into it!

[1] Entertainment battle goes beyond direct competition.

So as we mentioned, we dug into the Apptopia data to look at the battle inside entertainment, and is becoming more clear that entertainment now spans well beyond traditional media.

The newest form showing up in the mix is micro drama, which is essentially bite-sized vertical video, and you can see the overlap between Netflix users and micro drama apps starting to become meaningful.

This type of entertainment mimics the impulses we all get on things like Instagram. Shorter attention grabbing on a format that’s easier to consume regularly. You can see the rise in users using this that also use Netflix, going from less than 2% to 9% in about a year.

Something to monitor.

[2] Netflix competes with everyone.

The follow-on point is that Netflix isn’t just fighting the streamers anymore.

When you look at the overlap using Apptopia data, between Netflix and YouTube, the story flips from who has the best content to who owns your attention.

Attention is the market, and it’s fragmenting all over the place. Ai will probably accelerate this trend.

The stat is 72% of YouTube users are also on Netflix but that number is falling. YouTube is becoming more of a main consumption platform for all things. Which is why Netflix is fighting back by launching podcasts and more live events.

[3] Meta Muse puts Meta in the model race. Again.

Moving over to AI, we had continued model launches this week. As we get closer to an Anthropic IPO, I think you’re going to see even more of a race to launch models.

We also saw Hugging Face get acquired by Nvidia. Not going to spend time on that specifically, but it’s another signal that you’re starting to see some consolidation and a little bit of maturity taking place in parts of this market.

The one we do want to dig into is Meta Muse. Meta launched Muse Spark 1.3 and it puts them right at the front of the line, in real competition with peers again.

The important thing here is that Meta has a real cash-flowing business that can subsidize this. If you compare that to the AI labs, it’s a much different story. So if you are the labs you should be quite worried again.

[4] And engagement is showing up for Meta.

The model performance is one thing. The other thing worth considering is that Meta AI is climbing the mobile AI ladder.

Some folks aren’t going to love the exact dollar amount being spent here, but it’s showing up in advertising, it’s showing up in people using and engaging with the AI app, and the models themselves are getting more performant. Somewhat congruent with the Netflix section too. More AI = more content. More AI = more relevant content surfaced to users, and more AI = happier advertisers = more spend.

All three at once is pretty impressive.

Again this data captures Apptopia app usage. So pretty good proxy. And directionally is positive for Meta.

[5] Snowflake is a real beneficiary of the AI wave.

Branching slightly away from the models themselves, Snowflake reported and grew 37%. This is a company we really like and follow very closely. What we’re learning is that data is starting to find its place in the enterprise.

The ServiceNows and Salesforces of the world are keeping data on their platforms, and then Snowflake becomes a central hub for governance and data management across the enterprise.

This is so key to the defensibility for software names. Snowflake benefits from the same AI wave as rest of the ecosystem. Valuation is a little bit of a concern, but the demand they are seeing is impressive. Keep in mind it is a consumption model business already, so this isn’t ARR.

UiPath had a similar read on the print, AI orchestration finding real budget in the enterprise.

[6] But dilution is still a story.

Now I have to do it. But one thing worth flagging on Snowflake is that stock based comp continues to be very elevated. They’re funding some of that growth via equity for employees, which is fine for a while and can even be efficient, but it can work against you as we’ve seen before with other names. Watch this line as time goes on.

[7] Labor is stable enough and not cracking.

Zooming out to the macro, jobless claims continue to be modest. Running well below prior years and below historical seasonality. It’s not something we’re overly concerned about. But we also wouldn’t want to raise rates here.

What’s driving the inflationary impulses is less rate-sensitive than the market assumes at the moment, and hikes wouldn’t do all that much. In a strange way, a hike could actually push housing inflation higher. Higher rates raise landlord input costs, and in a steady economy that flows into rents.

[8] Waller signals a hold.

We got a Waller quote from the Fed this week, and the basic read is a desire to hold, with a willingness to move if inflation drifts against the mandate in a more meaningful way. But he flagged that the trend he sees is still favorable. Expectations for hikes in September dropped. Rose a little after the jobs report but still seesawing.

Working forward from here: no hike in October in our view, midterm elections make that a no-no unless you absolutely have to.

By December we’re close to when Chairman Warsh’s committee is going to come back with findings on how they want to manage parts of the Fed, and it’s not obvious you want to move policy right before or right after those findings are known.

So the path becomes challenging, and a lot of it is predicated on inflation staying where it is and slowly migrating in the direction we want.

We agree with a hold should it happen. Here’s why…

[9] Why we’re comfortable with a hold, shelter.

I sound like a broken record on this one, but the reason we’re comfortable with a hold is that the underlying components of inflation are a lot more muted. Shelter is doing the heavy lifting. It’s adding roughly 50 to 60 basis points to the underlying every month.

Current spot rates in housing are much closer to negative 1.5% to plus 1.5%. Net that out at zero and shelter shouldn’t be contributing anything.

Take out the energy impulses on top of that, and you quickly get into the 2.5% range on the underlying.

[10] And shelter is structurally a bad housing gauge.

The reason shelter is such a bad measurement of what’s actually happening in housing: in the last housing crash, housing prices fell double digits, roughly 15% at points during the financial crisis.

Yet the green line here, PCE shelter inside inflation, never actually went negative. You can see the lag, and you can see how it didn’t even turn negative, which is a little crazy and hopefully Warsh is solving for this.

Not even sure how this is seriously considered.

[11] And expectations are still anchored.

Now if you use breakevens to gauge whether inflation expectations are anchored, and Chairman Warsh and most others agree that’s a fair read, they continue to show the same thing. Anchored, just above target.

That gives the Fed cover to be patient.

Net Net

More the same here. Companies continue to report quality results, and we’re starting to see AI demand migrate into more interesting areas. Software in some cases will be a real beneficiary. Some of the AI shakeout has been needed.

The thing to remember: not all AI is good, and not all AI is bad. You want to be hyper-vigilant on what you own and why. We’ve always preached that motto, but in a world with this much euphoria and emotion, you want some grounded discipline underneath the decisions you’re making.

See you next week. Go Canes.

That’s all for this week!

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.

New Podcast Dropped Yesterday

Listen Here: Apple or Spotify

Big Picture

💡 Real wages positive, jobs stable, and the VIX regime is one of the strongest setups for forward returns.

So this week was filled with AI, earnings, inflation, jobs, and some data we analyzed heading into midterms.

The truth is we saw all of these come together, which is very different from just 6-12 months ago when these themes were competing. For example, names we own like Okta, Salesforce, and Zoom all put up numbers showing AI is starting to drive real usage and ROI, and Nvidia’s report help to confirm the whole ecosystem is leaning on them harder, not less. AI and software were not playing well together 6-12 months ago. It was, what’s good for AI is bad for software and vice versa.

Now we’re separating winners and losers. Not only in software, but also in AI where not all AI natives win. (woo!) All of this is very healthy.

We’re also monitoring real-time inflation data that still points lower. Jobs stayed stable (see charts below). And the VIX is sitting in the historical sweet spot for forward returns. There are still two forces we’re watching, one being midterm path risk and the other being what Iran does between now and November, but the current setup is more coherent than it’s been all year. See below.

Lets get into it all!

[1] Software and AI are finally working together.

For the last year, software and AI looked like they were pulling apart. AI was eating into legacy demand, and software names were caught in the middle. This quarter that started to flip.

Okta posted cRPO growth of 14% and RPO growth of 17%, both comfortably above the 11% revenue line, which is the shape you want to see for a steady-growth name.

For reference, cRPO is current remaining performance obligations and RPO is total remaining performance obligations. It’s essentially the contracted revenue a company has already signed but has not yet recognized as revenue. So a good leading indicator.

Backlog is running nicely ahead of recognized demand. Something we’ve been talking about for years now, but this is the first quarter it actually reaccelerated instead of staying in the range. Below is our chart of Okta.

[2] Zoom is starting to monetize AI at the enterprise level.

Zoom put up a record number of 7-figure ZCX deals, ZVA customers grew more than 250% year over year, and 9 of the top 10 ZCX deals in the quarter involved paid AI.

For reference, ZCX = Zoom Customer Experience and ZVA = Zoom Virtual Agent. They’re related, but ZVA is a product inside the broader ZCX platform.

High double-digit ARR growth on this side. This is what I would say monetized AI looks like when it’s actually working, not a demo or a pilot. Zoom also got named a leader in IDC’s MarketScape for agentic contact center platforms, which validates the direction.

Not to mention their Anthropic stake is now $3-4B, pre the IPO, which is said to be valued at up to $2-3T. That would send this stake to $5-6-7B, or 30-40% of Zoom’s market cap. Combined with $7-8B of cash and accelerating enterprise growth, this remains a conviction play for us.

[3] Nvidia continues to dominate the ecosystem.

Nvidia reported $96.2B in Q2 revenue, of which $59.7B fell through to net income.

That’s a 62% net margin on a business now scaled to nearly $100B a quarter. Data Center is 92% of the mix at +117% year over year. The 6 to 18 months of ecosystem investments they’ve made are showing up in demand security, and there’s no obvious replacement in sight. Nvidia still must execute as labs, hyperscalers, and sovereigns are all focused on creating their own purposed built chips. So moral of Nvidia story, do not stay still.

The main thing in the quarter was that they said next year could grow at 70%, which took the stock from negative after the report to positive.

Market sees this as demand that continues.

Now on the same day software stocks rose, which is great news and shows a market that is starting to actually act more rational… The main thing to monitor if you are a Nvidia bull is they guided to low 70% gross margins, this was near 80% a year ago. We think cycle median will sit in the 60% range.

[4] Codex is catching Claude Code. Kind of.

Now switching topics a little bit as this is very interesting… the AI application layer is showing another change of the guard.

Codex from OpenAI is closing on Anthropic’s Claude Code, which is another sign that leadership at the app layer is still being fought over. There’s also chatter that Anthropic may go public late September or early October. If that happens, it’s a boon for the whole AI landscape. More capital in, more build-out, more downstream demand for the names above. Our view remains these labs will have to continue to verticalize. Their own chips, some sort of model orchestrator, and either moving deeper into the infrastructure layer or attempt to go to the top of the stack.

Either way this data shows the growth of Codex at 550% vs 205%. The base is lower so growth is expected, but one is growing and accelerating and the other is decelerating. Both growing fast tho.

[5] Real wages are still positive, and it depends on which inflation you believe.

Now to the macro side. The government’s inflation figures keep getting headline attention, but when you look at real wages using real-time inflation measures, the picture is more constructive. Nominal AHE (wages) at +3.2%.

Real wages using BLS Core CPI: +0.7%. Real wages using real-time Core: +1.4%. Both measures positive, and neither has gone negative through this cycle. Hmm.

The “real wages are negative” narrative is on headline CPI, which is being juiced by energy prices.

[6] Oil futures are still pointing down and to the right.

WTI is $82.84 front month, $70.78 twelve months out. That’s a -14.6% discount to spot at the one-year mark. Not a perfect signal, but the slope keeps saying the market either sees ample supply or doesn’t believe the demand story that’s been used to argue for higher prices.

Either way, the curve suggests the inflation impulse continues to moderate, assuming this holds. This would help the headline CPI, which would improve sentiment on consumers.

[7] The consumer is not levering up.

NOW this one is important.

Household debt as a share of disposable personal income is sitting near multi-decade lows. Not many talking about this…

Despite all the “consumer is cracking” takes, the leverage story just isn’t showing up in the data. Combined with real wages holding positive, the setup on the consumer side is more stable than the headlines keep suggesting.

[8] Jobs stayed stable, and the ADP weekly pulse ticked up three weeks in a row.

The ADP weekly pulse posted three consecutive weekly hiring increases. Not a rip-your-face-off number, but the direction is right and it lines up with a labor market that’s cooled without breaking. Takeaway here is more of the same for us… jobs stable ENOUGH.

[9] Initial claims are running 10.7% below the 8-year average.

Then we got initial claims for the latest week, which came in at 206K, which is 10.7% below the 8-year average at this point in the year. Continuing claims are behaving similarly.

If layoffs were the real risk, this is where you’d see it first, and it’s just not showing up.

[10] Low VIX has historically been a healthy regime, not complacency.

VIX is sitting around 14 to 15 right now.

The take is usually that this is complacency, but the 50-year data says otherwise. When VIX is in Q2 (13.3 to 16.1), the median 12-month forward S&P return is +11.6%, and it’s positive 89% of the time.

That’s the best hit rate of any VIX cohort. The complacency read is looking at Q4 (VIX 19 to 24), which is where hit rates actually break down.

[11] Every midterm path diverges.

Now here’s some interesting data.

Over the 8 midterm cycles in the base, the average final path from now through election day is +1.5%, the median is -0.6%, and the group is positive 4 out of 8 times. But the range is +12.1% (2010) to -5.7% (2022). The aggregate is a coin flip. Again this is measured between today and this year’s midterm election.

The two forces we’re watching most closely are what Iran does between now and November, and whether the current administration can use the situation to cut a deal. Both outcomes are binary. And when the setup is that binary, we don’t want to do too much.

Net Net

Ok so the through-line this week is that the pieces are lining up. Software and AI are not competing as much anymore for interest or trading activity, real wages are positive, jobs are stable, and the VIX regime is one of the strongest historical setups for forward returns. The counter is that midterm paths diverge and we’ve got a binary sitting on top of a foreign-policy variable. So the base case is constructive, and we stay measured in the parts of the book where the outcome is binary rather than probabilistic.

That’s all for this week!

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.

Sean Emory of Avory & Co. argues the SaaS apocalypse narrative is broken. CRPO accelerations at Salesforce, ServiceNow, Okta, Palantir, MongoDB, and Zoom, plus NVIDIA's data center guide, show software and AI are compounding together, not cannibalizing. Sean walks the four layers of the AI stack, why distribution wins at the application layer, and where the real risks sit.

Chapters
00:00 The SaaS is dead narrative
01:07 Intro and disclosures
01:50 Is the SaaS apocalypse wrong
03:52 Backlog signals: CRPO
04:16 Salesforce and Agentforce
05:50 ServiceNow and Okta
06:52 MongoDB and Zoom
08:31 Four layers of the AI stack
12:37 NVIDIA's full stack strategy
16:31 Risks and what to watch
18:11 Wrap up

More from Avory & Co.

Newsletter: investingwithdata.com
Website: avoryfunds.com

Disclaimer
For educational purposes only. Not investment advice. Opinions reflect our views as of the recording date and may change. Avory & Co. and Sean Emory may hold positions in securities discussed. Do your own research and consult a professional before making investment decisions.

Big Picture

💡 Burry, Buybacks, and AI Squeeze. So many signals.

Happy Sunday. I wanted to wait for Friday close to occur to ensure we did not get more info post market close Friday after US Treasury bought back bonds + Yen intervention…

A lot happened this past week and more is coming next week with Nvidia reporting. Everything seems to connect given AI and financing are aligned, and yields tie to both. We heard Bessent went back on the mic last week reinforcing what has become a very clear playbook. Stability in yields matters a lot, and sometimes more than the absolute level, both on the yen side and for the US long bond (30 year). Buyers of bonds just want to know prices won’t fluctuate all that much, and if you zoom out, yields have actually been fairly range-bound over the last four years. We’re still sitting in that range.

The reason stability matters so much right now is what’s happening on the spending ledger. Net interest continues to move up the ladder of federal spending categories, which means steady to lower yields is critical to keeping the fiscal math in a good spot. But you have two forces pulling opposite ways. Capital gains tax cuts on the revenue side, a war occupying and increasing outlays on the spending side. Both are pressuring the balance, while the interest cost line itself is also pressuring yields.

On AI, the data this week was interesting. Adoption is growing, something we believe in and something we have clear evidence on. And we are seeing IT budgets grow not shrink or simply AI replacing spend, which is key to the software working story. Also open weights are now inside the frontier pack, Microsoft entered the model board with its own silicon, and the memory names are lining up buybacks instead of reinvesting, which makes me hmmmm.

But before all of that, one visual. Here you go.

[1] Burry brings up Etched. We already brought this up…

Michael Burry surfaced Etched this week, and the reaction has been a lot of “wait, what is this.” We’ve been talking about Etched for months.

The point isn’t that Burry is right or wrong…

The point is that disaggregating the current compute stack, breaking apart the layers that are backed up and expensive, is a real signal, and it’s closer than the market probably thinks. We learned during COVID that barriers break. What usually took a decade took months. I know it’s a soft topic for some, but I’m referring back to vaccines, which the world condensed the production time frame on as supply and demand were out of whack.

For AI, historically it takes about three-six years to build a fab that can compete. Given the money flowing into this space and the demand and supply imbalance, that timeline is likely shrinking. Something we’ve been explicit about.

Key to put this in the back of mind as it can impact margins everywhere.

[2] Domestic private is now the biggest holder of Treasuries.

Bessent hhas called himself the “primary caretaker of the Treasury market,” and his recent actions are starting to make that framing pretty clear.

Whether it’s intervention around the yen or, more recently, attention on the long end of the U.S. curve, the message has been fairly consistent: policymakers seem much more concerned with disorderly moves and volatility than with defending any specific level of yields. Bond buyers can live with higher yields. What becomes more problematic is when the market starts moving violently and price discovery feels unstable. And despite all the noise over the last several years, if you zoom out, Treasury yields have actually remained within a relatively defined range. We’re still sitting inside that range today.

What has changed much more dramatically is who owns the market.

Domestic private investors now hold roughly 54% of the U.S. Treasury market, up from around 30% for much of the last decade. Meanwhile, foreign official ownership has fallen to roughly 14%.

That is a very different buyer base than the Treasury market was accustomed to historically.

It also helps explain why Bessent appears increasingly focused on maintaining liquidity, orderly markets and confidence in the long end.

[3] Net interest is now the second largest federal spending line.

This is why stability and eventual direction of yields matter so much.

Net interest ran at $827B through June of FY26 and is now the #2 spending category, ahead of Medicare and Defense, sitting only behind Social Security.

Steady to lower yields is very critical to keeping the fiscal math in a good spot. But you have two forces at play. The revenue side is hinting at capital gains tax cuts, which the administration has floated over the last several weeks. And on the spending side, a war is occupying and increasing outlays.

Two sides of the ledger going in opposite directions, which we think is impacting yields as well. Tough job for Bessent with that in store, so proactively showing support I think is important.

[4] AI is already a real line item in the IT budget.

Now back to AI.

A study showed that 34% of organizations already spend 10% or more of their IT budget on AI.

The mode is 5-10%, at 32% of respondents.

Only 6% are still under 1%.

So we are making progress on real use cases here and it is showing in budgets.

[5] AI eats crappy software, but IT budgets are growing.

Now then the question that matters for investors is where is that AI spend coming from.

66% say they’re reallocating from existing software.

54% say they’re increasing the overall IT budget.

Both of those things are true at the same time. AI is eating the simple software categories, the ones that lack defensibility, regulatory burden, or governance burden. Some of those tools are being replicated and live-coded for single-purpose solutions.

Others are being absorbed by vendors that are expanding horizontally. But the biggest takeaway is that overall IT budgets are growing, not shrinking, because of AI.

We think entrenched software with APIs, MCPs, and headless architecture, like what Salesforce announced, benefits from this. Additional IT dollars will continue to flow to current software vendors that can absorb the new workload. We have been saying this for a while now.

[6] AI spend is concentrated at the top, and AI-heavy companies are moving faster.

Ramp data is pretty good here.

The median company using Ramp is now spending $11.95 per employee per month on AI, up from $2.32 a year ago. That’s a 5x increase.

The top 10% are at $650, and…

the top 1% are at $7,400 per employee per month.

That’s a massive gap between who’s using it at scale and who’s not. Some of that has to do with coding versus non-coding companies. But there’s a second story here. AI-heavy companies are also moving faster.

Block is a great example. AI-pilled in a good way, as we haven’t seen product velocity like this from them since their early days. It’s showing up in the data and it’s showing up in their roadmap.

This ties to the last data point on the survey of IT spend.

[7] Anthropic has held its lead over OpenAI.

We also got an update on spending by model provider.

Anthropic passed OpenAI on Ramp’s AI Index in May 2026, and now sitting at 43.5% versus OpenAI’s 39.7%. That’s a ~4 point lead, and from what we know, it’s held over the last several months.

The bigger point is what’s happening below.

[8] Microsoft took the crown on image editing.

Microsoft has entered the frontier.

MAI-Image-2.5-Pro sits at #1 on the Artificial Analysis Image Arena with an Elo of 1272, ahead of Reve 2.1, GPT Image 2, Qwen, Nano Banana, and Seedream.

This was self-built on their MAIA silicon (their own chip). That is key because Microsoft is now competing at the top of the model board with hardware they own end-to-end.

[9] Open weights are now inside the frontier pack.

So that last datapoint was closed source and built by the cloud vendor.

Now Kimi K3 from Moonshot scored 60 on the Artificial Analysis Intelligence Index. That puts it inside the top tier, behind only Claude Opus 5 and Claude Fable 5 and comparable to GPT-5.6 Sol.

This is the other side coming at AI labs.

This is the first open-weights model to get here. This is critical. If the compute and open layer keeps pushing up the leaderboards, that further squeezes the model labs.

Margin compression from above and below is now a real force in the model layer.

[10] OpenRouter usage keeps expanding.

Now this data also shows that demand in general remains. Not survey data, but token volume. The overall routed model volume continues to climb, and the mix keeps rotating. What’s interesting is not any single model win but the shape of the market.

Applications are increasingly agnostic about which model sits behind them and are optimizing for cost, latency, and task fit. That’s another data point suggesting the model labs face pricing pressure from every side, not just from the top.

[11] Nvidia reports next week. History says the bogey is 3-5%.

Nvidia has beat revenue estimates in every quarter of the last several years.

The last four quarters averaged a 5.2% beat, with a range of 3.5% to 8.5%. Historical average across 22 quarters is closer to 10%.

The market bogey is somewhere in the 3-5% range now, and the second question is how much they beat their own guide by. Guide magnitude matters as much as the print itself for Nvidia who’s stock really has not done much post 2024.

[12] Samsung is planning up to $79B in shareholder returns.

Samsung reportedly plans to announce a shareholder return package worth as much as 110 trillion won, or roughly $79B. That’s a massive number.

Business is running well. Cash generation is following. And the answer isn’t more capacity or more M&A. It’s returning cash to shareholders.

Makes me wonder. If the AI investment opportunity was so big and the returns would last for a long time, you’d think these companies would reinvest back in the business rather than tee up capital returns after their stocks have soared.

That combined with Burry surfacing Etched, and with what we know about the disaggregation of the compute stack, is worth pausing and thinking about.

[13] SK Hynix is buying back $29B. Hmmm.

Same story as Samsung. SK Hynix announced a $29B buyback after this run.

If AI is truly the greatest infrastructure boom of our time, and memory is sold out through 2026, again why aren’t they investing aggressively into more capacity or acquiring companies within the stack to strengthen their position?

Net Net

Busy week overall as we saw influx of important info. Some signal and some facts. But lead with Burry showing Etched, but honestly we were already first to that earlier this year. At the same time, yields increasingly need to stay contained because net interest is now the second-largest federal spending line. That puts real pressure on policymakers to avoid another sustained move higher in the long end. Meanwhile, AI is clearly real. IT budgets are growing, not shrinking, and companies leaning into AI, like Block, are moving faster both in product and in how they compete.

But the picture underneath that is getting more complicated. Adoption is still top-heavy. The model layer is getting squeezed by open-weight models from below. Microsoft is pushing further into the stack with its own frontier models and silicon. And after a massive run, some of the memory names are talking about buybacks instead of pouring every available dollar back into capacity So it it interesting. Next week we get Nvidia, I will have a podcast session on this so STAY TUNED!

That’s all for this week!

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.

Big Picture

💡 Inflation cooled, interesting AI signals.

Are you tired of talking about inflation concerns? I am, but we continue. This week both inflation prints came in soft. CPI which is what consumers see came in at 3.4% Y/Y, PPI which is a input to future CPI came in at 4.7% Y/Y, and both landed below expectations for the second month in a row. We think this pushes the next rate hike further out and gives Warsh room at Jackson Hole to sound patient. One more inflation report lands before the Fed meets in September so not totally in the clear, but good news.

On the AI side, the story we’ve been running with keeps showing up in different ways. First news broke that Anthropic is reportedly getting bid at $2T. That is interesting on standalone basis, but then add in some chaos signals coming from OpenAI, the AI debate is quite interesting.

To top if off labor is entering its seasonal slow period, another good thing for Fed doves, energy is calm (for now), and Meta got a call out.

This week is a little different. Showing not only do we use data, but you have to add a layer of qualitative info to help build your investing mosaic.

Here we go…

[1] Inflation cooled on both sides of the print. Here is inflation section.

Let's kick it off with Consumer Price Index. Of course.

CPI came in at 3.4% Y/Y, down from 3.5%. Producer Price Index or PPI came in at 4.7%, down from 5.5%.

Both measures of inflation and both at or below expectations, and… both moderating for the second month in a row after peaking in the spring. Something we've been talking about since the tariff pass-through started showing up in the goods line a year ago. So underlying inflation is not a concern overall. The main reason we are talking about it is due to the war, which led to oil prices surging which drove headline prices. The Fed historically looks through one off bursts like energy, and Warsh sounds patient so far.

Above we annualize all the key measures of inflation using 1-3-6-12 month figures. The takeaway? All continue to migrate lower with some speed bumps along the way.

Lastly on inflation and I think this is a cool index, so for my stat data nerds, you will like this. Here we split the inflation line items into rate sensitive and non rate sensitive. The point? To figure out whether raising or lowering rates would impact underlying inflation. For example if you raise rates will that stop health insurance prices from going up… probably not. Will it impact credit card spending on consumer items? Yes. Right now rate rate INSENSITIVE (ie energy and insurance) are higher than rate sensitive. Meaning nothing to do from the Fed.

[2] Anthropic reportedly gets bid at $2T.

Ok now we get into some cooler items. The FT is reporting Anthropic is in early talks for a round that would value the company around $2T. WOW.

That’s a step change from the last mark and it puts the frontier model race squarely in the trillion-dollar bucket. Money keeps chasing the frontier, but the compute bill keeps growing right alongside it.

This has two implications.

1. Everyone was all in on OpenAi just 12 months ago. To me calling a winner in such a transformative moment is dangerous. So much capital is flowing here that the winner will likely keep changing, especially as the way to shift from one vendor to the other can literally take 10 minutes.

2. We have exposure to Anthropic. More on this later but if you follow our work, there are interesting ways to get exposure to arguably current most important company in the world today.

[3] Polymarket odds keep shifting on the frontier.

If prediction markets say so, then it must be true right? That is me being facetious, but Polymarket has been a decent real-time read on what the market thinks about who leads next in AI. The odds keep moving as new models drop. We watch this the same way we’d watch positioning data. Not a signal on its own, but useful context when it moves fast. It says Anthropics valuation keeps moving up.

[4] Zoom is getting a sum-of-parts read.

We’ve broke out Zoom into pieces. Core video, contact center, phone, and now the AI-native workflows. That is its core business. Then you have cash. Then you have… drum roll… Anthropic stake. At $2T for Anthropic stake, that squarely puts Zoom’s stake probably somewhere near $6-10B. Some other companies have stakes too, so not the only way to play it, but as a percent of market cap, Zoom is right at the top. Do your own research.

[5] Chaos at OpenAI? COO Brad Lightcap leaving…

Brad Lightcap COO at OpenAI is leaving. He has equity, he knows what is going on there, he knows running frontier is expensive, exhausting, and the returns from each new model are getting harder to earn.

Now some will say hey he is moving on and he made enough money already.

My view is simple: Doesn’t Sam Altman keep saying we are hitting AGI by 2030. Same thing said at Google, and said at Anthropic.

Well if that is true. Who in their right mind would leave the #1/#2 leader up to AGI?

My take: AGI is either not 2030… or OpenAI is not in position to lead. Both can be true. Or maybe yeah he wanted a break. Signal to me.

[6] Then we have the Chief Revenue Officer.

And then this. Whatever OpenAI is going through internally, it’s not a company acting like everything is smooth. The CRO who just took over 8 months ago is leaving also.

More signal. Take it how you want.

[7] Bessent on Meta Muse.

This one is interesting and a signal the other way. Treasury Secretary Bessent flagging Meta Muse this week is not a small thing. When policy voices start naming specific AI products, it’s a signal that the surface area of AI is getting bigger, not just in the market but in DC too. Meta is spending like they know this.

Again interesting since I do not think Bessent is one who would tweet/x things on a fly.

[8] Suhail on the switching cost.

Suhail, one of the youngest YC members ever, had an interesting take. His take on switching between frontier models is worth reading. The cost of moving between them is dropping toward zero (as we have been saying for a while), which is exactly why the value in AI is moving up the stack toward workflow and distribution. Again this is a thesis we’ve been running with for a while now.

[9] ADP weekly hiring pulse stays soft.

Now some jobs data. ADP’s weekly hiring read keeps trending below where we were a year ago. Not falling off a cliff, but clearly cooling. That fits with the inflation prints and gives the Fed room to sit still. We’re watching this weekly, not monthly, because that’s where the real turn will show up first.

[10] Avory at BARR team event.

A great night with the BARR team this past week. Real relationships still matter in this business. We’re grateful for the partnerships and the people showing up in person. We laid out our views of the state of the markets, our views on ai, all data driven. If you want to go through the piece with us, let us know!

Net Net

Inflation is cooling, labor is hitting seasonal slowdown, and the AI dollars keep chasing the frontier while the value seems to be moving up the stack. We stay long the names that turn AI into cash, and we stay patient on the Fed. One more inflation report before Jackson Hole. Chairman Warsh gets his shot to set the tone. Let’s see.

That’s all for this week!

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.

Big Picture

💡 The jobs market was out this am, tariff refunds are quietly funding the consumer, and our companies keep executing (but volatile).

The July jobs number just hit and the print was actually negative for the month, on the surface not great but private sector hiring came in at +30k, inline with month prior. As always for us, the way that data gets calculated matters a lot. It takes a single reference week and roughly 25K jobs got impacted this month due to weather among some other nuances. We tend to look at data sources like Revelio that take a combination of various inputs to calculate the labor picture and they’ve proven to be fairly accurate once the eventual revisions hit. Our read is that the labor market is stable overall, not boom or bust. Good news is it likely keeps Federal Reserve more balanced.

Also, tariff refunds continue to be an interesting driver for the economy this year. That boost combined with tax refunds has completely offset any impact that elevated oil prices would have made, which is something we’ve been discussing for some time. It’s showing up in the data.

On the company side, Meta’s superintelligence team came out with a new model and it is impressing folks. Meta is climbing up the intelligence ladders and getting closer and closer to closed-source frontier models. Not only are they catching up in terms of intelligence, but the cost per token is impressive, and ultimately given their size and scale they are in a position where they can put financial strain on competitors that don’t have the cash machine that Meta has to help fund development.

Then the earnings, our favorite time of year. We saw lots of validating information in the quarter. Category leaders keep executing. Block, Paycom, Atlassian, First Watch, Xometry, Zillow, Airbnb, Clear Secure, Duolingo. Price reactions differed, with some strong moves higher and some not so much, but fundamentals will drive return overtime. We remain long some of those operators (not all). Lets dig in.

[1] Seems like category and foundational leaders are winning across the board.

The through-line this quarter is that the operators keep executing. Nine names we tracked and same financial outcome. Software, restaurants, marketplaces, real estate, travel, identity, education, payments. All accelerating or holding the line while the broader market worries about AI, housing, and the consumer. This is what we mean when we say category leaders keep executing.

[2] July jobs printed negative, Revelio tells a different story.

The BLS print for July came in at -23K jobs and revisions took May and June down another 103K combined. This is the first negative payroll month outside of the pandemic. But the methodology here matters. BLS uses a single reference week that includes the 12th of the month, and roughly 25K jobs got dinged from weather and other nuances that show up on that single week. Private payrolls grew +30k which is a better figure for what’s taking place outside of government. We tend to look at Revelio’s data because they pull from a wider set of inputs, and once BLS revisions eventually roll in Revelio has proven to be fairly accurate. Their July read shows Manufacturing +20.7K on the strong end, Leisure -9.8K on the weak end. Market for jobs remains pretty stable overall.

[3] Tariff refunds keep offsetting oil.

$71B of the $166B in tariffs collected has been refunded to companies so far, and Apple alone is sitting on $2.2B of refunds. Combined with tax refunds, this has completely offset the impact of elevated oil prices this year. This is something we’ve been talking about for a while now, and it’s showing up in consumer resilience across the data we track. We think it is only starting and will likely help companies and consumers throughout the rest of 2026.

[4] Meta is back in the AI game.

Now onto Meta. Meta’s superintelligence team just released a new model and our take was straightforward. They are climbing up the intelligence ladders and closing in on closed-source frontier models. Muse Spark 1.2 sits at #6 on the Artificial Analysis intelligence index with a score of 57, shoulder-to-shoulder with GPT-5.6 Terra and just behind Claude Opus 5, Fable 5, and GPT-5.6 Sol. The more interesting read is on the right panel. Muse Spark 1.1 lands squarely in the most attractive quadrant, delivering an intelligence score of 54 for $1.43 per task while Claude Opus and Fable are running $8 to $12 for a similar or slightly higher score. Given Meta’s size and scale, they are in a position where they can put real financial strain on competitors that don’t have the cash generation to fund development at this pace. We’ve been explicit about our thesis on Meta and this is exactly why.

[5] Block accelerated across the board.

Block put up a really strong quarter accelerating in many parts of their business. They beat and raised their guidance for the full year. They’re seeing efficiency gains driven by AI adoption, most of which has been built organically. They launched two open source projects within the last 6 months or so, Goose and Buzz, both of which have been met with real excitement. These may not be monetization drivers by any stretch, but one way to look at it is the internal talent inside Block that’s building the systems and tools, which is driving both internal efficiency and product velocity as we continue to see newer products.

The investor community is focused on Neighborhoods, which is really their method of connecting their ecosystem. Something they’ve talked about for nearly a decade. The disclosures they provided gave us a little bit of positivity on the fact that Neighborhoods has reached product market fit, and they provided various data points to support that in terms of payment volume flowing through neighborhoods and followers of certain stores. They’re iterating their way through this.

Which is why we think Block is much healthier and stronger than maybe even bulls expect. Everything they’re doing is very vertical, from Square Financial and their charter bank, to their hardware team building multiple hardware products. That’s only really possible if you have the talent inside. They own the consumer relationship and the merchant relationship, they have self-onboarding, and they’re also leaning into field sales. This is becoming a much more mature company on the sales and product side, but in many ways they still operate like a startup through product velocity and cadence. That keeps us excited.

[6] Paycom spiked 25% on record profits.

Paycom hit a record 44.2% adjusted EBITDA margin in Q2 26, and the stock ripped nearly 25% on the print. The company is finding ways to accelerate growth again, and AI is giving them a real tailwind to launch new products and modules that are driving results. Automation from Beti and GONE keeps compressing operating cost per client, which is exactly what you want to see in a software business at scale.

[7] Atlassian is proving the AI-impairment thesis wrong.

We called Atlassian out last time as one of the bullseyes for the “AI is going to kill this business” narrative, at least that’s what the street was implying by the drastic move in the stock over the course of the last year. Now they’re two of four quarters in a row proving they’re finding ways to re-accelerate growth.

RPO growth hit +44% Y/Y in Q4 FY26. We have our own views on stock-based compensation and some other factors inherent in the business, but from a pure revenue standpoint they seem to be doing just fine four and a half years after ChatGPT launched. There’s certain platforms out there that stand to benefit in a world of AI, and Atlassian is proving that now.

[8] First Watch keeps outpacing the industry.

First Watch continues to execute well in a market where investors remain on edge about the consumer. Q2 26 same-restaurant sales came in at +3.4% and the two-year stack is +6.9%, both outpacing their own guidance and outperforming peers again. This quarter should help alleviate any concerns about how they’re going to fund their future growth. On the call they echoed that they plan to fund future growth from organic cash generation and committed to opening around 50 stores a year. That’s the kind of self-funded discipline we look for and part of the original thesis.

[9] Xometry marketplace accelerated for the fourth quarter in a row.

Xometry’s marketplace revenue grew +45% Y/Y in Q2 26, extending the acceleration to four consecutive quarters at +26%, +33%, +40%, +45%. Guidance for the full year got raised to +33-34%. This is a name that has been quietly building a network effect between buyers and sellers of custom parts, and every quarter the flywheel gets more obvious. Reshoring, AI-driven design, and small-batch manufacturing are all tailwinds that fall right into their model. Valuations are getting up there though, as a source of caution.

[10] Zillow put up +18% revenue in a flat housing market.

Zillow continues to execute well in a housing market that’s essentially flat to negative, something we’ve also been discussing for quite some time.

Existing home sales TTM are running at just +6% while Zillow put up +18% revenue growth along with margin expansion.

Despite a housing market that is virtually seeing little to no activity, they’re growing three times faster than the underlying market. They have a couple of changes they’re making to drive the combination of rentals, for sale residential, and mortgages, and they are now a top 25 mortgage lender in the country. The model keeps working.

[11] Airbnb GBV kept accelerating.

Airbnb also continues to execute well. Bookings are growing and accelerating across the board, with Q2 26 GBV up +16% Y/Y and $27.2B in bookings for the quarter. Their newer categories like experiences and services are attaching to their existing base.

Brian Chesky articulated it well across the different phases they’re executing here. People come there for staying at homes, then they lean into experiences, and now hotels as a category is driving additional traction. Ultimately this is heading toward a platform that has AI infused in it, both inside the product and as the means of putting together your vacation travels in an AI-first way. If there’s anyone that can do it, Brian seems like one that can execute here, and the fact that their inventory is inherently much different than anybody else makes it interesting.

[12] Clear Secure hit a record margin on accelerating bookings.

Clear Secure reported $295.9M in Q2 26 bookings, up +33% Y/Y, and adjusted EBITDA margin hit a record 36.4%. That’s the first time the margin has run above the 35% target they laid out at IPO.

Bookings and margins accelerating together is the ideal combo, and it’s happening at a company that owns identity infrastructure across airports, and now moving that identity platform into other parts of business like healthcare and financial services.

[13] Duolingo DAU growth reaccelerated.

Duolingo is a name we’ve done well in in the past but sold on valuation exuberance, and since then it has gotten hit pretty hard. We still watch it closely. There are fears around AI and what it could do to the education and learning side, plus translation both across the internet and voice. I think those fears are generally overblown. Duolingo continues to prove recently that they can drive daily user growth and re-accelerate that number, which is key to the durability story.

Q2 26 DAUs grew +23% Y/Y to 58.7M, up from +21% in Q1 26.

Net Net

Big week with lots of earnings. Tariffs plus tax refunds are acting as a tailwind, and that likely continues throughout 2026. Jobs are stable enough, but not too strong, which still gives the Fed some pause on hikes. Meta keeps climbing the intelligence ladder in a way that could change the competitive shape of the entire AI stack. Time will tell what the strategy is with some of these models and how it all comes together. And the operators we own or watch kept doing what they do. Block, Paycom, Atlassian, First Watch, Xometry, Zillow, Airbnb, Clear Secure, and Duolingo all executed. We have some portfolio changes coming over the next week or so. More on that soon!

That’s all for this week!

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.

Big Picture

💡 The AI names shook out, the Fed stood pat, hyperscalers delivered, Buffett was right.

This was a very interesting week. From AI funds blowing up, to earnings, to the Fed.

The gyrations across AI names finally started to show their hand. It was tough watching random stocks rise 100% in a month simply because of AI exuberance, but being disciplined means not chasing short-term returns and instead asking what is sustainable.

If 1999 is any guide, most of the perceived AI winners will ultimately be losers, while the smaller group of real winners becomes clearer over time.

We are starting to see that with the hyperscalers. Amazon and Meta (which we own), along with Microsoft and Google, are already benefiting from AI. They are ALL designing their own chips, building their own data centers, using the compute internally and selling excess capacity to others at a premium. Not bad.

Meanwhile, many AI names fell 30% to 50% in two weeks, and a prominent hedge fund reportedly blew up, forcing Citadel to step in and take over the portfolio. It was 4x levered, what did you expect. Those are not normal market conditions. It reinforces the point we have made for a while: leverage and speculation were helping drive many AI-related stocks higher.

Again to be clear, we believe deeply in AI and have investments tied to what it is today and what it can become. But this was another reminder of what happens when speculation gets too far ahead of fundamentals.

Couple other aspects this week. The Fed also held rates steady. There had been some discussion of a surprise hike, but we did not think one was warranted or that surprising the market would have been productive. Holding rates was the right decision.

Then we got hyperscaler earnings, which were constructive across the board. Advertising remained strong, Amazon’s retail business grew at one of its fastest rates since COVID, and the consumer continues to look healthy. Enterprise demand through Microsoft and Apple was also solid, with little evidence that AI is eating away at core software spending.

One point we made on X received more than 300,000 views. Hyperscalers increased capex, but nowhere near the increase we are seeing in memory prices. Makes you think.

Memory prices are up 2x to 3x. If those increases were fully flowing through to the cost of the buildout, capex growth should be closer to 20% to 40%. Instead, it got moved up by roughly flat to up 8%. May be time and place, or it may be signal. Either unit demand is slowing at the margin, hyperscalers are finding ways around the cost increase, or someone else in the supply chain is absorbing it. Given hyperscaler margins are still rising, you can draw your own conclusion.

Lastly, the labor market remains stable. Initial claims are holding up, while the alternative data we track and apptopia data continues to show improvement in both blue-collar and white-collar hiring.

Let’s get into the data!

[1] This was a wild week. A $45B AI fund got washed out.

Ok lets start here. Yikes…

Situational Awareness, the AI fund founded by former OpenAI researcher Leopold Aschenbrenner, was 4x levered to the AI trade. Doing so with tens of billions of dollars was part of what drove up AI names in the first place.

That leverage and speculation is exactly what led to the fund being washed out this week and taken out at a discount by Citadel.

This is what we mean when we talk about speculation running ahead of the fundamentals, and it’s the setup for everything else in this weeks issue.

Remember this quote we shared just 2 weeks ago… Good ole Warren dropping truth bombs

[2] Hyperscaler capex is nowhere near the +20-40% memory pass-through implies

Now after getting the earnings this week and knowing price of memory we wanted to make some hypothetical assumptions…

Here you can see the disconnect that’s shaping our read of the entire quarter.

Memory is 20 to 40% of the cost of an AI buildout, and prices are up 2 to 3x. If hyperscalers were passing that fully through and holding unit growth flat, capex should be up +20 to +40%. Instead Alphabet is +8%, Amazon +10%, Meta +2%, and Microsoft actually cut guidance -8% (accounting change).

Either the impulse in unit demand is fading, or the hyperscalers are eating a lot of the memory cost themselves. Something we’re going to be watching closely but their margins are not getting worse while memory names have 80-90% margins.

[3] DRAM prices are ripping, with 2026 up +245% Y/Y in the forecast.

Here is the evidence.

The Y/Y price change in DRAM is the memory story in one chart.

After -19% in 2022 and -41% in 2023, prices flipped hard in 2024, and the current outlook has 2026 up +245% Y/Y with another +36% in 2027 before rolling over. The 2027 is an estimate from Nomura, but still.

[4] AWS growth is holding, another sign the enterprise is in good shape.

Now to earnings.

AWS is clearly winning. It delivered another strong quarter, and the read-through for enterprise demand remains constructive.

Growth is very strong for a business of this size, and margins remain healthy.

AWS is also winning in AI for a simple reason: AI is difficult to implement. At the compute layer, companies need access to infrastructure powerful enough to run these workloads. AWS and the other hyperscalers take much of that complexity off the customer’s plate.

Then you have the model layer. Amazon Bedrock allows companies to run AI workloads on AWS, choose between Amazon’s own chips or hardware from Nvidia, AMD and others, and use the model that best fits their needs.

This allows businesses to focus on what they do best rather than managing compute infrastructure, model deployment and delivery.

We call this orchestration, and Amazon is winning.

[5] Meta’s ad engine keeps compounding.

Meta’s earnings showed the advertising business continues to run very well.

Revenue grew 28%, and the company guided to roughly 22% growth in Q3, despite facing its most difficult Y/Y comparison.

The read through is that the consumer remains healthy enough for advertisers to keep spending.

Meta also saw both impressions and price grow at double-digit rates. To us, that suggests the AI story is working for them. Engagement is improving, while advertisers are generating strong enough returns to support higher pricing. Sounds good to me.

Is Meta actually using AI? Yes. The company said Instagram Reels recommendations are now entirely powered by large language models. More than 50% of recommended feed content is also less than one day old, more than double the level from last year.

AI is not a side project for Meta. It is increasingly what runs the core platform.

There is still plenty of noise around Meta’s AI spending, but the actual capex numbers were more measured than what I keep reading. Meta seems to always be punching bag despite strong growth and profits.

Here’s some math… Capex came in $3 billion below expectations this quarter and was guided $1.5 billion above expectations next quarter. On a two-quarter basis, that is still $1.5 billion below what the Street had expected.

Meta also increased its full-year capex guide by only 1.9%, one of its smallest increases.

[6] Tariff refund payouts are running well above the prior run-rate.

Now to some other news.

Economy is getting a boost.

The refund bucket keeps expanding, and the pace is a good marker for what’s actually flowing back into corporate P&Ls.

Tariff refunds are happening. We heard it from Starbucks and many others this week in their earnings calls. We tracked BofA data to small businesses, and the story is same. Money back to economy.

[7] White-collar hiring is turning back up, now +5% Y/Y.

Now to jobs.

The Bullhorn Professional Index has been in a two year drawdown since the 2022 peak, and it’s now running +5% Y/Y for the four-week trailing.

This is the cohort that everyone has been worried about on the AI-displacement narrative, and the data is so counter to the narrative. Hiring is coming back you just have to look at data not feelings.

[8] Blue-collar staffing is accelerating, +10% Y/Y.

Another datapoint.

The Commercial Index is running +10% Y/Y and accelerating hard off the 2023-2024 lows.

Combined with white-collar turning up, this is a very clean signal that the labor market impulse is picking up on both sides. Good read for the consumer and for the broader economy.

[9] The consumer is shopping smart across four different app categories. Apptopia Data.

This is a cool dashboard.

We like to bring in Apptopia download and engagement data to check consumer intent, and this week the reads are mostly constructive.

Deal apps accelerating to +35% Y/Y , most of this coming from one app DealSeek. So consumers looking for deals, nothing new but something to watch.

Job apps back positive in July after two negative months.

Gig Worker apps off the highs but slightly re-accelerating.

And Gas apps decelerating hard as the fear of a gas spike comes out of the system.

Most moves are in constructive directions but cool way for us to measure real-time consumer health… (IF YOU LIKE THIS DASHBOARD, LET US KNOW IN COMMENTS, WILL ADD MORE OFTEN)

[10] Kimi had a viral moment in July, briefly surpassing DeepSeek.

Now you may or may not have heard of Kimi. So we looked at the data.

Kimi went from ~37K daily downloads to a peak of 218K in about two weeks. Impressive trial spike, and it briefly took over DeepSeek. But context matters. Kimi is still well behind Grok, and both are miles behind where ChatGPT and Claude sit on the consumer side.

Our read is that the consumer AI market remains ChatGPT and Claude territory, and the interesting battle for Kimi and the challengers is on the enterprise side, not consumer.

[11] Snap’s fastest-growing cohort is Age 46+, nine straight positive quarters.

We were looking at data this week and we came across this. Asking ourselves whether Snap is aging up, and the data says yes. Time spent per DAU for the 46+ cohort has been positive Y/Y for nine consecutive quarters, with Q2’26 running +17.9%. The 4Q average sits at +9.2%. Snap’s platform is broadening its user base into the older cohort, and that’s a durable shift, not a one-quarter print. Lets see how it shows in Snap earnings coming up…

[12] Venezuelan crude is flooding back in.

The 2026 annualized run rate for US imports of Venezuelan crude is 111M barrels, up from 51M in 2025 and 48M in 2023.

It’s another data point that adds supply into the global energy picture, and it’s part of why energy has stayed calmer than the war headlines would suggest. We continue to see evidence that neither side truly wants to escalate. If they did, we’d have already gone further by now.

Net Net

The AI shakeout was overdue, and the leverage that got flushed this week was a healthy reset. The Fed did the right thing by staying put. Hyperscalers delivered, and their earnings tell you the consumer and the enterprise are both in good shape. But the one thing we keep coming back to is the move in memory prices. Memory prices are up 2 to 3x, and capex is up single digits. Something has to give, and forsure that will be the read we’re watching into next quarter.

That’s all for this week!

MAKE SURE TO SUBSCRIBE TO AVORY AROUND THE DESK PODCAST

Spotify: Here

Apple: Here

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.

Chapters

00:00 Meta Earnings Debrief
01:14 Show Intro & Disclosures
01:55 Revenue & Guide Highlights
02:55 AI Monetization Flywheel
06:49 Open Versus Closed Models
08:46 Enterprise AI On WhatsApp
11:01 CapEx Outlook And Bottlenecks
13:04 Wrap Up & Key Takeaways

This discussion is for educational purposes only and should not be considered personalized investment, legal, or tax advice. Views reflect our opinions as of the recording date and may change.

Big Picture

💡 Earnings reviews for Tesla and Google. Earnings previews for next week. And software is starting to fragment winners from losers.

Happy Friday! This week we got earnings from the likes of Google and Tesla, along with ServiceNow. You’ll hear more about one of these names in a week or so. But overall the conclusion was that earnings in software are proving winners and losers in the AI era. Also Google suggested that the economy is healthy while AI spend continues. Tesla was a different story as the stock is now down -14% over the last ~5 years.

At the same time, the macro backdrop is back in flux, not making this easy. Oil ripped ~30% off the recent low. The 10Y sits at the top of a four-year range. But then jobless claims fell to 187K, a new cycle low, and Jensen keeps saying AI creates jobs. So the “AI is job destroying” narrative isn’t quite lining up with the tape yet. That’s the Avory view, and we’ve been leaning into it. Question remaining is when do we actually stop this war so that the rotation we saw can continue.

Meta reports next week and Tesla already gave us the reminder that scale without a working narrative doesn’t buy you much. So let’s see what they share. We breakdown some previews using data, on Roblox, Lemonade, and Duolingo.

Let’s get into it…

Before That: If not already sign up to our podcast Avory Around the Desk: Spotify, Apple, YouTube.

[1] Google delivered. Cloud is the story now.

Google’s quarter checked every box we cared about. Search revenue grew +17% Y/Y, YouTube ads up +13%, and Cloud came in at +82% with an operating margin over 20%. Cloud is the story here. It’s now big enough and profitable enough to matter, and it’s growing faster than any other segment. Impressive. Someone asked is it cheap now? Our view if you assume they get back to $100B in free cash flow by 2030, they are then trading at 37x that out year number. Not exactly cheap. But great asset, one we have owned before, but valuation matters to us.

But we have been in the Google camp for a while now. Regular readers won’t be surprised by this. The AI overhang was overdone, and the earnings continue to settle it.

[2] AI Mode isn’t killing Search. It seems to be boosting it.

More on Google.

Search revenue grew +17% Y/Y in Q2, the fifth straight quarter of teens growth. That doesn’t happen if AI were cannibalizing the core.

If anything, more searches, more queries, more monetization surface. The bear case on Google Search has been that AI would eat it. The data says the opposite so far and don’t forget this is year 4 of AI.

That’s the read from our desk. Off the Q1 high but still very healthy. The story here is durability and that’s good for Google but good for economic views also.

[3] ServiceNow: scaled software thesis in one chart.

Subscription revenue +21% Y/Y.

Free cash flow margin north of 30%.

Current RPO growing +25%.

This is what a scaled software business looks like in an AI world. AI while easy to use, isn’t easy to implement especially at scale. Soooo yes you lean on trusted, innovative, and full spectrum platforms.

NOW is not a story stock anymore. It’s a cash machine that happens to be growing 20%+ and is embedding AI into every workflow it touches.

We’ve been beating this drum for a while, not just for them but broadly in certain software names.

read somewhere this week “what if after all this software is the biggest beneficiary of AI”

my views is yes. But who…

[4] Duolingo: AI power users churn far less.

So we are heading into some earnings this week.

Duolingo being one. Here’s a name we owned in the past. A successful investment, and got out on valuation. Now the name has fallen from graces on AI fears plus some social media issues.

But Apptopia data we track is helping inform us of some interesting trends.

The chart tells the story. Users who engage with AI-powered features churn at roughly half the rate of the rest of the base. Read that again.

That’s a big deal for a subscription business. It means AI isn’t just a feature, it’s a retention mechanism.

This is why we care about which companies can actually operationalize AI inside their product. It’s more of a loop.

If we can see users pick up, and their AI-native products increase retention and avg revenue per sub, then the story gets going again…

[5] Lemonade downloads are reaccelerating.

Now onto Lemonade. A name we have covered in the past, spoke with management several times. General theme here is that insurance will continue to move closer and closer to the end users. Listen to our Prediction Markets Meets Insurance Podcast Episode.

So back to Lemonade.

Monthly downloads using Apptopia data for Lemonade are running well above last year. Y/Y growth turned back positive after a rough stretch, and the reaccelerating is clean. Small sample still, but worth flagging. Insurance is one of those categories where AI-native players get to compound for years before it shows up in the headlines really.

This shows top of funnel for Lemonade is working again. Heading into the quarter, this feels constructive.

[6] Roblox: the 17 to 25 cohort is going deeper.

Now onto Roblox.

They have been treated like a kid’s game for years. Every earnings we want to know when and will they move up market.

The data keeps saying yes here.

Weekly active penetration in the 17 to 25 age group is climbing steadily, and it’s now well past where the 9 to 12 cohort was a few years back. The platform is aging up, not aging out. This should support future ads monetization and also assist and create larger market for more developed game formats.

This is one we own, yes, newer position this year.

Nothing new from us here, we’ve said it before, but the data keeps supporting the thesis of aging up, cleaning up the legacy issues around child safety, new game formats, AI allowing for more creation not less, ads monetization, and new platforms.

This quarter may not be an inflection point as they work through their verification process which impacts engagement, but that will ultimately prove to allow a safer more monetizable platform to exist.

[7] Tesla holders are down 14% over five years.

Crazy but if you bought TSLA 5 years ago, you are down 14%. Also every EV maker is down 95% or more. So reminder, cool ideas does not equal nice returns. Stay disciplined when hot stocks or hot ideas pop up.

They reported this week and the company continues to look more like traditional automotive than anything else. With gross margins around 16%, street was at 18%, this peaked at high 20’s%.

The multiple is still well above traditional automotive tho so the robotaxi story has to actually show up in the numbers eventually, or holders keep grinding. We shared last week how Amazons Zoox is not in the autonomous game and taking download share. Waymo we know exists and doing well, plus some others are coming to market soon. Optimus is cool idea, will come to market at some point. We will probably see SpaceX combo with Tesla here soon enough.

Actually, and I just wrote about this. There’s a weird incentive for Tesla shares to go lower so that SpaceX can merge or acquire Tesla, which will then provide them with in-house AI and compute to help assist in the robotics side, along with their energy business. Musk wins either way tho…

Time will tell.

[8] Meta preview: capex, AICAAS, and user growth all matter.

Meta reports 7/30. Four things we’re watching:

Capex. The Street is at ~$33B for Q2, more than double last year. That’s a huge step-up and the market wants to see the payoff cadence.

Daily Actives growth. We want to see stabilization or reacceleration, not further deceleration.

ARPP. Consensus is $16.70, which would be +22% Y/Y. This is where AI-driven ad monetization has to show up. Another move higher says AI working.

Revenue and capex Y/Y. The gap between top-line growth and capex growth is interesting. Consensus has revenue at +27% and capex at +100%. That’s fine for one quarter. It’s not fine forever. But in 2 quarters from now the Q/Q increase in CAPEX will be surpassed by the Q/Q increase in revenue, an important cross over.

This one has our attention.

[9] Oil ripped ~30%, and the market shrugged.

WTI bottomed near $54 in December and touched $119 in March on the Iran spike. Today we’re back around $90. That’s a ~30% move off the low, and equities moved lower especially in macro and consumer areas. But not as bad as back in Feb / March. That’s welcomed news.

We’re watching this one, but not repositioning around it as we think all this subsides at some-point here as incentives are for both to have oil flow. One wants oil to flow for profits, other wants it for consumers and politics.

Let’s see but we saw what happens when yields and oil and inflation expectations fall, consumer and macro names rip.

[10] The 10Y is sitting at the top of a four-year range.

Continued from last chart. The 10Y is at ~4.8%, right at the top of the range it’s been carving for four years now. Not a breakout, not a breakdown. Just sitting up there. This is the level where the equity market gets nervous historically, but so far the tape is holding.

Bottom line: rates are elevated, but they’ve been elevated. Nothing new here. Same story, different week. All tied to oil and inflation concerns. Plus some deficit talks due to war $.

watching closely.

[11] Claims fell to 187K. Jensen says AI creates jobs.

All of the prior was market moves. But jobs, ie macro, are looking good. Something we called for heading into 2026, was a pick up in jobs activity. Its showing.

Initial jobless claims fell to 187K, a new cycle low. Not what you’d expect if AI were destroying jobs at scale. Jensen said it again this week: AI is creating jobs, not killing them, and the data isn’t disagreeing with him yet. He said “ai will add millions of jobs”. All these AI-native companies, do we think humans don’t work there?

We’ve been in that camp for a while. The doomer narrative on labor doesn’t square with the facts on the ground, again we are in year 4 of AI. Something to keep watching, but for now the ground truth is what it is.

Net Net

Scaled software is starting to separate itself, and this week made it obvious. Google and ServiceNow showed what happens when AI shows up in the numbers. Meta gets its turn next week and Tesla already showed the counterexample. On the macro side, oil is higher and rates are elevated, but claims at a cycle low keep the labor market off the table as a bear case for now.

That’s all for this week!

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.

He frames three eras of software: individual tools, integrated platforms, and an emerging era where AI orchestration acts like a new operating system that selects the best capability for a task regardless of UI or vendor.

In this model, winning depends less on distribution, brand, and long contracts, and more on reliability, latency, accuracy, security and governance, API quality, cost per call, and task success rates.

Sean argues platforms won't disappear. They may become command centers for permissions, compliance, and governance, while features "escape" suites. He walks through examples from Shopify, Shop Pay, Stripe, Twilio, Zoom, Box, Salesforce, Adobe, and HubSpot.

Chapters:
00:00 AI changes software buying
01:44 Three eras of software
03:10 Composable software era
03:55 Agents versus humans
05:06 Orchestrator as new OS
07:46 Capabilities escaping platforms
08:33 Examples: Shopify, Stripe, Zoom
12:28 New scorecard for winners
15:32 Two questions for companies
17:22 Where this could be wrong
19:48 Big picture takeaways
23:22 Closing and subscribe

Listen on:
Apple Podcasts: https://podcasts.apple.com/us/podcast/avory-markets-and-investing/id1504555573
Spotify: https://open.spotify.com/show/3A8acTyfhhxpUFoFLEKeMJ
YouTube: https://youtube.com/@avoryco

Follow Sean and Avory & Co.:
X: https://x.com/avoryco
X: https://x.com/_SeanDavid
LinkedIn: https://www.linkedin.com/company/avory-co
Website: https://avoryfunds.com

Disclaimer

The content on this channel is for informational and educational purposes only. It does not constitute personal investment advice, a solicitation, or an offer to buy or sell any security. Views expressed are those of Sean Emory and Avory & Co. as of the recording date and are subject to change without notice.

Avory & Co. and Sean Emory may hold positions in the securities and companies discussed. Any references to specific companies, products, or services are for illustration only and should not be interpreted as recommendations.

Investing involves risk, including possible loss of principal. Past performance does not guarantee future results. Consult a qualified financial advisor before making any investment decision.

Big Picture

💡 Earnings season kicked off this week and the market is doing exactly what we’ve been calling for.

Happy Friday! On a plane here so wrapping up some of the data. This week was packed with info. My view remains pretty simple. We’re constructive on what’s been taking place. The rotation we’ve been talking about is happening, driven by a healthier macro views and people becoming less concerned about the Middle East. Oil isn’t responding as aggressively as it was. That can change, but it doesn’t change our view that the economy can withstand it either way.

Also if you look around, there have been a large swath of companies treading water for 6, 12, 18, even 24 months while continuing to grow revenues and bottom lines. That is opportunity. The banks reported this week and delivered clean prints. Always look at banks to give you a sense of economic activity. For the banks, net interest income and margins were positive across the board, deposit and lending growth were constructive, and the messaging on the economy and consumer was uniformly healthy. That is good and coincides with our new bank sentiment reading. See below.

A few other things caught our eye this week. The autonomous space is getting interesting. Amazon’s Zoox is gaining, literally have not heard anyone put Zoox in the same category as Waymo or Tesla, but I think they should start, see the data, we are. Venmo as an asset remains interesting, and ironically PayPal was reportedly bid for by Stripe, Block and Advent. That is a big deal and I suspect the offer goes higher after speaking around, but something to watch as it relates to competition for other fintechs. Credit card data below shows the consumer is fine, and the Venmo engagement data speaks to why the asset remains attractive.

Lots to share, let’s get into it.

[1] Netflix engagement data down sequentially. Data we tracked flagged this ahead of time.

Netflix reported last night and first it is great to have earnings season back. Love it.

What we saw was generally healthy top line and bottom line. However, viewing hours per member fell sequentially for the first time in five quarters. On the earnings call Netflix’s answer was a constructive counter, not every hour is the same. Live shows and ad-supported content pull in more ad dollars per hour even if the time spent is lower, different from the traditional binge format of old.

We have been looking at Netflix for a while here as it is down 50% from highs, but Apptopia data was showing time spent per DAU falling Q/Q. So again great way to manage fundamental risk. We will be looking for a rebound, and Netflix may get interesting…

Under the hood, older cohorts actually moved higher, 36-45 hit 87 hours and 46+ hit 84 hours. The under-35 cohorts pulled the aggregate lower. Could that be social + YouTube? Not a large concern until it is, but again likely why the stock is down ~50% from highs.

[2] Quality Factor is doing what it’s supposed to do.

The Quality Factor long/short has been on a run, and this is the setup we’ve been positioned for. Not for a trade, but for an investment in specific names that have been “placed” in that factor.

Rotation into quality is happening because the macro is healthier and has stayed that way. Also feels like middle east risk premiums are compressing, which is a positive, and rates fell over the last week or so on lower CPI + PPI falling

So again a thesis we’ve been explicit about. When Quality outperforms it usually says the market is rewarding earnings power, not just multiple expansion for high flying momentum names, and that lines up cleanly with what the banks showed also.

[3] Banks came in clean across the board.

Consumer comments

Economy Comments

Let’s dig into that for a second. The big banks reported this week and the message was clear. Net interest income and margins were positive across the majority of them. Deposit growth and lending growth were both healthy. Each has their own nuances so we won’t get lost in the weeds.

The cleanest take is that the banks that control the majority of deposits and loans in the U.S. all delivered clean messaging on the economy and the consumer. All of them said the same thing. Hard to find any negative tilts.

[4] Our bank sentiment reading keeps perking up.

We now track CEO/CFO sentiment reading across the big banks. Every quarter we pull their earnings commentary and score whether their views on the consumer and the economy skewed positive or negative relative to the prior print.

This quarter’s reading continues to perk up.

Both the Economy and Consumer heatmaps are green across the board, and the quotes panels underneath show exactly why. Also read the main quotes, very constructive read as the messaging is remarkably consistent across the biggest banks in the country. In aggregate, many things are healthy, spending is fine, credit is holding, and there’s no sign of imminent stress in the underlying data. Good news.

[5 ] Financial Index, XLF just posted six consecutive up days.

On top of that, Signal 42 flagged XLF for six consecutive up days into 7/16. Financials leading the markets is never a bad signal. We have only seen this 3 times over the last 20 years where financials were up 6 days in a row while consumer discretionary was up 3 days in a row, and the market 12 months later has been up 100% of the time with an average return of +20%.

[6] Investors haven’t been this bearish on consumer in 20 years.

Then looking at the Fund Manager Survey that shows net allocation to consumer stocks (staples + discretionary) is at the lowest level since May 2006. Investors haven’t been this under-exposed to consumer names in 20 years. People are getting less concerned about the economy and rate hikes, and more concerned about AI. That leaves a real contrarian setup here. But again, this echoes our ongoing view that this area is ripe for activity.

[7] Real consumer data says the “consumer is cracking” story is wrong.

We have two charts here that are essentially saying the same thing, but differently. BofA’s card spending data by income group shows spending accelerating across every tier.

Higher income +5.8%

Middle +4.8%

Lower +4.8% Y/Y as of Jun ‘26.

Same pace, all three moving together. Then their after-tax wage growth chart shows the same convergence. Higher +4.2%, Lower +4.2%, Middle +3.5%.

When you tax-adjust, wage growth across income tiers is essentially running at the same rate. Real wages are positive across the board despite what others suggest, we shared in prior pieces. This is the data the underweight consumer positioning is missing.

[8] Business formations and AI adoption both say the underlying economy is healthy.

Business formations continue to run at elevated levels. We have pointed to this for a while now. That’s a leading indicator of economic activity. Underneath that, AI adoption is broadening out.

Small firms adopt AI at lower rates than large firms, but when they do adopt they use it more intensively. AI is a net positive for the economy in the way it’s diffusing, and it shows up in the productivity of businesses that are actually deploying it.

[9] Even Buffett noticed the speculation. We agree.

Warren Buffett said this week that it’s tough to find value when everyone is gambling. We agree. There is speculation in the markets right now, and it’s real. But it’s in pockets, not the whole market. Really around AI, Space, Chips etc. There are still names to own in those areas where the fundamentals are compounding and the multiple hasn’t gotten silly. That’s where our attention sits. Speculation somewhere doesn’t mean overvaluation everywhere. So good to separate those two views clearly.

[10] Robotaxi share is shifting fast.

Robotaxi is one of the pockets where actual operator progress is showing up in the data. Apptopia data shows that Zoox is taking share from Waymo faster than most people appreciate.

Pay attention…

[11] Venmo engagement is re-accelerating.

PayPal, which owns Venmo, was offered this week to be taken private. Suitors are a consortium of Stripe, Block and Advent. Now why? One part is their Braintree processing business, another part is their PayPal branded business. Lastly is Venmo, their consumer engagement layer.

Now how is Venmo doing?

Apptopia data shows that Venmo’s engagement metrics are re-accelerating after a stretch of flat reads.

[12] Semis are the crowded pocket. Etched shows competition is coming.

More data on investment positioning. Long Mag 7 and AI semis came in as the most crowded trade globally. Probably why, when there is some selling pressure, we see exaggerated moves. But there are some things to think about within the AI trade. A company like Etched just showed their Frontier Inference Clusters, purpose-built inference silicon designed to run transformers faster and cheaper than general-purpose GPUs. This is what we’ve been saying about the inference layer. As models stabilize, margins show up in the infra layer. Companies will look to either create their own silicon to sell, or use custom silicon. This does not mean that it kills companies like Nvidia, but it does re-shape the margin structure over time. Something to keep tabs on.

Net Net

So earnings season is back and the first reads are constructive. Banks are healthy, the consumer is spending, wages are positive after tax, and the economy underneath continues to form new businesses and adopt AI. Positioning in consumer is washed out at 20-year lows, which is exactly the setup we’d want to see for a contrarian move. Speculation is happening in certain pockets, we agree with Buffett there, but the broader market is doing what we’ve been calling for. Rotation into quality, financials leading, and the companies that have been treading water while growing are starting to work. That’s the story of this week.

That’s all for this week!

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.

Big Picture

💡 The rotation we have been calling is starting to show up.

We’ve been talking for months about how the market has been an “AI or nothing” story at the index level, and how underneath that surface there’s a lot of durable, cash flowing business getting left on sale. This week the data builds on that view, plus some other company data points worth looking at.

For one, non-AI earnings are accelerating. Cash App just put up two of the three biggest 2-week engagement gains ever, back-to-back. Prediction markets pulled clear separation from the legacy sportsbooks. And Consumer Discretionary on an equal-weight basis is sitting near great financial crisis relative lows despite the sentiment recovery. That’s a nice setup.

Over the last 10 days we’ve seen the rotation start to move. The ceasefire question creates a little bit of headwind, but the magnitude of that as of today looks less impactful than it did a couple weeks back. This market has been very different, momentum driven so lets dig in.

[1] Cash App may be finding its mojo back.

The wand?

Cash App launched the wand, an interesting looking way to pay.

We went back and looked at the history of Cash App engagement, and the two-week window around the Pay Wand launch on June 4 was arguably the strongest engagement move we’ve seen in over a year.

Using Apptopia data which allows us to get pretty granular, the two-week move was roughly 45% bigger than the same window last summer, and it’s the largest 2-week absolute move in the entire 16-month dataset we looked at.

This is important since Cash App has historically been well known for its ability to create energy and demand off of their social tactics, almost like growth hacking, and the Wand, which is a payment solution that has a cool look for the right individual, seems to have hit. We expect another hardware drop here shortly as Cash App employees on X continue to hint at it.

[2] Kalshi and Polymarket pulled away from the sportsbooks.

More data from Apptopia to help inform us on the prediction markets.

We just had our podcast on prediction markets (Prediction Markets Aren’t Really About Betting… Maybe) , click that link to listen.

But with that we wanted to dig into the data, so this is what we learned.

Since June 1, Kalshi is +185% and Polymarket is +180% on US mobile daily active users, both accelerating after the World Cup started on June 11. That is massive.

DraftKings is +65% and FanDuel is +22% over the same window, both fading after their peak. Now to be fair, Kalshi specifically is spending a massive amount on sales and marketing, so yes, that’s creating real demand for the platforms. But ultimately they’ve opened up the door for something bigger as we express in the pod. Since we use their data ALOT here, they were kind enough to provide 30 days free to anyone out there that wants to cut up their data too.

[3] Meta Muse ranks #2 and #3 on the model leaderboards and Muse Spark 1.1 Leading in Many Categories.

Meta just launched a couple of new models. Muse Image ranks #2 on the Arena text-to-image leaderboard with a 1,280 score, sitting behind only GPT Image 2. Muse Video ranks #3 on text-to-video at 1,459, above Sora 2 Pro and every Veo variant Google has out. It just shows you Meta is back in the game as it relates to model delivery, and these models are going to be built inside of their apps specifically.

Good timing as it relates to people questioning Meta’s AI story. We had our podcast on this recently along with the Pitch to the PM session, worth a listen if you want the deeper view.

[4] Corporate priorities are lining up with our identity thesis.

Two data points, same thread. On the agentic AI side, 57% of respondents say agents acting outside intended context or policy is their top identity concern, and 56% cite over-privileged agents.

On the info security budget side, Identity Security is the #1 priority for the next 12 months at 68, up from 64.

AI-focused Security jumped from 49 to 58, the biggest mover on the board. If you’ve followed our work, this is further evidence of how we’re playing artificial intelligence outside of chips and energy and infrastructure.

The explosion of artificial creation of people, voices, and agents is going to require a massive amount of verification layers sitting between agents and agents, agents and humans, avatars and humans. We’re seeing that with Okta and Clear. This is corporations telling us the same thing.

[5] AI is an accelerator of jobs, not a replacement.

Firms that adopt AI with high intensity grow headcount 10% over the two years following adoption. Low-intensity firms see no difference from the control group.

There will be companies and industries where AI is a clear replacement, but our view has been that AI is more of an adjacent enhancer of productivity inside of a firm.

Productivity per person rises as they’re able to do more, which means less potential hiring for that specific role. But eventually that same person, who is now multiples more productive, runs into their own bottleneck. The lower bar of what you’re able to do rises. The upper bar of what you’re capable of rises. But eventually you still run into a productivity gap and you need more people. Long-time readers know we’ve been on this for a while.

[6] The S&P Equal Weight vs S&P 500 correlation is at 78%, lowest since 1991.

The 1-year rolling correlation between the equal-weight S&P and the cap-weighted S&P 500 just printed 78%. Long-run average is 96%. This is further evidence that this has been a one-way highway for the index.

Reference last week’s newsletter where we had our quarterly investment letter and our investment podcast session with Luis, my partner here at Avory, walking through the major themes. Over the last 10 days we’ve seen a pretty significant shift and the rotation we’ve been talking about start to show up.

Does it hold? We hope/think so. Now you have the ceasefire question, so maybe that creates a little bit of headwind, but the setup is there.

[7] Consumer Discretionary equal-weight is back at cycle lows.

This one is wild.

On a relative basis, discretionary equal-weight sits back near great financial crisis lows despite the economy doing good enough. This is further evidence of what we’ve been talking about for the last couple of years. The market has been pretty fearful of what’s happening in inflation and the economy, and yet inflation and the economy have gone counter to those negative views.

Our read has been fairly constructive. To us as long-term investors, we see opportunity here from a valuation perspective and a quality perspective, and we continue to be long certain pockets of it.

Since the ceasefire took place, discretionary started to levitate. While the ceasefire is somewhat in question, the magnitude of the implications look less impactful today, and we should start to see the broadening trade continue, including discretionary.

[8] Non-AI earnings are accelerating.

Why broaden? well.

AI earnings growth is still running well above the rest of the index, no argument there. But the story here is the other bar. Ex-AI earnings growth has been accelerating for four straight quarters and has flipped clearly positive. That underpins the broadening trade, and it’s probably a big part of why we’re seeing the transition. Doesn’t guarantee it, but the setup is in place. Same story, different week.

[9] The AI complex is now roughly a third of the S&P.

This one pairs directly with the last chart. From January 2 to June 30, the AI complex ex-hyperscalers grew from 23.8% to 31.1% of the S&P 500’s market cap. Semis and semicap alone went from 13.3% to 18.0%. Hyperscalers ticked slightly lower, from 19.9% to 16.9%. Other S&P, which includes Apple, dropped from 54.2% to 49.5%. So if non-AI earnings are accelerating, and AI dominates the index, then the question becomes: should you be considering the non-index? Or a little of both. That’s where our head is at.

[10] Good example of dislocation.

Someone asked me the other day, “what does dislocation mean?” Here you go.

This is a really clean reflection of how you can have a detachment between a company’s fundamentals and its stock price.

Roblox bookings are up 220% since Q1 21, a 26.2% CAGR. Stock is -27% over the same window. Part of it is nuances specific to the Roblox story, but this is less about Roblox directly and more about the idea that over the last year you’ve had people questioning what AI is going to do to certain digital assets, and questioning what’s going to happen in the economy.

These things tend to resolve favorably if the fundamentals continue to persist in a constructive way.

There are a bunch of these divergences out there right now.

[11] From Signal 42… Oil surged 7.2% in a single day.

Oil moved 6% to 7% roughly Wednesday. Signal 42 pulled the history of 1-day moves of 7.2% or more going back to 1986, and typically when you get these type of large moves they mark transitions from one period to another.

We saw the rise in oil during the conflict, and now we’re simply seeing moves on the way down as people question what’s happening in the Middle East. Overall, if you look over the last 5 years, oil prices have pretty much stayed in this range. In theory, if you weren’t concerned about $70 or $80 oil 5 years ago, you shouldn’t be all that concerned today either.

Net Net

Rotation is showing up. Non-AI earnings are accelerating, real-time engagement data at Cash App and the prediction markets is running hot, corporate identity spend is validating the Okta and Clear thesis, and Consumer Discretionary is sitting at relative lows even as economic conditions have remained steady. That is opportunity. Meta is back in the model game. Roblox is one of many examples of fundamentals detaching from price. That tends to resolve favorably over time.

That’s all for this week!

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.

Sean Emory, Founder and CIO of Avory & Co., asks whether prediction markets like Kalshi are really about betting, or whether they could expand society's capacity to transfer risk. Using Florida's hurricane insurance crunch, the Generac hedge, and AI-enabled micro-insurance, he argues insurance is capacity, and technology may finally be able to expand it. 

Chapters
00:00 Welcome and setup
01:19 Why prediction markets matter
02:49 Markets as information engines
05:55 From betting to bigger thesis
07:17 Insurance prices uncertainty
09:28 Capacity and risk transfer
11:41 Florida hurricanes case study
13:06 Hedging with opposite exposures
15:04 Micro risks and AI automation
18:31 Tech stack enables new markets
21:15 Liquidity, market makers, regulation
22:56 Wrap up and key takeaway

Disclaimer
This conversation is for educational and informational purposes only. Nothing discussed is personalized investment, legal, tax, or insurance advice. Views reflect our thinking as of the recording date and may evolve as new information becomes available. Do your own research and consult a qualified professional before making financial decisions.

More at avoryfunds.com