Part 2: Meta bets on Nuclear, while Waymo share grows.

This week’s issue is part two of our broader “State of AI” conversation. Last week focused on adoption and infrastructure trends. This week, we extend that story by tracking how product pull, energy demand, and ecosystem dynamics are playing out in real time. From Meta’s nuclear deal to Google’s retention data and the rise of the agent layer, we’re seeing early signs of how platforms and users are adapting.

At the same time, we heard from dozens of public companies this week during tech and consumer conferences. Our conclusion? The demand environment remains firm. From First Watch and Boot Barn to Paycom, Salesforce, and JAMF—companies are talking less so about macro headwinds and more about execution. The consumer remains steady. Enterprise spending is holding.

The economy continues to moderate with jobs gains becoming less and less, which is strengthening the case for rate cuts. Something we continue to believe is less about fighting crisis, and more about enabling momentum. Now lets get to the data this week.

Here is the summary if you want just that:

• ChatGPT vs Google retention data: +2,259 bps

• +27% Waymo share

• 20 Years of nuclear power for Meta. Why?

Last week, we shared a wave of data on the state of AI—from Nvidia’s earnings to multiple industry surveys.

This week, we’re doubling down. The space is evolving quickly, and we want to stay close to the signals that matter most.

First up. Meta just signed a 20 year deal with Constellation for clean nuclear energy. The move isn’t just about going green, it’s about securing the raw energy compute infrastructure AI demands. 1,121 megawatts isn’t trivial. That’s “always-on” power for a company training trillion parameter models. This should be taken as signal as Zuck rarely misses with directional bets.

To understand the magnitude of this shift, consider the compute eras. We went from ~1M mainframes to billions of mobile devices. Now, the AI era could scale to tens of billions of intelligent endpoints—a full order of magnitude larger.

And it’s not just a hardware cycle, it’s faster. GenAI is tracking ahead of both the internet and PCs at similar points post launch. The user ramp is happening in real time, with infrastructure still catching up. This is why I keep talking about this every couple of weeks to update you as we are tracking things daily.

What drives this scale? One great product. In every tech cycle, a “killer product” pulls demand forward and lays the groundwork for a product suite. AI will be no different, winners will start with one dominant experience and build out from there. Simple image via Rex Woodbury.

Even the "physical world" AI story is moving fast. In just 20 months, Waymo went from 0% to ~27% rideshare share in San Francisco. The takeaway? AI isn’t hypothetical, it’s commercial, it’s scaling, and it’s taking share from incumbents. Yes Uber owns millions of users across food, grocery and rides, but the data is the data.

Product adoption. On the software side, ElevenLabs’ traffic continues to surge, hitting 20M+ visits. AI-native apps with tight focus and clear user value are seeing exponential demand. Last week we spoke how chat bots are leading the use cases today.

Meanwhile, job postings for white collar analytical roles like Market Research are falling sharply. Down nearly 2x faster than the baseline. AI isn’t just adding new categories, it’s absorbing old ones. However Ai as of now looks like an enhancement rather then the end of the world scenario for jobs as we spoke last week 80% of those surveyed are seeing job growth from ai.

And yes, Google remains sticky. But ChatGPT’s retention has been quietly rising +2,259bps over the past 15 months. Users are starting to split habits, especially for information seeking tasks. Are we entering a world, like we saw with Facebook when TikTok or Snap emerged, where users aren’t replacing, but simply adding? A shift from default to dynamic.

Still, Google isn’t easily disrupted. With nearly 55% of desktop traffic going to the core search domain, and additional share across mail, docs, calendar, and more, its ecosystem gravity remains massive but that search component is at risk unless they evolve.

We saw what happens when Google launches new products like VEO 3. Daily visits to DeepMind’s site spiked to 900K+.

Google’s also working on “AI Mode” and usage is creeping up too. 1.25% of desktop users are now opting in. While early, this opt-in trend could be an indicator of them being able to shift.

The story loops back: as AI accelerates, so does the need for infrastructure. Meta’s nuclear bet isn’t about ESG, it’s a capex bet on where the world is going. We're moving from broad “AI capabilities” to real product led pull. The next leg of growth won't come from hype only (maybe some), it’ll come from usage. From retention. From energy. From memory. From the agent layer replacing app navigation. I will talk more about agents next week as I am collecting data on that as we speak!

That’s all folks. See you next week!

---

About Avory & Co.

Investing where the world is headed. 

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

Speak to us: Schedule a Brief Zoom Meeting

Send us an email: Team@avoryco.com

Want to invest? We are on most platforms. 

Want More

🎥 Avory YouTube Channel

🎙️ Avory Podcast 

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