Altimeter Capital

Jensen told Brad (@altcap) on his @BG2Pod: "We've gone from pre-training to inference time reasoning. Inference is about to 1 billion X." Not 10x. Not a million x. A billion x. And our compute systems weren't built for it.

Altimeter did not vote or nominate any of the companies referenced herein. Please see the article for the selection criteria used and full list of companies selected. time.com/collection/tim…

We’re proud to see several companies in Altimeter’s private and public portfolio recognized on the 2026 @TIME 100 Most Influential Companies list. Congratulations to @AnthropicAI, @discord, @Meta, @nvidia, @OpenAI, and @SpaceX for shaping the future across AI, frontier technologies, and global platforms.

Cloud Giants Update: AWS (Amazon): $150B run rate growing 28% YoY (last Q grew 24%) Azure (Microsoft): ~$108B run rate (estimate) growing 39% YoY (last Q grew 38%) Google Cloud (includes GSuite): $80B run rate growing 63% YoY (last Q grew 48%, neither are cc)

Some good research from the @inngest team - Got early access to their 2026 AI in Production report. One stat that stood out: 20% of teams building AI now spend up to half their engineering time just keeping things running. Twice the rate of non-AI teams. Full report next week!

Another aha from MS&E 435 with @alighodsi last week when asked what's working in enterprise AI Enterprise AI does not need smarter brains. We are already there (Ali says we are at AGI) It needs the body. Brain = intelligence = models, reasoning, coding, agents. Body = context = workflows, permissions, systems, exceptions, and organizational memory. History rhymes: Electric motors took decades to show up in productivity statistics because factories first just swapped steam engines for electric ones. The unlock came when companies redesigned the factory around electricity. AI is similar: Every company has a “go ask Jane” person. Jane knows the customer nuance. Jane knows the exceptions. Jane knows why the process exists. Jane knows what is written nowhere. That context is not in the model. So even a very smart model makes dumb mistakes. Ali’s point: the bottleneck is not just intelligence. It is organizational memory. But the real unlock is workflow rewiring.

Exciting milestone from @ProfluentBio today: a major partnership with @EliLillyandCo that validates what frontier AI can do in drug development. This team has been building toward exactly this. AI will massively accelerate the drug discovery space

One aha from class with @alighodsi at MS&E 435 this wk: Open source closing the gap with closed source may be inevitable. Why? Distillation. The old training substrate was the internet. Common Crawl is roughly 2T tokens (10^12). The new training substrate is AI-generated output. If OpenAI + Anthropic have produced ~$50B of tokens at ~$5 per million tokens, that 10^16 of proprietary-model tokens in the wild. That is four orders of magnitude more than Common Crawl. (10^4) At some point, every training run is learning not just from humans, but from the exhaust of the best closed models. The gap may close because *the teacher cannot stop teaching*.

Castelion's hypersonics are delivering an asymmetric strike and deterrence capability to the @DeptofWar Another major contract from the @USNavy to integrate Blackbeard with core carrier air wing platforms.

@apoorv03 with @CrusoeAI Founder @ChaseLochmiller at Stanford: In Abilene (pop. ~120k), building an AI data center brought ~9,000 workers at peak and ~2,000 long-term jobs to operate it If you break down a “$100” of AI infrastructure spend, a huge portion goes to labor, electrical systems, and power, not just chips And despite the scale, these facilities use minimal water through closed-loop systems. AI infrastructure is often misunderstood, it’s as much about jobs and energy as it is about compute.

Altimeter’s Apoorv Agrawal (@apoorv03) with @CrusoeAI Founder @ChaseLochmiller at Stanford: “A gigawatt is basically what powers the whole city of Denver.” This isn't abstract compute. A single AI campus now consumes city-level electricity. The bottleneck has shifted. It’s not just GPUs, it’s whether you can actually access and deploy power at this scale.

As AI shifts software from seat-based access to API-based workflows, what happens to professional info & research terminal companies? First, customers add API spend while cutting seats. Then, vendors will likely reprice API higher, or moving toward consumption-based pricing? Most businesses are not built around a world where one internal AI layer can serve an entire team... E.g. Today - A hedge fund have 20 analysts, each with a @Bloomberg Terminal ($30K+ per year), or $600K+ annually. New Model - Buy 1 API key, pipe the data into an internal $Claude / LLM workflow. Cost fall by 50%+. This is what AI really changes: it centralizes information retrieval. The value of the UI + search / query layer disappears. And this goes far beyond finance. Anything below are at risk: - historically sold by seat - much of the value comes from search, retrieval, summarization, basic analysis, or light workflow - the underlying data or functionality can be exposed via API - the end user does not truly need the native UI The key questions are: -> Are customers paying for the interface or for the answer and the data? If just for the data, that is more exposed. -> Do users actually live inside the product all day? If yes, more defensible. -> Is the data truly proprietary or mainly an aggregation layer? Aggregators are more at risk. -> And if you plug the product into Claude, does 80% of the value still remain? More broadly, finance has always been strangely fragmented. - @Bloomberg has real-time data and some consensus. - @VisibleAlpha has the deepest consensus detail, but not all mgmt. guidance. - @AlphaSenseInc is a strong aggregator for sell-side research and expert content. - @tradingview has great charting, but no market caps or financials. - the list goes on... Maybe AI is finally the force that breaks those walled gardens. The real risk is never expensive data. It is expensive UI.

Congrats to @AnthropicAI, @baseten, @CrusoeAI, @databricks, @glean, @OpenAI and @physical_int on being named to the @Forbes AI 50 🎉 These teams are building great companies and shaping what's possible. Honored to support each of their missions. The future is bright, the momentum is real, and we are so proud to be on this journey with you. 🚀 And a huge congratulations to all 50 companies on this year's list. 💥

Brad's (@altcap) advice to the next generation: "Being the smartest person and solving the problem faster than every other human — that gets commoditized. You're not going to beat the machine." Network. Persuasion. Leadership. Make yourself bionic.

Altimeter Founder & CEO Brad Gerstner (@altcap) & Nvidia VP Sunny Madra (@sundeep) were guest speakers of Apoorv Agrawal (@apoorv03)'s Stanford class. The cost of inference has dropped 99% in two and a half years. That's not a rounding error; it's a fundamental shift in what's possible to build.

Most CEOs set 2025 people budgets before they saw the banger results of the new models / coding / co-work. Companies now making real time adjustments to lower tech headcount growth for 2026 & increase spend on tokens / intelligence. 🧐🧐

"Anthropic was literally counted out of the game last year.” “They executed their butts off. They took the lead. 2,500 people tight pulling on the ore in the same direction." At the same time, OpenAI isn't standing still — Codex is already one of the fastest ramps we've seen. But the bigger point: "This is not zero sum. The TAM of intelligence is dramatically larger than any TAM we've ever seen in our investing careers over the last two decades." Two great companies. One massive market. This is a multi-winner race, where intense competition is accelerating innovation and reinforcing a massive moment for Team America. 🇺🇸