$GOOGL clicks grew 8% and 15% on a 2-Y; continued mid-teens growth.
I’m more likely to believe management’s argument that FX, FIFA timing, Asian advertisers, and AI Overviews were the culprit affecting monetization — vs. any type of meaningful headwind from agentic/LLM.
Probably an overreaction here.
What's difficult about Jensen is he is right on the issue but loose with facts and tends to exaggerate to make his point.
There has been one sizeable deal signed that valued compute at $50B/GW – Google paying SpaceX for 200MW of compute for what is effectively constrained spot/on-demand pricing.
However, you don't need $50B/GW to earn attractive IRRs.
H100s earn double-digit unlevered IRRs at ~1.45 an hour.
@SemiAnalysis_ GPU rental pricing index still has H100s at $3.26 4 years into their product life.
If an H100 averages $2.50/hour over its life, the unlevered IRR is north of 50%.
Now suppose they do start winning enterprise inference workloads. What stops a hyperscaler from shipping a bare-metal, hypervisor-less offering – similar to the neos' current offering?
Has AWS not done it because they lack the talent? Or because it isn't competitively differentiated and they can replicate it the moment it matters?
The benefits neoclouds currently offer, while real, do not constitute a sustainable competitive advantage (h/t Bernstein).
And 70% understates it.
Look at the contracted book. Even within the neo mix, at least half (probably 70%+) is hyperscaler + AI lab offtake.
The "synthetic hyperscaler" is a good narrative, but doesn't pass the sniff test.
Here's my capacity build to 2030. I don't think I'm being unfair to $CRWV, $NBIS, the other neos, or $SPCX.
Hyperscalers still control 70%+ of the watts in 2030.
$AMZN $MSFT $GOOG $META $ORCL
The reality is what we are seeing unfold is Nvidia speedrunning the creation of a synthetic hyperscaler.
Apologies in advance to all the investors who are stuck in their priors that this will trigger.
But what is a hyperscaler? Strip it down and it’s a scaled infrastructure collective of CPUs, networking, storage with a development platform on top. It fulfills two purposes. Financial: it pools and smooths the financial obligations of its users, renting infrastructure as opex instead of capex. And Operational: it builds software that makes consumption the underlying primitives simple by abstracting them away.
The hyperscaler makes a healthy 35-40% operating margin by buying hardware at bulk pricing, pooling scale to get a lower cost of capital, and driving utilization of that hardware with software that shares and shards workloads across many customers.
But in the age of AI, the atomic units of compute changed. Training (massive coherent clusters) and inference (agentic workloads) - require a fundamentally different configuration of resources. These new workloads require dramatically more accelerated compute, shifting the design target from multi-tenant utilization (the cloud era) to absolute workload performance (the AI era). The economics of the data center inverted. A giant, redundant fleet of Amazon Basics CPUs and storage doesn’t work when the job is synchronous training and one straggling node stalls the entire cluster. For inference, tokens per watt and time to first token dominate the economics, not how many VMs you can pack in a box. And none of it works in a world of limited power (at least in the West. Maybe in China).
As Nvidia built more compute and sold it to the hyperscalers, it faced a fundamental problem. The hyperscalers had classic innovator’s dilemma - expecting 35-40%+ op margin, along with an underlying desire to commoditize Nvidia's 75% GMs with their Amazon Basics equivalent. Pay an ASIC vendor a 25% margin instead of Jensen’s 75%, then stack your own 40% on top! They owned the customer relationships too, enterprises developed on AWS, Azure, GCP and their data was captive there too. But most important of all, these companies moved at their own pace. They were not scrappy or hungry to operate at the pace Nvidia or the AI labs felt was necessary to build out compute to fulfill the demand in front of them. They would never look at retrofitting a 35MW site outside of Ashburn, Virginia!
Meanwhile, a group of hungry entrepreneurs noticed the fat margins the hyperscalers earned renting what was basically stock Nvidia hardware with limited software on top, and started building businesses around it. Nvidia - skeptically at first - recognized that working with these partners would lead to faster development cycles and competitive fires and pressures for the ecosystem. Thus the neoclouds were born.
The software these neoclouds co-developed with Nvidia were purpose built for the new workloads. They solved the new problems and requirements operating the new infrastructure needed. They were ready with hotswaps, they did predictive maintenance, they built new storage software that was built for training with cheaper ingress and egress fees, because their competitive drive was to win workloads, not to lock in enterprise data on their platform. And it was working - AI labs started preferring to work with them over the hyperscalers. Common complaints on the incumbents: too slow, too particular with how their clusters were built, virtualization and networking overlays that made GPU clusters underperform stock Nvidia reference designs. Neocloud bare metal was cheaper too as their teams built AI software, not a cloud data warehouse business. And they were happy to run at half the margin (~20%) that the big guys would never accept.
But the hyperscalers still had one structural advantage: their balance sheets. Investment grade. Able to fund speculative capacity ahead of demand and rent it out at much higher spot rates. The neoclouds couldn’t play that game as lenders would only finance hardware that was already contracted with offtake. And more expensive if that offtake were the labs which at an earlier point were much more speculative. If only they could build ahead of demand, they could maybe earn the kind of returns Elon is achieving on Colossus.
But the twist is that balance sheet edge is eroding in real time. Google just printed its first negative-FCF quarter and raised $50B equity. Microsoft is carrying $329B of leases signed but not yet commenced. Even the IG balance sheets hit the wall - more capital had to come from somewhere else.
And that's how we got to where we are today. Look at what Nvidia has actually built. The operational half of a hyperscaler: DSX OS and Mission Control to run and operate GPU fleets, DSX reference designs and Omniverse digital twins as hardened playbooks for building a data center itself. Dynamo for inference serving. All the old secret sauces of the hyperscalers built specifically for new age data centers that they have led the way in architecting. Offered to any hungry, technically competent team with a serviceable site.
And then the financing half: the revenue share and credit support model that smooths utilization across a distributed fleet the way multi-tenancy used to. Support the operator, release capacity to demand, share in upside, and now bring $500B of third-party capital to the table. Nvidia standardized the asset with reference designs, proved the compute was “fungible and transferable across customers and operators” and showed infrastructure investors DD unlevered yields across 7-8% hurdles. Those investors wet their beaks on early special situation financings, saw the paybacks, and understood the demand was global. That’s why Jensen spent 2025 flying around Europe, the Middle East, and Southeast Asia - these are the ground zero for new compute sites. The reality is that this didn’t happen just over the last 3 months. CoreWeave master agreement in 2023, the $6B spot reserve backstop in 2025 (to sponsor capacity for the inference clouds), the Blackrock AI Infrastructure Partnership in 2024, Brookfield’s $100B fund with Nvidia in 2025, KKR Helix with Nvidia in 2026. And now six independent financing platforms. Chess!
So now the three fears by name. Circularity? Monday was the opposite with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR bringing third party capital, independently underwritten apart from one another, replacing Nvidia’s balance sheet rather than just extending it. Useful lives / underwritability of these assets? CoreWeave just disclussed A100s, 6 year old silicon contracted through 2029 and pushed 25% price increase on its fleet in July. The collateral is aging more like an aircraft than a smartphone as feared. Market share? If you don’t see that the platform of Nvidia and the fungibility of this compute is the reason why this is even possible - the skeptics themselves are making the bull argument. The complaint that these platforms keep capital tethered to Nvidia and away from other ASICs / accelerators… $500B that can only buy Nvidia reference architecture is a moat dressed up as a risk.
So what were you doing when the first synthetic hyperscaler was built under your nose? :)
All views expressed are my personal views. Does not reflect the views of Altimeter or Nvidia or anyone else. Full disclosure I/we may hold positions in companies mentioned. Purely for discourse and thinking - no financial advice.
Torsten Slok at Apollo (whose economic impulses I usually agree with) argues the S&P 493 is spending heavily on AI with nothing to show for it. I strongly disagree.
Slok: "The S&P 493 is spending heavily on AI. But it is not showing up in profit margins, see the first chart below. Looking at the individual sectors of the S&P 500 outside tech shows that the picture is flat, cyclical or worse, see charts 2, 3 and 4. Health care margins have halved since 2015; consumer staples are stuck near 6% and consumer discretionary near 8%; energy and materials have given back most of their 2022-23 gains; real estate is going sideways; and the only genuine improvements look like an ordinary cyclical recovery rather than a technology-driven step change. The bottom line is that the AI capex boom is so far only showing up in the sellers' margins, not the buyers'. This is important because the longer it takes the S&P 493 to generate ROI, the bigger the downside risks to an economy and a market this concentrated in the AI trade." (Link: apollo.com/wealth/insights-n……)
I strongly disagree with this reasoning.
Margins are one component of returns on capital, not the returns itself (NOPAT margin × invested capital turnover = ROIC). The deeper issue is that the argument implicitly assumes AI productivity gains should accrue to the adopting company as higher margins. They absolutely do not have to... and in a competitive market, they won't.
Buffett described almost exactly this scenario in the 1985 Berkshire letter, on textile machinery:
"Over the years, we had the option of making large capital expenditures in the textile operation that would have allowed us to somewhat reduce variable costs. Each proposal to do so looked like an immediate winner. Measured by standard return-on-investment tests, in fact, these proposals usually promised greater economic benefits than would have resulted from comparable expenditures in our highly-profitable candy and newspaper businesses.But the promised benefits from these textile investments were illusory. Many of our competitors, both domestic and foreign, were stepping up to the same kind of expenditures and, once enough companies did so, their reduced costs became the baseline for reduced prices industrywide. Viewed individually, each company's capital investment decision appeared cost-effective and rational; viewed collectively, the decisions neutralized each other and were irrational (just as happens when each person watching a parade decides he can see a little better if he stands on tiptoes)."
That is my (strong) base case for AI adoption. Each deployment pencils out, everyone deploys, costs fall industry-wide, prices follow costs down, and the economic benefit lands with the customer vs. shareholders. Aggregate ROIC doesn't move because capitalism is designed to compete excess returns away. Returns will only durably improve for the handful of businesses whose moats let them capture and hold the surplus.
So flat 493 margins aren't evidence AI isn't working. They're roughly what you'd expect if it is working and the gains are being passed through as consumer surplus. Berkshire kept funding those mills for years precisely because each project looked rational in isolation. CapEx that has a competitive-necessity doesn't require a margin payoff to begin (or to persist).
