Nvidia is forecasting 70% sales growth next year, telling investors the AI infrastructure buildout has years of runway left rather than months. The guidance, delivered as the company briefed markets on demand visibility, lands as the loudest signal yet from the industry's principal beneficiary that hyperscaler and enterprise spending is not slowing into 2026.
The number is unusual for a company already operating at Nvidia's scale. Guiding a base of nearly $100B a quarter to another 70% expansion implies tens of billions in incremental annual revenue that has to be absorbed somewhere — by hyperscalers, sovereigns, neoclouds, and enterprises still in the early innings of buying compute.
The forecast follows Nvidia's most recent quarter, when the company posted $96.2B in revenue and guided the current period to $108B. Data center is the engine behind those numbers, and management has repeatedly said order books are booked out well beyond the next release cycle.
Key facts
- 01Nvidia forecasts 70% sales growth next year, signaling the AI capex cycle has years of runway left.
- 02The guidance follows a $96.2B quarter and $108B forward guide disclosed earlier this month.
- 03Amazon recently tripled its Nvidia GPU order for AWS, adding 2M Blackwell and Rubin chips.
- 04Anthropic committed to a $45B compute deal with Nscale built on Nvidia's Vera Rubin architecture.
The 70% figure also lines up with the deals being signed around it. Amazon recently tripled its Nvidia GPU order for AWS, committing to roughly 2M Blackwell and Rubin-class chips. Anthropic signed a $45B compute agreement with Nscale built on Nvidia's Vera Rubin architecture. Those two deals alone represent a level of forward commitment that supports the growth trajectory Nvidia is now guiding to.
Jensen Huang has spent much of 2026 arguing that the market is misreading the shape of the cycle. His pitch: the industry is not overbuilding for a demand curve that has already peaked, it is underbuilding for one that keeps re-rating higher as models improve and inference workloads scale. The 70% guide is that argument expressed as a number.
The counterweight is the growing chorus asking whether hyperscaler capex can keep compounding at this pace without stretching balance sheets or producing returns that justify the outlay. Microsoft, Meta, Google, and Amazon have collectively guided to combined AI infrastructure spending well north of $300B in the coming year, and a slowdown from any one of them would show up quickly in Nvidia's order book. Circular financing arrangements — where Nvidia invests in customers who then buy Nvidia chips — have drawn scrutiny from analysts questioning whether demand is as organic as the top-line numbers suggest.
There is also the competitive frame. AMD's MI400 series is beginning to appear in real deployments, Google's TPU generation is being marketed aggressively to third parties, and hyperscalers continue to pour money into in-house silicon. None of that is currently large enough to dent Nvidia's share, but the 70% forecast implicitly bets that it will not become large enough next year either.
For now, the guidance reframes the conversation. The question is no longer whether AI spending peaks in 2026 — Nvidia is telling investors, on the record, that it does not. The question is whether the customers on the other side of those orders can keep writing the checks and generate returns fast enough to justify the next round.
Skeptics will point out that Nvidia has every incentive to project confidence, and that a company selling picks and shovels benefits from prospectors believing the gold is still in the ground. The order book is real, but so is concentration risk: a handful of customers drive the bulk of the revenue, and any pullback in hyperscaler capex, sovereign AI budgets, or neocloud fundraising would compress the growth curve fast.
If Nvidia hits 70%, it settles the near-term debate about whether AI infrastructure is a bubble in real time — the demand is being underwritten by named buyers with signed contracts, not sentiment. If it misses, the miss itself becomes the story of the year, because the entire AI capex complex has been priced on the assumption that Nvidia's guidance is conservative. Either way, the next four quarters are the ones that decide how the 2020s AI buildout gets remembered.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
