Nvidia told investors on August 26 that it expects 70 per cent revenue growth in fiscal 2028, a forecast so far above Wall Street’s 44 per cent consensus that it sends the chipmaker’s projected annual sales to roughly $673 billion. That would put Nvidia ahead of Apple and Alphabet by revenue, behind only Amazon among US technology companies. CFO Colette Kress delivered the figure on the Q2 fiscal 2027 earnings call — the first time Nvidia has provided a forecast this far out.
The Numbers
The forecast builds on a quarter that doubled year-over-year. Q2 fiscal 2027 revenue reached $96.2 billion, with data-centre revenue up 117 per cent to $89 billion. Nvidia shares rose roughly 4 to 5.6 per cent in extended trading, the variance reflecting different market data providers rather than any change to the forecast itself.
Based on the current Wall Street consensus of $396 billion for fiscal 2027 (ending January 2028), a 70 per cent growth rate puts fiscal 2028 at approximately $673 billion. CNBC noted that this would move Nvidia past Apple and Alphabet, leaving only Amazon — still primarily a retailer — ahead among US tech companies.
The fiscal 2028 figure is not a revenue guidance number in the traditional sense. Nvidia has not previously given a full-year forecast that far in advance. CEO Jensen Huang said the company did so now because it has visibility into next year’s compute requirements and wanted its supply chain partners — those providing land, power, and complementary data-centre infrastructure — to have the same information.
Demand Outruns Supply
The constraint is not demand. It is components. Huang said shortages in memory and other parts prevented Nvidia from guiding higher. “Our demand is much greater than 70 per cent,” he said on the call. “Our supply allows us to confidently deliver 70 per cent.”
This matters because it reframes the AI infrastructure story. For the past two years, the question was whether hyperscaler spending would hold. Now the question is whether the supply chain can physically deliver enough chips, memory, and networking gear to meet the order book.
The for(geeks) analysis highlighted European semiconductor stocks rising after the results — ASML up about 2.5 per cent, STMicroelectronics, Infineon, and BE Semiconductor each up 2 to 4 per cent — suggesting investors expect the spending wave to lift the entire component chain, not just Nvidia.
The Customer Base Is Broadening
The other shift worth paying attention to: Nvidia’s customer base is no longer a handful of hyperscalers building for a few frontier labs. Huang grouped the new buyers under the label ACIE — regional AI companies, neoclouds, startups, and enterprises. He described them as “largely invisible” a year ago but now a growing category deploying AI for production work, not just model training.
“This time last year, one lab alone was driving the buildout,” Huang said. “Today, we have a golden age of new AI labs and startups, multiple frontier labs scaling in parallel, a thriving open-model ecosystem and physical AI coming online.”
A broader customer base reduces dependence on any single hyperscaler’s capital budget. It also gives Nvidia more opportunities to sell networking, systems, and data-centre components alongside its processors. Neither the earnings release nor the call broke out how much revenue comes from those adjacent products versus GPU sales.
Nvidia as Banker
The financing dimension is the part that deserves scrutiny. Nvidia is investing in model developers including OpenAI and Anthropic, backing neocloud providers, and helping arrange credit for data-centre construction. Earlier this month it announced a program with six major financial firms to arrange up to $500 billion in data-centre financing. It has also committed $105 billion in financial support for the Ohio compute campus where OpenAI is expected to be the tenant — a deal we covered in detail when it was confirmed.
The circular risk is obvious: Nvidia helps fund customers who then buy Nvidia hardware. Huang’s defense is that these companies need tens of billions before they can borrow cheaply on their own. “This is the first generation of startups that needed tens of billions of dollars to get funded,” he said. That is Nvidia’s position, not an independent assessment of the financing risk.
What stands out is the contrast with the earlier $250 billion guarantee talks reported in July. Three weeks later, that number became $105 billion in credit plus a $1.5 billion energy investment. The trajectory is real money flowing into real infrastructure — but the financing structure makes Nvidia both a supplier and a direct participant in the capital required to buy its systems. The unanswered question is whether the ACIE customer base can sustain the projected growth without Nvidia’s balance sheet continuing to subsidize it.
The Hugging Face Angle
The forecast landed the same week Nvidia agreed to buy Hugging Face for $12.9 billion, the platform where most open-source AI models are hosted. The acquisition and the revenue forecast tell the same story from different angles: Nvidia is positioning itself as not just the chip supplier but the infrastructure layer underneath the entire AI economy — chips, cloud, model hosting, and financing.
What This Means
A $673 billion revenue company selling AI infrastructure is not a chipmaker in any traditional sense. It is closer to a utility — the entity that powers the compute layer everyone else builds on. The forecast assumes demand keeps expanding faster than supply can catch up, and that Nvidia’s financing strategy does not come back to bite. Both are reasonable bets. Neither is guaranteed.
— CJ Murden, editor of Singularity.Kiwi. Former digital technologies teacher, author of AI-focused books. Writing with a New Zealand focus.
