Nvidia’s “hidden” data-center business beyond hyperscalers is growing faster, raising questions about demand and concentration

Nvidia’s data-center business has long been closely associated with the largest cloud providers, the so-called hyperscalers that built much of the early demand for AI accelerators. But a new market report argues that a major part of Nvidia’s opportunity now sits “beyond big tech,” describing a hidden $40 billion business that is growing at roughly twice the pace of the rest of the company’s ecosystem.

The report, published by Yahoo Finance on August 27, frames the change as a fast-moving evolution in Nvidia’s data-center customer base. It says the list of buyers is “changing fast” and that the “new buyers look nothing like” the hyperscalers who were central to Nvidia’s early dominance in AI compute.

At the center of the claim is Nvidia’s dependence on demand for accelerated computing, particularly for AI training and inference. Nvidia sells the GPUs, networking and related software platforms used to build AI data centers, and it benefits when more organizations deploy AI systems at scale. In that context, a broadening of customer types would be consistent with the idea that AI compute is spreading beyond the biggest cloud platforms into industries and organizations that are not traditionally lumped into “big tech.”

Still, the report’s most important point is not only that the customer mix is changing, but that the “hidden” slice is growing quickly. The article uses the $40 billion figure to describe the magnitude of this segment and characterizes its growth rate as about twice as fast, implying Nvidia may be benefiting from a demand wave that is not fully captured by the market’s usual focus on hyperscaler spending.

Nvidia does not typically comment on customer-by-customer revenue splits in a way that lets outsiders precisely map these dynamics in real time. Companies in this position often disclose aggregate segment revenue and high-level market demand trends, but not the internal breakdowns necessary to independently verify a specific “hidden business” estimate like the one cited in the report. As a result, the Yahoo Finance framing is best read as a market interpretation rather than a confirmed internal Nvidia metric.

Beyond demand breadth, the shift in who buys Nvidia’s data-center hardware matters because it can change how predictable revenue is across product cycles. Hyperscalers tend to place large orders tied to platform roadmaps, while non-hyperscaler customers can adopt AI more irregularly, driven by procurement timing, regulatory constraints, and project-based budgets. If the growth described in the report is coming from new buyer categories, Nvidia may face a different pattern of buying, even if long-term adoption rises.

There is also a competitive dimension. Nvidia operates in a crowded landscape of AI hardware, including alternative GPU suppliers and custom silicon used by cloud providers and enterprise systems. A move away from hyperscalers toward a wider set of customers could benefit Nvidia if it means more users need its broader software and systems approach. But it could also expose Nvidia to different adoption hurdles if the new customers choose different architectures, rollouts, or deployment strategies.

The key uncertainty, therefore, is what Nvidia itself is willing to disclose about the drivers behind the “hidden” growth figure. The report does not, in the information available here, provide a primary-source explanation from Nvidia detailing whether this $40 billion segment corresponds to a particular customer group, geography, workload type, or product mix. Investors and analysts will likely look for corroboration in Nvidia’s future disclosures, including commentary about data-center demand breadth, customer onboarding, and any changes in the mix of its platform deployments. Next, the market’s attention should turn to whether subsequent Nvidia updates support the idea that non-hyperscaler demand is scaling sustainably or whether the growth rate is partly a temporary spike tied to specific customer rollouts.