#Nvidia's Case for Life Beyond Hyperscalers
Nvidia's stock market narrative has centered on a slow bleed. Various analysts expect its share of the AI accelerator market to erode from near 90% into a range roughly between 70% and 80% by the end of 2026, estimates vary but the direction is consistent, mainly because hyperscalers are building their own custom chips. Nvidia's latest earnings call offered a different way to think about that number.
Management argued that hyperscalers are only part of the picture. A broader group, made up of governments building sovereign AI systems, regional cloud providers, so called neoclouds, and enterprises and startups deploying AI in their own operations, now accounts for roughly half of Nvidia's data center business and is growing 100% a year. If that group keeps expanding, hyperscaler market share matters less than investors assumed.
#Who Is Actually Buying the Chips
Each category already has real customers behind it. On the sovereign side, India's L&T, Yotta and Netweb are building government backed AI factories, Australia's Sharon AI has launched the country's first Cisco Secure AI Factory with Nvidia, and Nvidia has flagged additional sovereign deals across Asia, Europe and Africa.
Regional neoclouds such as CoreWeave, Nebius and Nscale lease out GPU clusters to customers who do not want to build their own data centers. On the enterprise side, Samsung uses Nvidia's cuLitho software to speed up chip manufacturing, Bristol Myers Squibb is investing in a Vera Rubin AI factory for drug research, and Red Hat has packaged Nvidia's software into a deployable AI Factory platform for large companies. Among AI native startups, nearly 20 companies, including Cursor, Figma and Together AI, now exceed $1 billion in annualized run-rate revenue, up from 13 companies in Q4 of last year.
A sharper way to see the markets in just 5 minutes.
Same news, different lens. We cut through the noise and hand you the overlooked ideas and the deeper read the crowd misses. Join 38,000+ investors seeing the markets differently.
#A Guide That Broke With Precedent
Management's confidence showed most clearly when Nvidia broke its usual habit of guiding one quarter at a time. It told investors to expect roughly 70% revenue growth in fiscal 2028, implying revenue close to $700 billion, well above the roughly $570 billion analysts had penciled in. Sovereign AI and regional neocloud revenue alone grew 35% sequentially and more than tripled from a year earlier.
#The China Front Nvidia Doesn't Control
Nvidia accounted for an estimated 55% of China's AI chip market as of September 2026, down sharply from near monopoly levels a few years ago. Nvidia's own quarterly filing acknowledges the mechanism, stating that being locked out of China's data center market has let rivals build larger developer ecosystems that now challenge the company worldwide.
Four homegrown chipmakers, nicknamed China's little dragons, are driving the shift. Shanghai Enflame Technology, founded by two former AMD engineers, just priced a Shanghai IPO worth close to $900 million, backed by Tencent as its top customer and 20% owner. Enflame's first quarter sales jumped nearly 1,475% year over year, even as its net loss widened to roughly $66 million. Rivals Moore Threads, MetaX and Shanghai Biren have all gone public over the past year too, alongside state backed national champion Huawei, and several are building their own CUDA alternatives to chip away at the software lock-in that has long kept customers on Nvidia hardware.
#The Case for Caution
None of this erases the underlying risk, and it now comes from two directions at once. Custom silicon from hyperscalers is still coming, part of the new customer segment relies on financing structures that are harder to evaluate than a straightforward sale, including a conditional guarantee of up to $105 billion supporting OpenAI-related leases at SB Energy's Ohio data center campus, and inside China, funded domestic rivals are chipping away at a market Nvidia once had almost entirely to itself.
The bull case is that a more diversified customer base outside China makes Nvidia less dependent on any single hyperscaler's spending plans. The bear case is that some of this growth is circular, with Nvidia's own capital helping generate the demand it later reports, while the China opportunity keeps shrinking regardless of how the rest of the business performs. Watch whether the non-hyperscaler segment keeps growing on its own merits, and whether China's share stabilizes near 55% or keeps sliding.