Semiconductor analyst Jay Goldberg of Seaport Research delivered a sober warning on Marketplace Morning Report’s segment Chipping away at Nvidia’s chip dominance: “There is the potential for this to entirely disrupt Nvidia, so I think it is a pretty significant risk.” The comment lands at a sensitive moment for NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) shareholders, who have ridden the stock to a $4.87 trillion market cap on the back of the AI buildout.
The starting point matters. Almost the entire AI ecosystem currently runs on NVIDIA silicon, with CUDA, developer mindshare, and software lock-in forming a moat that has translated into $68.13 billion in Q4 FY2026 revenue and a Data Center segment of $62.31 billion, up 75% year over year. CEO Jensen Huang told investors that “Grace Blackwell with NVLink is the king of inference today.” The financial details are in NVIDIA’s most recent press release.
Why hyperscalers keep building their own chips
Goldberg’s thesis rests on the scale and capital of NVIDIA’s largest customers. Tufts economic historian Chris Miller, author of Chip War, framed the structural logic on the same segment: hyperscalers want to reduce dependence on NVIDIA and “specialize it to your particular needs.”
The evidence is on the income statements. Alphabet (NASDAQ:GOOGL) reported Q1 FY2026 Google Cloud revenue of $20.03 billion, up 63% year over year, with backlog nearly doubling quarter over quarter to over $460 billion, much of it served by in-house TPUs. Amazon (NASDAQ:AMZN) CEO Andy Jassy said “our chips business topped a $20 billion revenue run rate (growing triple digits year-over-year)”, anchored by Trainium. Meta Platforms (NASDAQ:META) is scaling its MTIA program while raising 2026 capex guidance to $125 to $145 billion. Google and Amazon are already offering their accelerators to outside customers, extending the threat beyond captive workloads.
The bull case for NVIDIA
Bernstein Research’s Stacy Rasgon pushed back on the horse-race framing, arguing on the same segment, “I don’t really care who’s winning or losing right now. I think it’s the wrong question.” His point: AI agents have sent compute demand off the charts, so the binding constraint is capacity. Every credible chipmaker, NVIDIA included, sells what it can make. NVIDIA’s $95.2 billion in supply commitments and partnerships covering OpenAI, Anthropic, CoreWeave, and Meta support that view.
What investors should watch
- Hyperscaler capex commentary on earnings calls, where combined 2026 commitments approach $500 billion-plus.
- CUDA stickiness as alternative software stacks mature.
- Whether external customers adopt Google’s TPUs or Amazon’s Trainium at scale.
- Whether AI compute demand keeps outrunning supply, which has propelled NVDA 92% over the past year.
NVIDIA still sits at the center of this buildout. Goldberg’s warning is a reminder that the customer list is also the competitor list.
Top Gaining Stocks
Top Losing Stocks