SambaNova Is Worth $11 Billion — in Nvidia's Shadow
SambaNova has reportedly raised $1 billion at an $11 billion valuation, per CNBC. The same week, Nvidia sits on roughly $500 billion in AI chip bookings alone. The round isn't a price — it's a wager that the inference market can fragment, and the numbers could settle it within a few quarters.
An $11 billion valuation for a chip challenger, set against a competitor holding roughly $500 billion in AI chip bookings alone. That is the contrast SambaNova now has to live with after its new funding round.
According to CNBC, the company has raised $1 billion at an $11 billion valuation, with a possible IPO in 2027 in its sights. The CNBC article was not accessible when this piece was written; the figures are relayed via TLDR AI's summary of the story, and details such as the lead investor and the exact round format therefore remain unconfirmed here. What remains is still enough to pose the question the round actually represents: What, exactly, are investors buying for $11 billion?
The Number That Measures Every Challenger
Nvidia's order book is the yardstick. The company reportedly holds around $500 billion in AI chip bookings covering both 2025 and 2026, according to The Motley Fool — a backlog investors can track quarter by quarter as a reliable indicator of future revenue. None of Nvidia's challengers comes anywhere close to combining revenue of that magnitude.
That makes the SambaNova round less a funding story than a priced prediction. Investors are not betting that SambaNova beats Nvidia on flagship models or training runs. They are betting on something narrower and more concrete: that inference — running already-trained models — is an economically different game from training, and that specialized silicon and full-stack systems can win that game.
SambaNova has positioned itself precisely there: as a full-stack vendor of rack-based systems aimed at inference workloads, not at Nvidia's training fortress. The company's own account of that positioning is just that — a claim. The delivery and contract evidence behind it is not documented in the sources this piece rests on.
Why the Capital Flows Anyway
The obvious reading — "Nvidia is untouchable" — doesn't explain the funding. Two factors do.
The first is the scale of demand. Hyperscalers are estimated, per figures cited by The Motley Fool, to spend roughly $725 billion in planned capex on AI infrastructure. At that scale, buyers actively want an alternative, regardless of benchmark results. A second supplier reduces pricing and delivery risk. Demand for a "source two" exists whether or not the challenger is technically superior.
The second is the supply scarcity that follows from the same number. When the backlog alone covers two years of production, many buyers have to queue. Queues create market room. That is the mechanism behind the $11 billion: not that SambaNova is better, but that the market is so supply-constrained and capital-rich that even a distant second place can be worth a fortune.
The capital stack around Nvidia illustrates just how concentrated all of this is. Goldman Sachs has reportedly mobilized banks, wealth managers, insurers, and private credit firms to finance Nvidia's infrastructure push, according to PYMNTS — with the investment bank placing the debt into private credit funds and public vehicles. Even the money flows around the challengers, in other words, run largely through the incumbent's ecosystem.
If Investors Are Right — and If They're Wrong
If the prediction lands, we won't notice it as a dramatic transfer of power, but as marginal shifts: Nvidia's pricing power erodes somewhat, inference costs per token fall faster, and enterprises gain negotiating leverage in procurement. For anyone building AI products, that is a concrete consequence — cheaper runtime and more genuine supplier choice.
If the prediction fails, the round becomes a marker of AI-infrastructure froth. The comparison is sitting right there: Databricks recently closed a strategic round of $5 billion at a $190 billion valuation, according to Forbes. Lovable raised at a $13.3 billion valuation, per the WSJ. Cognition is negotiating a valuation of at least $40 billion, according to PYMNTS. August's funding cycle is pricing the entire stack aggressively — and a valuation is not a delivery.
The counterargument deserves to be stated without softening. Nvidia's software moat (CUDA) is real and expensive to leave. Its grip on the supply chain gives it a power no challenger valuation dissolves. And SambaNova's own revenue figures, customer contracts, and shipped systems have not been disclosed in the sources available here — there is no confirmed delivery record to point to. Which means the $11 billion, today, prices a hope of fragmented inference, not a confirmed share of the market.
The Falsifiable Question
The best thing about this story is that it can be settled with numbers, not opinion polls. Nvidia's backlog is reported quarter by quarter. Hyperscaler capex is reported too. And SambaNova's path — delivered systems, signed contracts, and eventually a possible 2027 IPO — will reveal whether the round was a prediction or a price bubble.
The next few quarters of hyperscaler procurement decisions, not the press releases, will price this wager.