Wall Street Wanted 45% Growth From Nvidia. It Guided to 70%, and the Real Number Is Higher.
Nvidia guided to 70% revenue growth and the stock jumped, but a KeyBanc analyst revealed the real demand figure is significantly higher, and understanding why the company cannot meet it changes everything about what you are actually buying.
John Vinh, Senior Research Analyst at KeyBanc Capital Markets, went on CNBC after NVIDIA (NASDAQ:NVDA | NVDA Price Prediction) reported and said what actually moved the stock. “The outlook for next year was very impressive. Street was looking for up 45%. That was what was, I think, put NVIDIA over the top here.”
The company guided fiscal 2028 revenue growth to roughly 70% against a Street bar near 45%. Shares rose 8.74% on the day and closed at $227.98.
The interesting number is the gap Vinh described between real demand and deliverable supply, because that gap changes what kind of company you are underwriting when you buy the stock.
How the 100% Demand, 70% Supply Gap Reframes the Risk
Vinh said the quiet part directly. “NVIDIA yesterday said that they actually have demand for 100% growth next year. But they’re actually supply constrained. And as a result is only currently able to grow 70%.”
Jensen Huang confirmed the same picture on the call, saying, “Our entire supply chain is challenged. And everybody is really running flat out.”
A company that cannot produce enough product is facing a manufacturing and supply chain problem, which carries different risks than a demand problem.
The bullish read is that the reported 70% growth understates the true demand, suggesting the backlog persists into 2028. Vinh believes it does. “I do think that demand is clearly durable through all of next year and probably into a good part of 2028 right now.”
The hidden risk is that supply-constrained revenue depends on partners NVIDIA does not control, including memory suppliers, foundry capacity, and data-center power (we profiled seven of the infrastructure suppliers behind that buildout in a free report). A demand figure the company cannot fulfill is also unverifiable from outside, so you are taking the guidance on trust.
Supply obligations climbed to $279.00 billion, largely for memory tied to Vera Rubin, which shows how big the check has to clear before the revenue does.
What 20 Times Forward Earnings Actually Means
Vinh set a price target of $330, saying “It’s based on 20 times our numbers.”
A forward earnings multiple is the price you pay today for each dollar of expected earnings next year. Twenty is roughly what an investor pays for a solid large-cap growing at a much slower rate than the one NVIDIA just guided to.
The stock currently trades at a forward multiple of 24x, which is not stretched for a company that reported 105.85% revenue growth last quarter.
Vinh’s target rests on the earnings estimate coming true at roughly the multiple the market already assigns. That is where real skepticism belongs. If fiscal 2028 EPS comes in below plan because memory prices squeeze margins to the guided 71%-72% trough, or because supply slips, the multiple is fine and the earnings number is not.
Merchant Alternative Question
Vinh also said, “There is no other merchant AI chip solution out there on the market that is a credible alternative to NVIDIA right now.”
As stated, that is true. AMD’s Instinct line is real but not at the same rack-scale system maturity, and Intel’s Gaudi has not taken hold at hyperscale.
The word doing the heavy lifting is merchant. Hyperscalers building their own silicon, including Google’s TPU, Amazon’s Trainium, and Microsoft’s Maia, are not merchant vendors and do not appear in that framing.
NVIDIA disclosed its Q2 revenue of $96.22 billion in an 8-K filing and, in the same release, announced an AWS expansion of 2 million GPUs, even as Amazon promotes its own Trainium accelerators.
Custom silicon is a slower threat than a competing merchant part because it steals the internal workload of one buyer over years rather than the incremental buyer, which is why Vinh’s claim holds for now.
What Would Break This Thesis
The thesis breaks if memory pricing does not normalize and gross margin gets stuck below 72%, because that is the assumption inside the earnings estimate.
It breaks if a hyperscaler’s custom chip reaches the point where it displaces NVIDIA at the margin for inference workloads, since inference is where volume compounds fastest.
It breaks if China Data Center compute does not return and the assumed zero contribution becomes a permanent absence, though management already excludes it from the $108.0 billion Q3 guide.
The most likely point of fracture is the financing structure. NVIDIA has invested nearly $50 billion in frontier AI labs and organized more than $500 billion of third-party capital with Apollo, BlackRock, Blackstone, Brookfield, Goldman Sachs, and KKR.
Huang addressed the framing head-on. “We recognize the scale of this support, and we know some will call this circular financing. We see it differently.” The thesis holds while that capital keeps clearing. It breaks the moment it stops.
Contact [email protected] for any questions or corrections.
Omor Ibne Ehsan is a writer at 24/7 Wall St. He is a self-taught investor with a focus on growth and cyclical stocks that have strong fundamentals, value, and long-term potential. He also has an interest in high-risk, high-reward investments such as cryptocurrencies and penny stocks.
