The $490 Nvidia Math Is a Multiple Bet, Not a Growth Bet

On August 26, Nvidia did something it has never done before: it issued a forecast for the year ahead. CFO Colette Kress told investors the company expects revenue to grow about 70% in fiscal 2028 — the year that ends in January 2028 — a far bigger jump than the roughly 44% growth the average analyst had been modeling. It is the first time management has spoken that far into the future, and it is the reason the stock is suddenly circulating as a "$490 by the end of 2027" story.

That number deserves a fair hearing before it gets any skepticism. The growth is real. In the fiscal second quarter, revenue came in at $96.2 billion, up 106% from a year earlier, with data center revenue up 117% and gross margin at a gaudy 75%. Non-GAAP earnings were $2.22 per share. This is about as good as business economics get, and the $490 talk attaches to a company that keeps delivering on what it promises.

The number that changed the argument

The old story was that AI chip demand was enormous but nobody could see it clearly enough to trust it. Nvidia just spent a decade's worth of credibility on making that visible: management's first-ever year-ahead outlook implies fiscal 2028 revenue of roughly $670 billion to $700 billion, based on that roughly $396 billion consensus for the current year. The stock jumped about 8.7% on the news.

That "expectation reset" cuts the opposite way from the ones I usually look for. My normal hunt is a beaten-down name where the market still prices the old risk profile while the numbers quietly improve. Nvidia is the mirror image: the market already prizes the improving story to the hilt, and management just raised the bar to a level Wall Street hadn't reached. The good news is no longer a secret waiting to be priced — it is a guidance number everyone already read.

Saying "$490" is saying "double the multiple"

A price target is rarely a mystery. It is expected future earnings multiplied by a price-to-earnings ratio, and the $490 headline resolves into exactly that. Consensus earnings for the year ending January 2028 sit near $15.50 per share. Divide $490 by that and you get roughly 32 times.

Now hold that against today's price of about $218. At the very same earnings number, the stock trades at roughly 14 times. The $490 target is not primarily a bet on whether Nvidia hits its growth forecast — that forecast is already guided, and much of it is already in the price. The target is a bet that the market roughly doubles the multiple it is willing to pay for those guided earnings, to about 32 times, in roughly 15 months, while a company sitting at roughly $5.3 trillion of market value also happens to deliver the growth. Both halves have to cooperate.

It is worth emphasizing what the stock is not: a broken story being rescued by cash flow. Nvidia's hard proof is real — on a trailing basis the data I can pull shows on the order of $127 billion of free cash flow, a roughly 47% free-cash-flow margin, and a net-cash balance sheet. The trouble is that the price already pays for all of it, close to 40 times trailing free cash flow. A beginner should see the distinction: a fantastic business and a cheap price are not the same thing, and here only the first is on offer.

The proof, and the tripwire

The strongest bear argument is not that demand is fake. It is that the company's own outlook is deliberately conservative in ways that hide how much has to go right. The 70% growth figure is supply-constrained — management says it would be "a lot higher" were memory components more available — and it excludes China data center revenue entirely. Meanwhile the customer base is concentrated: hyperscalers drove roughly $49 billion of the quarter's data center revenue alone. A handful of the world's biggest spenders are the engine.

So name the condition that would break the case. If that 70% guidance starts being walked back toward the 44% consensus, if the memory shortage persists long enough to choke the ramp, or if one hyperscaler talks about pausing AI spending, the pain lands on the multiple — the roughly 32 times the $490 already assumes. That is the number to watch first, because it is the part of the story that is genuinely unresolved.

I can be wrong. Memory supply could loosen, demand could stay this hot, and the whole thing could keep compounding for another year. That is precisely why I would not bet against it. But turning "$490 by the end of 2027" into earnings times a multiple shows that the work is being done by the multiple holding near its high, not by a discount being discovered. You would be paying full freight for a near-certain outcome — and with a company this visible, certainty is exactly what has already been priced.