NVIDIA’s Falling Multiple Is A Bill, Not A Bargain

NVDAYTD+16.8%SPYYTD+13.7%QQQYTD+17.5%
Analyze NVDA →

Analysts have the earnings growing into the price, so what a holder is really underwriting is the build-out that has to produce them.

A multiple of roughly 32.1 times the last twelve months of adjusted earnings is why NVIDIA (NVDA) reads as expensive at about $218 a share. On the earnings analysts expect for fiscal 2028, a measure not defined identically to the trailing one, that same price is a 47% lower multiple. That fall looks like a discount arriving, but it is closer to a bill: the cheaper multiple is growth being paid for in advance.

Image from Pixabay

Forty-Seven Percent Cheaper, And Still No Gain

That trailing multiple is not on plain GAAP profit but on adjusted, non-GAAP earnings with stock compensation added back, intended to sit nearer the basis the forecasts use, though the two measures are not defined identically. The path runs through about 24.2 times on the fiscal 2027 forecast to about 17.1 times on the fiscal 2028 forecast. That fall is not a payment. Leave the price where it is, and a holder simply owns NVIDIA at 17.1 times its fiscal 2028 earnings, which proves the buyer did not overpay, but delivers no gain. Being paid needs the market still valuing the company above that: at about 20.6 times, roughly halfway between 24.2 and 17.1 times, the shares would be worth about $263, some 21% above today’s price.

Analysts Are Already Underwriting A Slower NVIDIA

None of that assumes an acceleration. Consensus has revenue growing about 49.1% a year across the two years, against 70.7% actually delivered over the trailing twelve months, on a business already past $253 billion of revenue. Consensus also has earnings growing more slowly than revenue over that same two-year span, at about 37% a year, which means analysts are assuming some margin give-back. Management’s own May guidance held the fiscal Q2 2027 gross-margin midpoint where the prior guide had it and raised spending instead, with fiscal 2027 operating expenses set to grow in the upper forties percent year over year.

The Ramp You Are Funding Runs Through A Packaging Line

That spending is what the forecast is made of. NVIDIA’s stated strategy is a new AI infrastructure generation every year, and fiscal 2028 depends on the next one landing: production shipments of Vera Rubin, the generation after Blackwell, begin in fiscal Q3 2027. Behind that sits total supply, including inventory purchase commitments and prepayments, which management increased to $145 billion in fiscal Q1 2027, plus money committed up front to expand Amkor Technology’s U.S. advanced packaging capacity under a multi-year partnership. So the forward multiple is one wager on one capital cycle, the kind of exposure the Trefis High Quality Portfolio does without since it does not rely on the handful of largest technology names to produce its returns.

The Arithmetic Is Fine, The Schedule Is The Risk

Not overpaying, then, provided a two-year consensus is roughly right. What decides the return is not the arithmetic but whether the market keeps paying up as the earnings land, which turns on whether that ramp arrives on schedule. The market has stopped paying up before, and hard: past market shocks, in 2008-2009 and again in 2022, have taken the stock down as much as 84% from peak to trough. Anyone leaning on the arithmetic alone should see how wide that trailing-to-forward gap is across the rest of the market.

What A Stock Is Worth And How Much To Own Are Different Questions

Valuation says what a stock might be worth, it says nothing about how much of your wealth should ride on it. A position that has grown large enough to matter is worth sizing deliberately rather than by accident. What a position that size would do to your net worth is exactly what the Trefis Wealth team computes, with the same rules-based systematic discipline that runs our High-Quality Portfolio. Request a free vulnerability audit of your biggest positions.