Nvidia Is Not Priced for Perfection. It's Priced for a Slowdown.

Nvidia is worth about $4.9 trillion, trading places with Apple for the title of the most valuable company in the world. What neither of them shares is how fast Nvidia got there: revenue went from $27.0 billion in FY2023 to $215.9 billion in FY2026, an eightfold rise in three years.

A rise like that attracts a stock phrase, and Nvidia's is "priced for perfection". The idea is that everything has to go right from here, and that at this price there is no room left for disappointment.

It is a phrase people reach for rather than check. So it is worth asking what the price actually assumes about Nvidia's future, and whether that assumption is a demanding one. That is what a discounted cash flow model is for.

What analysts expect Nvidia's revenue to do

Start with the input that matters most. Our model does not guess at Nvidia's revenue, it takes analyst consensus.

Table 1. Analyst consensus revenue estimates for Nvidia, as we hold them.

Fiscal year	Revenue	Growth	Analysts covering
FY2026 (actual)	$215.9B	+65%	-
FY2027	$393.2B	+82.1%	39
FY2028	$561.3B	+42.8%	40
FY2029	$686.6B	+22.3%	26
FY2030	$774.2B	+12.8%	13
FY2031	$1,005.0B	+29.8%	16

Read the growth column. The Street is not forecasting perpetual hypergrowth. It has Nvidia going from 82% growth to under 13% in four years, which is a severe deceleration by any standard, before a late reacceleration in FY2031 that rests on 16 analysts against the 39 and 40 covering the near years. We present the estimates as they stand, but it is worth noticing which years are thinly covered.

Our model uses these figures directly for the first five years, then tapers growth down toward 3.5% over years six to ten. Taken together they work out to a five-year revenue growth rate of about 36% a year. That number is not an assumption we imposed. It is simply the compound rate of a curve that is already bending down hard.

What Nvidia is worth today

Table 2. Nvidia DCF inputs and output, 18 July 2026 model run.

FY2026 revenue	$215.9B
Revenue growth (analyst consensus, 5-year CAGR)	36.0%
EBITDA margin (normalized)	62%
Free cash flow (FY2026)	$96.7B
Discount rate (market-derived WACC)	12.4%
Terminal growth	3.5%
Fair value	$194.84
Price	$202.81
Upside	-3.9%

Start with the bottom three rows. Our model puts Nvidia's intrinsic value at $194.84 against a price of $202.81, so the stock trades about 4% above what we think it is worth.

For a company with this reputation, that is a remarkably ordinary answer. It is not the yawning overvaluation the bubble talk implies, and it is not a bargain either. There is no margin of safety at this price, but there is nothing here that looks like a mania.

What makes it interesting is what produced it. That $194.84 is not the output of a model assuming the AI boom runs forever. It is built on the consensus path in the table above, the one where growth falls from 82% to 13% in four years. The slowdown is already inside the fair value, and the stock still comes out roughly where it trades. Nvidia is not priced for perfection. It is priced for a slowdown.

Which leaves two assumptions doing the work: that revenue follows the consensus path, and that margins hold near 62%. The rest of this piece tests both.

What if the deceleration is steeper?

Nvidia's revenue is, to a first approximation, its customers' capital budget. Microsoft, Alphabet and Meta spent $225.7 billion between them on capex in FY2025, close to Nvidia's entire revenue for the year, and the four largest hyperscalers have collectively guided to roughly $725 billion across 2026. When we valued Microsoft, Meta and Alphabet, we assumed that spending eventually normalizes rather than compounding forever. That single assumption is set out in how we model AI capex, and we also ran it backwards to ask what those three prices already assume. If we are right about them, Nvidia's growth has to come down too.

Which brings us to Michael Burry, whose short thesis we worked through in December. His argument was never about Nvidia's accounting. It was that its customers are misjudging the economics of GPUs, that the chips go obsolete faster than the depreciation schedules assume, and that once the returns disappoint, the hyperscalers rein in spending.

Note what that claim actually is. It is not that Nvidia will slow down, because the Street already says that. It is that the slowdown will be sharper than the Street thinks. Burry has not published a revenue forecast, and we are not putting one in his mouth. His thesis is directional, so the scenarios below are ours, not his.

Table 3. Nvidia intrinsic value across revenue growth scenarios. 18 July 2026 model run, price $202.81.

Scenario	5-year CAGR	Implied FY2031 revenue	Fair value	vs price
Build-out runs hot, Street too cautious	40.0%	$1,163B	$227	+12%
What today's price requires	37.0%	$1,045B	$203	0%
Analyst consensus, our default model	36.0%	$1,005B	$195	-4%
Modestly steeper deceleration	32.0%	$864B	$167	-18%
Meaningfully steeper	28.0%	$739B	$143	-30%
Sharply steeper, the direction Burry argues	25.0%	$655B	$126	-38%

If analysts have it right, Nvidia is worth roughly what it trades for. If they are too cautious and the build-out keeps running hot, there is real upside: at 40% growth the shares are worth about $227, some 12% above today's price. And if they are a little too optimistic, it is worth a good deal less. Four points off the growth rate takes about 14% off the value, and eleven points off takes more than a third.

None of these are disaster scenarios. Even the lowest row still has Nvidia's revenue tripling by FY2031. That is what makes the stock difficult. You do not need anything to go wrong, you only need Nvidia to be a little less spectacular than the Street expects.

What if margins come down?

The second assumption is profitability, and it deserves as much attention as the first. Our model runs Nvidia at a 62% EBITDA margin, which is a normalized figure drawn from several years rather than the current peak, and already sits a few points below the 66.9% it earned in FY2026. It then holds that flat for a decade.

There are two reasons to wonder whether it holds. The first is competition. Even if the hyperscalers spend every dollar they have guided to, they can choose to spend less of it with Nvidia, and all of them are now designing their own AI accelerators, with help from rival chip designers, specifically to reduce what they hand over. The second is that Nvidia's own margins have never been stable. Its gross margin has already slipped about four points from its peak over the past year, during a boom, and in FY2023, the last time demand paused, its EBITDA margin fell by almost half in a single year.

So it is worth asking what Nvidia is worth at the margins it has actually earned before.

Table 4. Nvidia intrinsic value at different EBITDA margins, revenue held at consensus. 18 July 2026 model run, price $202.81.

EBITDA margin	Fair value	vs price
66.9%, its FY2026 level	$204	+1%
62%, our default model	$195	-4%
58.4%, its FY2024 level	$186	-8%
55%	$174	-14%
50%	$157	-23%
42.2%, its FY2022 level	$130	-36%

Margins are the gentler of the two levers, but not by much. Holding today's 66.9% is worth about $9 a share over our default. Slipping back to the 58.4% Nvidia earned in FY2024, which was hardly a bad year, costs about the same again. Returning to FY2022 margins would take a third off the value.

The important caveat is that these two tables are not independent. We have flexed growth and margins one at a time to keep each effect visible, but in the real world they move together: the demand slowdown that pulls revenue below consensus is exactly the environment in which pricing power erodes and margins compress. If both happen at once, the damage is worse than either table shows on its own.

So where does that leave Nvidia

At $202.81, about 14% below its May peak of $236, Nvidia is roughly fairly valued on our model. Not cheap, not obviously expensive, with no margin of safety.

What that fair value rests on is worth being clear about. It does not assume the AI boom lasts forever. It uses a consensus that already has growth falling to 13% by FY2030, and it holds margins a few points below where they are today. Both are reasonable assumptions. Neither is a fact.

Move either one and the answer changes quickly. A few points off the growth rate, or a return to the margins Nvidia earned as recently as FY2024, and the shares look expensive rather than fair. Push both the other way and there is real upside. At this price you are not paying for perfection, but you are paying for the Street being about right on two things at once.

Those are the numbers to argue about, and they are the ones you can change yourself on Nvidia's valuation page. A DCF is one lens, not a verdict. Nvidia is a candidate for your own research, not a recommendation.