Company spotlight
Is Nvidia stock cheap or expensive?
“Cheap or expensive” is the wrong question to ask in isolation, and Nvidia is the clearest example of why. At about 45 times earnings it is plainly not a value stock. At a 0.7 PEG against 65% growth, it is arguably underpriced. Both are true. So instead of arguing, we pulled the actual filings and the actual price, put the numbers on a chart, and worked out exactly what you are paying for — and at what prices the math gets friendlier.
The short version
- Expensive on absolute multiples: ~45x trailing earnings, ~25x sales, a ~1.8% free cash flow yield. Nothing about that is “cheap.”
- Reasonable on growth: revenue grew ~20x in six years and ~65% in the last year alone, putting the PEG near 0.7.
- The whole bet is durability. Pay 45x and you are underwriting years more of AI-infrastructure demand — the exact thing the bubble skeptics question.
- The “entry levels” below are valuation arithmetic, not buy signals. This is analysis, not advice.
The business, in three numbers
Nvidia (NVDA) designs the GPUs that train and run modern AI. The financial story is one of the most extreme in public-market history: in fiscal 2026 (the year ended January 2026) it reported about $215.9B in revenue, up roughly 65% year over year, on which it kept about $120.1B in net income — a net margin north of 55%, with operating margins above 60%. Free cash flow was roughly $96.7B. For context, six years earlier the whole company did about $10.9B in sales.
That FY23 dip in net income is real — an inventory and crypto-hangover year — and worth remembering: this is a cyclical chip business, not a smooth annuity. But the trajectory since is the steepest in megacap history.
So is it cheap or expensive?
At a price of $222.82 (June 2, 2026), Nvidia is worth roughly $5.45 trillion. Run that against the fiscal-2026 results and you get a snapshot that is genuinely two-faced.
The expensive case is simple. A 45x earnings multiple and a sub-2% free cash flow yield mean the market has already priced in years of strong results. If growth merely slows to something normal, the multiple has a long way to fall. A buyer at these levels has almost no margin of safety in the traditional sense — you are not getting paid to wait.
The reasonable case is just as real. Valuation is always relative to growth, and a 45x multiple on a business compounding earnings at ~65% is not demanding — the PEG near 0.7 says so. By the standard growth-investor yardstick, that is cheap, not expensive. The catch is the word “compounding.” A PEG is only meaningful if the growth rate in the denominator is durable.
The real question isn’t the multiple, it’s the durability. At 45x you are not betting on whether Nvidia is a great company — clearly it is. You are betting that AI-infrastructure spending keeps compounding for years. That is the same demand chain the circular-financing skeptics are poking at. Your view on Nvidia’s valuation is really a view on that.
The price, for context
The stock is up more than 13x over five years, split-adjusted, and round-tripped a brutal 2022 drawdown along the way. It is currently near the high end of its 52-week range (about $137 to $236).
What you pay at each multiple
Here is the part people actually want: at what price does the valuation get more comfortable? The honest answer is that it depends on the multiple you are willing to pay. So rather than invent a target, here is the arithmetic — the price implied by a range of P/E multiples on the trailing FY2026 EPS of $4.90.
| P/E multiple | Implied price | vs. $222.82 today | What it would mean |
|---|---|---|---|
| 25x | $123 | -45% | Priced like a maturing, slowing chipmaker |
| 30x | $147 | -34% | Premium, but growth clearly fading |
| 35x | $172 | -23% | Strong growth, de-risked a notch |
| 40x | $196 | -12% | Roughly the market’s current comfort zone |
| 45x | $221 | ~0% | About where it trades today |
| 50x | $245 | +10% | Re-acceleration priced in |
These are not buy targets. They are valuation arithmetic on trailing earnings, and they deliberately ignore future growth — if earnings climb another 40%, a “35x” price moves up with them. The point of a ladder like this is to pre-decide what multiple you would find reasonable, so the choice is made with a calm head and not on a green or red afternoon. Nothing here is a recommendation to buy or sell.
The bull and the bear, stated plainly
Bull: Nvidia owns the picks-and-shovels of the defining technology shift of the decade, with 60%-plus operating margins, a software moat (CUDA) that locks in developers, and a customer base racing to spend. At 45x earnings growing 65%, you are paying a fair price for a generational compounder, and the multiple can hold if growth does.
Bear: This is a cyclical hardware company at a cyclical peak, selling chips that depreciate fast into a buildout that may be partly circular-financed. Margins this high invite competition and customer in-sourcing. If demand normalizes, both earnings and the multiple compress at once — the dreaded double-whammy — and there is no FCF yield to cushion the fall.
Notice that neither case argues about whether Nvidia is a good company. They argue about durability and price. That is almost always where the real disagreement lives.
How we built this (and how you can)
Every figure here came from primary sources in about a minute, not a spreadsheet you maintain by hand: the fundamentals are straight from Nvidia’s 10-K filings, the price is the split-adjusted daily series, and the multiples are simple arithmetic on top. With a markets database connected to an AI agent over MCP, the whole workup is a conversation:
“Pull Nvidia’s last six years of revenue, net income, operating cash flow and capex from the filings. Get the current split-adjusted price and shares. Compute market cap, P/E, P/S and FCF yield, and show me the price implied by 25x to 50x trailing earnings.”
The same recipe works for any large-cap. Run it on the names you actually own, and you replace “it feels expensive” with a number and a level. For the systematic version — ranking a whole universe on cheapness and quality at once — see the Magic Formula screen; to check whether the smart money agrees, the 13F view.
This is analysis, not financial advice. Figures reflect filings and prices as of the dates shown and use trailing, not forward, numbers. Valuation is a judgment, not a fact; do your own work and mind your own risk tolerance before acting.