Those results use a cash baseline reconstructed from NVIDIA’s latest filings, with an allowance for stock compensation. They are conditional calculations, not company guidance or a price target. Their purpose is to make an investment assumption visible: a different required return, cash baseline or growth duration produces a different answer.
Updated 16 October 2026. Financial results run through 26 July 2026; the price is the previous trading day’s close. This replaces our pre-results discussion of the $91 billion revenue guide.
The latest quarter gives the valuation a stronger business
NVIDIA’s second-quarter fiscal 2027 release reported $96.221 billion of revenue, compared with the $91 billion midpoint it had guided. Data Center contributed about $89 billion. The company then guided to approximately $108 billion of revenue for the following quarter, with a 74% gross margin outlook. The new outlook remains a management forecast; the $96.221 billion is an actual reported result.
That scale explains why an old fiscal-year multiple can be unhelpful. NVIDIA’s business has grown substantially since the year ended in January. Using the $4.90 diluted earnings per share reported for fiscal 2026 alone would put the dated share price at roughly 44.3 times earnings. Reconstructing the trailing twelve months through July gives approximately $7.91 per share and a P/E of 27.5 times. The latter incorporates the two newer quarters.
An investor using only the latest quarter’s $2.46 GAAP earnings per share, multiplied by four, would get about 22.1 times earnings. That is an annualised quarterly snapshot. It assumes the quarter can represent a whole year and gives considerable weight to income outside the operating business. The choice of denominator changes the apparent valuation before anyone makes a forecast.
The earlier $91 billion guide assumed no China Data Center compute revenue. The subsequent filing says Data Center Hopper shipments to China were less than 1% of Data Center revenue in the quarter. That is a limited product-specific disclosure; it does not establish a complete geographic bridge. The actual result now replaces the preview, while export access remains one input to future growth.
Profit and cash tell different parts of the story
The quarter’s $59.688 billion GAAP net income included a significant contribution from equity investments. NVIDIA’s reconciliation removes $7.771 billion of net gains on equity securities and applies the other adjustments and their tax effects, arriving at $53.954 billion of non-GAAP net income, or $2.22 per diluted share. The entire gain cannot simply be subtracted from EPS without considering tax and the other adjustments.
Stock compensation also needs the correct period definition. Beginning in fiscal 2027, NVIDIA includes stock-based compensation in its non-GAAP financial measures. Treating the current adjusted EPS as though it automatically excludes that expense would misread the company’s presentation. At the dated price, annualising the $2.22 adjusted quarterly EPS produces approximately 24.5 times earnings.
Cash conversion presents another consideration. The July quarter generated $24.077 billion of operating cash and $21.341 billion of company-defined free cash flow. The cash-flow statement shows a $22.346 billion use of cash from the increase in receivables and another $5.784 billion from inventories. These are large uses of working capital while the business expands. They help explain why a strong earnings quarter did not produce an equivalent cash inflow.
The Form 10-Q, pages 30–33, connects the receivables increase with extended payment terms on large, multi-quarter agreements involving certain investment-grade customers. It also says those financing arrangements will continue to affect operating cash-flow timing. That disclosure gives investors a specific issue to follow: how quickly reported sales become collected cash as larger agreements mature.
The half-year view is less sensitive to one quarter. NVIDIA generated $69.895 billion of free cash flow in the first six months, against $118.010 billion of net income. A future collection can improve cash conversion without requiring another sale, while continuing growth can keep absorbing cash in new invoices and inventory. Neither outcome can be inferred from the revenue growth rate alone.
Build a consistent cash baseline
For the twelve months ended 26 July, start with the $96.575 billion of free cash flow NVIDIA reported for fiscal 2026, subtract $39.584 billion from the corresponding first half of that year and add the latest $69.895 billion half-year figure. The result is $126.886 billion. This uses the issuer’s definition, which deducts purchases of property, equipment and intangible assets plus related principal payments. A data provider using operating cash less asset purchases alone can report a slightly different number.
The same period substitution gives $7.241 billion of stock-based compensation: $6.386 billion for fiscal 2026, less $3.099 billion for its first half, plus $3.954 billion in the latest half. Operating cash flow adds this non-cash expense back. For the sensitivity exercise, deducting it from free cash flow leaves a more conservative $119.645 billion cash proxy. That is an analytical allowance for compensation paid in equity, rather than another company-reported performance measure.
Dividing by the latest quarter’s 24.285 billion weighted-average diluted shares gives approximately $4.927 per share. The model holds that share count constant. It does not simultaneously impose further dilution or assume that buybacks will increase per-share cash, either of which would require a separate forecast. The denominator is a recent diluted-share proxy, rather than a claim about the exact future number of shares.
On that basis, the dated price represents about 41.6 times trailing company-defined free cash flow per share, or 44.1 times the cash proxy after the compensation allowance. These cash multiples answer a different question from the 27.5-times GAAP earnings multiple. An investor needs to choose which resources the valuation is expected to deliver.
How long must the faster growth last?
The model projects the $4.927 starting cash amount at a constant growth rate for either five or ten years. It discounts each annual amount at the chosen required equity return. From the following year onward, it assumes 3% annual growth indefinitely. The terminal value is the next year’s cash divided by the required return minus 3%, then discounted back to today. Adding that amount to the discounted initial years produces the model value.
Working backwards from $217.23 gives the following required growth rates. At an 8% required return, cash needs to grow 21.4% annually for five years, or 12.6% for ten years. At 10%, the corresponding rates are 31.1% and 17.8%. At 12%, they become 39.3% and 22.2%. Every case switches to the same 3% terminal growth rate after the initial period.
The difference between the two columns is economically substantial. A business that can sustain attractive cash growth for a decade can support a higher price at a lower annual growth rate than one whose exceptional period ends after five years. NVIDIA’s position in accelerated computing, networking and software has to be evaluated partly through that duration question. Being central to today’s AI infrastructure does not settle how long unusually strong per-share cash growth will last.
The 10% input is a chosen return hurdle for the exercise. It is not an estimate of NVIDIA’s cost of capital. Investors accepting a lower return require less future growth to justify the same price; investors requiring more need stronger cash outcomes. The sensitivity makes that trade-off explicit without assigning a universal required return.
Much of the value arrives after year ten
In the ten-year, 10%-return case, approximately 66% of the calculated value comes from the terminal value. That calculation assumes growth has already slowed to 3%; it still represents cash arriving over a very long period. The first decade of projected cash accounts for the remaining third. Confidence in the enduring economics of the business therefore matters even when the near-term revenue outlook is strong.
The model also simplifies capital allocation. It values a future equity cash proxy, rather than building an enterprise-value model and adding a separate cash-and-debt bridge. It assigns no extra liquidation value to NVIDIA’s equity holdings, forecasts no borrowing or acquisitions and deducts no separate programme of strategic investments. Those choices define the exercise’s scope. A complete investment valuation would need to model those assets, obligations and uses of cash consistently.
This limitation has practical significance. In the latest half-year cash-flow statement, purchases of equity securities used $42.404 billion. Such purchases can have strategic or financial value, but the issuer’s free cash flow measure is calculated before that investment spending. The cash proxy should therefore not be read as a promise that its entire amount can be distributed to shareholders every year.
Customer investment has to support the cash assumptions
NVIDIA’s 26 August earnings call describes demand across large cloud providers, AI companies, enterprises and sovereign projects. An investor can accept that broad demand while separately examining the terms under which the systems are sold and financed. The relevant question for the model is whether a larger installed base produces repeatable cash after the resources needed to support its growth.
The 10-Q describes long-term supply and capacity commitments as well as financing arrangements and infrastructure support. These can help customers obtain systems and secure future production. They can also leave NVIDIA exposed if customers’ build-outs, demand or funding develop differently from expectations. Commitments are not all immediately due cash expenses, so adding their total to current debt would overstate the present balance-sheet obligation. Their timing and conditions still matter for future cash.
This is also why NVIDIA’s valuation cannot be copied from a supplier’s margin story. Our Marvell analysis follows how custom silicon changes its product mix. The Dell analysis asks what server growth contributes to operating-profit dollars. NVIDIA’s own cash-growth assumptions need its own customer terms, capital allocation and share count.
Use the next filing to revise specific assumptions
Three developments would change this exercise in useful, measurable ways:
- Receivables convert into cash. Compare subsequent operating cash flow with the receivables balance, new sales and the company’s explanation of payment terms. Collection of earlier invoices can raise the cash baseline; continuing extensions can postpone that improvement.
- Growth requires a different amount of capital. Follow inventories, asset purchases, principal payments and the cash cost of strategic commitments. Higher revenue supports the valuation more strongly when it produces cash after those needs.
- Capital allocation changes the per-share result. Compare repurchases with stock compensation and the diluted share count, then include strategic investments separately. The amount spent on buybacks does not itself establish growth in each remaining share’s cash claim.
Our Costco capital-allocation analysis uses the same discipline of connecting a potential shareholder payout with the resources supporting it, although Costco’s business and growth assumptions are very different.
To see the sensitivity within one consistent set of assumptions, hold the required return at 10%, the faster-growth period at ten years and terminal growth at 3%. A 15% annual increase in the cash proxy gives approximately $176.50 per share; 20% gives approximately $255.06. Both use the same starting filings. The investment judgment is whether the business can deliver the assumed cash path, including the spending and customer financing needed to sustain it.
