Nvidia Stock After Q2 FY2027: $96 Billion of Revenue—and the Working-Capital Shadow

There is a moment in every gold rush when the sound changes. At first, one hears picks striking stone, then the shouts of prospectors, then the clatter of entire towns being built around the promise of buried wealth. Nvidia’s latest quarter sounds like all three at once. Revenue reached $96.2 billion. Data Center revenue alone was $89.0 billion. The company is guiding to roughly $108 billion for the next quarter—without assuming any China data-center compute revenue. These are figures of such scale and speed that the ordinary vocabulary of semiconductor analysis begins to feel too small.

And yet the detail that kept me at the table after the headline numbers had faded was quieter: accounts receivable rose to $63.1 billion from $38.5 billion at the end of January, while inventory climbed to $31.6 billion from $21.4 billion. Neither number proves that demand is weak. In a business doubling at this scale, working capital must expand. But they remind me that Nvidia is no longer merely selling chips into a booming market. It is becoming part of the financial architecture that allows the market to keep booming.

That distinction is the center of my Nvidia stock analysis. The operating achievement is extraordinary and, on the evidence available today, real. The valuation question is also real. At about $5.09 trillion in market value, investors are not being asked whether Nvidia will remain a superb company. They are being asked whether an unprecedented concentration of capital expenditure, supplier credit, customer financing and infrastructure partnerships can produce enough durable cash earnings to justify one of the largest equity values ever assembled.

The quarter in one view

The slope of that chart is why the stock cannot be understood through a conventional “expensive versus cheap” argument. A static multiple applied to a business whose quarterly revenue has risen from $46.7 billion to $96.2 billion in one year will almost always look backward. But extrapolating the slope indefinitely is equally dangerous. Nvidia’s numbers require two thoughts to coexist: the present acceleration is historic, and no exponential curve is immortal.

What is genuinely excellent

The cleanest part of the story is the income statement above the line marked “other income.” Data Center revenue grew 117% year over year, and total revenue grew 106%, while operating income grew 124%. That is operating leverage in its purest form: each new dollar of sales is carrying more profit than the average dollar before it. A 75% GAAP gross margin in a hardware-centered business also tells us that Nvidia is not being paid merely for silicon. It is being paid for an integrated system—accelerators, networking, software libraries, developer tools and the accumulated habit of an ecosystem built around CUDA.

This is why a simple chip-cycle analogy misses something important. Traditional semiconductor cycles often end when capacity catches demand and products become interchangeable. Nvidia’s defense is not the physical chip alone. It is the cost and risk of moving a production AI workload away from the full platform. Competitors can offer attractive hardware and still confront the less visible moat of software, trained engineers, optimized models and operational familiarity.

That does not make the moat permanent. It makes it costly to cross. I see that distinction as more useful than declaring Nvidia either invincible or doomed. The company’s competition with AMD will be decided partly by performance per dollar and energy efficiency, but also by whether customers can genuinely multi-source without sacrificing developer velocity. Nvidia’s ability to move Rubin into full production is therefore not just a product milestone. It is an attempt to keep the software and networking ecosystem moving quickly enough that the destination itself never stands still.

The Q3 outlook reinforces the demand signal. Management expects $108 billion of revenue, plus or minus 2%, while assuming no data-center compute revenue from China. That removes one easy explanation for the forecast and highlights the strength elsewhere. Edge revenue of $7.2 billion, up 27%, also suggests that inference is beginning to travel beyond the largest training clusters. The AI economy is not yet broad enough to remove concentration risk, but it is becoming wider.

The cash-flow sentence deserves its own paragraph

Nvidia reported $59.7 billion of GAAP net income and $24.1 billion of operating cash flow in the quarter. A gap of that size naturally attracts suspicion, but suspicion is not analysis. The quarter included $7.8 billion of net gains on equity securities, which lifted accounting profit without creating the same amount of operating cash. Working-capital movements also matter: receivables and inventory expanded sharply as the business scaled.

For the first six months of the fiscal year, operating cash flow was $74.4 billion, which gives a less distorted view than one quarter alone. I do not read Q2 cash conversion as a red alarm. I read it as the place where investors should keep their eyes. Receivables have grown about 64% since January and inventory about 48%. Revenue growth can rationally require both, especially when complex systems are assembled and shipped on a massive schedule. Still, the balance sheet is where exuberant demand eventually reveals whether it was paid for promptly, financed generously or ordered ahead of need.

The distinction matters because Nvidia’s ecosystem is increasingly linked to capital providers, cloud companies and AI infrastructure specialists. Management said it is working with capital partners to mobilize more than $500 billion over time, subject to agreements. That may be an elegant solution to a genuine bottleneck: customers need vast sums before they can turn computing capacity into revenue. It also makes Nvidia’s demand environment more reflexive. Financing enables data centers; data centers buy Nvidia systems; the systems support financing narratives and asset values; those values can enable still more investment.

Reflexivity is not fraud, and it is not automatically a bubble. Railways, electricity grids and mobile networks were all built through financial systems that pulled future utility into present construction. The analytical task is to ask who ultimately bears the utilization risk. The balance sheet of a hyperscaler is different from that of a highly leveraged specialist. Our CoreWeave analysis explores the leveraged end of that spectrum. For Nvidia shareholders, the key question is whether credit support and capital partnerships broaden a healthy market or postpone the moment when marginal projects must prove their economics.

China is absent from guidance but present in the risk

The Q3 forecast assumes no China data-center compute revenue. That is conservative in one narrow sense: any permitted shipments would be upside to the stated assumption. Strategically, however, zero is not a victory. China represents customers, developer mindshare and a large domestic incentive to build alternatives. Export controls can protect advanced technology in the short term while accelerating local substitution in the long term. Investors should resist turning a political constraint into a simple quarterly modeling adjustment.

The deeper risk is that the AI hardware market becomes geographically and technologically fragmented. Nvidia could remain dominant in the United States and allied markets while losing participation in a parallel ecosystem. At the same time, major customers are designing more custom silicon. These efforts need not “replace Nvidia” to matter. If they absorb stable inference workloads or provide credible bargaining leverage, they can influence mix, pricing and the duration of 75% gross margins.

That is why I watch gross margin more closely than unit market share. Nvidia can cede low-value volume and remain economically dominant. Conversely, it can preserve shipment share while the profit pool begins to migrate toward customers, networking alternatives or application owners. The current evidence—75% GAAP gross margin and accelerating operating income—still favors Nvidia. The burden of proof lies with the challengers. But the burden embedded in the stock price lies with Nvidia.

A $5 trillion valuation changes the question

At an intraday share price around $209.66 on August 27, Nvidia’s market capitalization was approximately $5.09 trillion. The quoted trailing price-to-earnings ratio was about 26.3, based on trailing earnings per share near $7.97. That multiple is not extreme relative to the current growth rate. The denominator, however, contains a business operating in an unusually favorable phase and quarterly investment gains that should not be capitalized like recurring chip profit.

Annualizing Q2 revenue produces roughly $385 billion. The market capitalization is therefore about 13.2 times that simple run rate. This is not a forecast and ignores the Q3 step-up, but it usefully shows what the market already assumes: Nvidia must convert extraordinary revenue into extraordinary profit for years, not quarters.

I prefer a scenario frame to a single price target. The table below is deliberately transparent. It asks what Nvidia might be worth in three years under different combinations of sustainable annual earnings and terminal earnings multiples. The inputs are analytical assumptions, not company guidance.

Even the bear case requires annual earnings far above the current quarterly run rate once unusual investment gains are stripped out. That is the subtlety. A “bear” outcome for the business could still involve enormous profit growth; it would merely be insufficient relative to the expectations already capitalized. The base and bull cases demand not only continued AI spending but sustained high margins, successful product transitions and enough customer economics to keep the infrastructure cycle funded.

There is also a reverse valuation lesson. If Nvidia can genuinely earn $300 billion annually within three years and retain a high-twenties multiple, today’s valuation can still produce attractive returns. The stock is not mathematically doomed by its size. Size simply removes forgiveness. A delayed architecture, lower utilization at customers, tighter credit conditions or a modest margin reset can subtract hundreds of billions of dollars from reasonable value.

Capital returns are large—and still secondary

Nvidia returned $26.0 billion to shareholders through repurchases and dividends during the first half and ended the quarter with $99 billion remaining under its repurchase authorization. In almost any other company, that program would dominate the capital-allocation discussion. Here it sits behind the scale of operating profit and the market value of the equity.

I judge the buyback by two tests. First, does Nvidia retain enough flexibility to protect supply, accelerate software development and support strategically useful ecosystem investments? The present balance sheet suggests yes: cash, marketable debt securities and marketable equity securities together substantially exceed conventional debt. Second, is the company buying below a conservative estimate of intrinsic value? At more than $5 trillion, that answer depends on the same long-duration assumptions investors are debating. A repurchase can lift per-share results while still destroy value if executed above intrinsic worth.

The dividend remains symbolic at $0.25 per share each quarter relative to the share price. That is appropriate. Nvidia’s highest-return opportunities are still likely inside its platform and supply ecosystem. I would prefer disciplined investment and selective repurchases to a larger payout designed merely to signal maturity. The danger is not that Nvidia lacks cash. It is that abundance can lower the hurdle rate for projects intended to sustain demand around the company.

This is another reason to read ecosystem investments and customer financing terms carefully. Capital allocation now extends beyond factories and research. It includes the health of the buyers, infrastructure developers and software companies that make Nvidia’s addressable market real.

The strongest bull and bear arguments

The bull case

• AI infrastructure is becoming a new computing layer. If accelerated computing replaces a meaningful share of general-purpose workloads while new agentic and physical-AI workloads emerge, the addressable market can continue expanding.
• The platform is deeper than the accelerator. Networking, systems, software and developer lock-in allow Nvidia to capture more value and defend margins.
• Execution is exceptional. Product cadence, supply coordination and gross margin show a company converting technical leadership into financial results at unusual speed.
• China is excluded from guidance. Any lawful reopening could add revenue, while current growth demonstrates demand elsewhere.
• Balance-sheet capacity remains formidable. Nvidia can invest, repurchase stock and support ecosystem development without depending on fragile external financing for its own operations.

The bear case

• Customer economics remain incompletely proven. Infrastructure revenue is immediate for Nvidia; monetization may arrive more slowly for the buyers.
• Capital is becoming part of the sales system. Partnerships can expand supply but may blur the boundary between organic demand and financially enabled demand.
• Working capital is rising quickly. Receivables and inventory are rational areas to monitor as order sizes and system complexity grow.
• Custom silicon and rival platforms can pressure the profit pool. They need not defeat CUDA outright to reduce Nvidia’s pricing power at the margin.
• The valuation assumes sustained excellence. At $5 trillion, even a healthy normalization can produce disappointing shareholder returns.

What I will watch next

1. Receivables relative to revenue. I want to see collections grow with shipments rather than a persistent lengthening of the cash cycle.
2. Inventory and purchase commitments. Growth is expected ahead of Rubin, but the mix and pace should remain aligned with sell-through.
3. Operating cash flow over several quarters. The six-month figure is strong; a multi-quarter view separates timing effects from structural change.
4. Gross margin through the product transition. Holding the mid-70s would confirm that system value remains with Nvidia.
5. Customer utilization and monetization. Cloud disclosures, AI service revenue and enterprise deployment evidence matter more than model demonstrations.
6. The terms of capital partnerships. Investors should distinguish introductions and capacity finance from guarantees, take-or-pay commitments or balance-sheet exposure.

My conclusion

I came away from Nvidia’s quarter more impressed by the company and more demanding of the stock. Those positions are not contradictory. Revenue of $96.2 billion, 75% gross margin and a $108 billion outlook describe an operating machine without a modern peer. The rise in receivables and inventory, the gap between quarterly net income and operating cash flow, and the growing role of capital partnerships describe the expanding shadow cast by that machine.

The shadow is not evidence that the light is false. It is evidence that the object has become very large.

For a new investor, I would not reduce the decision to whether AI is “a bubble.” The better questions are narrower: Can Nvidia preserve the economic value of its platform as customers develop alternatives? Can the buyers of its systems earn adequate returns? Can cash conversion keep pace with accounting profit? And can all of that happen long enough to support a $5.09 trillion starting valuation?

Today, the operating evidence says yes more loudly than it says no. The valuation says that a quiet yes will not be enough.

Sources and methodology

• Nvidia Q2 FY2027 earnings release, August 26, 2026.
• Nvidia Form 10-Q for the quarter ended July 26, 2026.
• Market price, market capitalization, trailing EPS and P/E are intraday market-data snapshots as of August 27, 2026. Ratios and scenarios are The Kapital calculations and may differ from vendor figures because of timing and share-count conventions.

This article is independent analysis, not investment advice. Securities can lose value, and forward-looking scenarios are uncertain.