Nvidia’s $91B Test on August 26: AI Stocks at Risk
Nvidia can post $92 billion in sales this week and still cop a hiding. That is what happens when investors stop pricing a business and start pricing a miracle.
Nvidia can post $92 billion in sales this week and still cop a hiding. That is what happens when investors stop pricing a business and start pricing a miracle.
On Wednesday, August 26, Nvidia reports its fiscal second-quarter results after the US market closes. In May, the company guided investors to roughly $91 billion in quarterly revenue, plus or minus 2%. Analysts now expect about $92.2 billion. That sounds like a comfortable setup until you understand the problem: the market is not asking whether Nvidia is excellent. It is asking whether excellence can keep accelerating forever. ([nvidianews.nvidia.com](https://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-first-quarter-fiscal-2027?utm_source=openai))
The $91 billion number is not the target. It is the trap.
Nvidia’s last quarter was absurdly good. Revenue hit $81.6 billion, up 20% on the previous quarter and 85% on the year before. Data Center revenue was $75.2 billion. The company produced $58.3 billion in net income and expanded its quarterly dividend from one cent to 25 cents a share. It also added $80 billion to its share-repurchase authority.
Those are not the figures of a company struggling to find buyers. They are the figures of a company sitting in the middle of the biggest capital-spending race since the internet build-out. ([investor.nvidia.com](https://investor.nvidia.com/news/press-release-details/2026/NVIDIA-Announces-Financial-Results-for-First-Quarter-Fiscal-2027/default.aspx?utm_source=openai))
But markets are funny beasts. A business growing at 85% can disappoint if investors bought the shares expecting 100%. A company can beat the published estimate and still fall because the serious money was quietly expecting more than the estimate.
That is the risk hanging over Nvidia this week.
The published consensus is around $92.2 billion in revenue. Nvidia itself told the market to expect $91 billion, plus or minus 2%. So a result within the guided range may be operationally fine but psychologically weak. If the company delivers $91.5 billion, it has done exactly what it said it would do. Yet some investors may treat that as a problem because they have become accustomed to Nvidia making impossible growth look routine.
That is not analysis. That is entitlement dressed up as a spreadsheet.
Wall Street has turned one earnings call into a referendum on AI
Reuters reported that Asian shares slipped on August 25 as investors got twitchy ahead of Nvidia’s result. The nervousness was not limited to one stock. Semiconductor names across the region were under pressure, with investors also weighing Alibaba’s discounted $10.2 billion share sale to fund its AI ambitions and Samsung’s shareholder-return plans. ([ca.investing.com](https://ca.investing.com/news/economy-news/asia-shares-fall-on-tech-nerves-oil-prices-slip-4814829?utm_source=openai))
This is the real story: Nvidia is no longer merely Nvidia.
It has become the quarterly lie detector for the entire AI trade.
If Jensen Huang says demand remains strong, supply is tight and customers are still ordering more compute, every data-centre landlord, power supplier, memory maker, cloud provider and AI software hopeful gets a confidence boost. If the language gets even slightly cautious, the market will not calmly assess the nuance over a nice cup of tea. It will sell first and ask questions later.
That is because a ridiculous amount of capital has been committed on the assumption that AI demand will justify an equally ridiculous amount of infrastructure. Chips are only the first invoice. After the chips come data centres, electricity, cooling, networking, debt, leases, software teams and years of operating costs.
Nvidia is the cleanest early winner because it sells the picks and shovels. Its customers carry much more of the execution risk.
The overlooked issue is not demand. It is who is paying for it.
Most commentary will focus on whether Nvidia can sell more Blackwell systems, how quickly its next-generation Rubin products ramp and whether China remains a drag on the numbers. Those matter.
But the better question is simpler: are Nvidia’s customers earning enough from AI to justify this spending?
Nvidia’s numbers prove that hyperscalers, cloud firms and sovereign projects are spending. They do not prove that every dollar being spent will generate an attractive return.
That distinction matters more than people think.
A company can buy truckloads of GPUs because it fears being left behind. A chief executive can approve a giant AI budget because every competitor is doing the same thing. A board can call that “strategic investment” and nobody gets fired immediately.
I have seen versions of this movie in business. When the market tells everyone that a new category is inevitable, people stop demanding normal commercial answers. They stop asking: What does the customer pay? What is the gross margin? How long until payback? What happens if demand is merely good rather than biblical?
That is how good ideas become bad investments.
Nvidia is not the obvious weak link here. In fact, its reported first-quarter gross margin of 74.9% tells you exactly why it has become the tollbooth on this spending boom. The potential weak links are further down the chain: companies funding huge compute fleets before they have proven recurring revenue that can carry the cost. ([nvidianews.nvidia.com](https://nvidianews.nvidia.com/_gallery/download_pdf/6a0e17dc3d633295d45282e6/?utm_source=openai))
A beat could be bearish if the guidance is merely normal
Here is the contrarian bit: the most dangerous outcome for the stock may not be a miss. It may be a perfectly strong quarter paired with guidance that is only solid.
Markets can absorb bad news when expectations are low. What they struggle with is a tiny crack in a flawless narrative.
Nvidia’s prior guidance implied revenue growth from $81.6 billion to about $91 billion in one quarter. S&P Global’s Visible Alpha consensus now sits slightly above that at $92.2 billion. The gap is not enormous in percentage terms, but at this scale it represents billions of dollars in investor expectation. ([spglobal.com](https://www.spglobal.com/market-intelligence/en/news-insights/research/2026/08/nvidia-earnings-preview-q2-2027?utm_source=openai))
Suppose Nvidia reports $92 billion or $93 billion. Great result. Then suppose its forward guide suggests the growth rate is settling into something more normal as capacity catches up, customers digest prior purchases or product transitions create a pause. The company would still be growing at a rate most businesses would tattoo on their forehead.
The share price could still get belted.
Why? Because the market has already moved beyond asking whether Nvidia is a great company. It is debating the duration of greatness. That is a much harder thing to value.
I would rather own a good business at a price that allows for a few ordinary quarters than a brilliant business priced as though ordinary quarters have been abolished by technology.
What matters on Wednesday night
Forget the breathless headlines about an earnings “beat” or “miss.” They are often useless within five minutes. I would watch four things instead.
First, revenue versus the $91 billion guide. Did Nvidia clear its own number decisively, or merely land around it?
Second, the next-quarter revenue outlook. This matters more than the reported result. Markets buy the future, not the press release.
Third, product-transition commentary. Investors will want clarity on the Blackwell ramp and the timing of Rubin. Big technology transitions can produce demand pull-forwards, delivery bottlenecks or short pauses that get wildly overinterpreted.
Fourth, customer economics and financing. Listen for any indication that Nvidia is taking on more commercial risk to help customers buy systems, or that demand is broadening beyond a small group of very large buyers. Strong demand is good. Demand that requires increasingly creative financial engineering is a different animal.
Goldman Sachs analysts have flagged investor interest in Nvidia’s financing platform, its potential capital commitments and the product roadmap, precisely because the market is trying to work out how much of the AI build-out is organic demand and how much needs assistance to keep moving. ([finance.yahoo.com](https://finance.yahoo.com/markets/stocks/articles/four-key-things-investors-watch-110302048.html?utm_source=openai))
The boring truth investors need to hear
Nvidia may keep winning for years. The company’s May results and guidance make that completely plausible. It has extraordinary products, extraordinary margins and customers with enormous budgets. Calling the business a bubble just because the stock has become important would be lazy.
But calling the valuation risk-free because the business is extraordinary would be even lazier.
The hard part of investing is accepting that a wonderful company and a wonderful stock are not always the same thing on the same day. Nvidia’s Wednesday result may confirm that AI infrastructure demand remains ferocious. It may also remind investors that when the whole market expects perfection, “very good” gets punished.
Both can be true. Usually, they are.
What this means for you
If you are an investor, do not make a major portfolio decision because you think you can outguess one Nvidia earnings call. That is not investing; it is punting on a coin toss with a Bloomberg terminal attached.
Know your position size before the result. If a 10% or 15% overnight move would make you panic, your position is too large. Fix that before the bell, not after it.
If you run a company, steal the useful lesson from Nvidia without copying the madness around it: make yourself essential to a large spending trend, but stay savage about customer economics. Revenue is lovely. Revenue from customers making no money is eventually somebody else’s problem — and then it becomes yours.
And if you are building in AI, do not confuse buying infrastructure with building an advantage. Every competitor can order hardware. Far fewer can create a product customers happily pay for, retain them for, and improve fast enough to stay ahead.
That is the game after the GPU invoice lands.