Nvidia (NVDA) Earnings Aug 26: The $92.18B Test for Your Retirement
Nvidia (NVDA) reports on August 26, 2026, with $92.18 billion in expected quarterly revenue. That still doesn’t make it safe at any price.
Nvidia is expected to pull in $92.18 billion in quarterly revenue, nearly double a year earlier — and that still may not be enough to keep the market happy.
That’s not a typo. The company at the centre of the AI boom reports after the US market closes on August 26, 2026, and plenty of investors are treating the result like a referendum on whether their portfolio gets to keep going up. That’s daft.
I’m not saying Nvidia is a bad business. Quite the opposite. It is one of the most extraordinary businesses ever built. I’m saying the more everyone agrees something is wonderful, the less room there is for even a minor disappointment. Wealth is lost when people forget that distinction.
The $92.18 billion number is the market’s problem
Analysts surveyed by LSEG expect Nvidia’s fiscal second-quarter revenue to reach $92.18 billion, nearly twice the result from the same period a year ago. Nvidia itself guided for $91 billion, plus or minus 2%, when it reported its previous quarter.
For perspective, Nvidia generated $81.6 billion of revenue in the quarter ended April 26, 2026. Its data-centre business alone produced $75.2 billion, up 92% year-on-year. The company’s gross margin was roughly 75% — a frankly outrageous number for a business that sells physical computing hardware.
That is the good news. It is also why the hurdle is stupidly high.
When a business grows at 85% year-on-year, investors stop asking whether it is growing. They ask whether growth is accelerating, whether margins will hold, whether the next product is arriving on schedule, whether management has found a new way to spend money, and whether every customer signing a giant order can actually pay.
Nvidia has become less a stock than the market’s AI confidence meter. A strong result could soothe nerves across chips, data centres, cloud infrastructure and the broader growth-stock trade. A merely good result could still be punished if investors decide the best part of the boom is behind it.
That is not rational in the ordinary-person sense. But it is exactly how markets work when expectations are inflated.
The real story is not chips. It is who is financing the party.
The clean version of the Nvidia narrative is simple: Big Tech needs enormous amounts of computing power for AI, Nvidia sells the most sought-after kit, and everybody gets rich.
The messier version matters more.
Big Tech’s data-centre spending is forecast to exceed $730 billion this year. Nvidia has also helped arrange financing platforms targeting more than $500 billion for customers building AI infrastructure. Last week, it agreed to guarantee up to $105 billion to support OpenAI leasing an Ohio data centre over 20 years.
Read that again: the company selling the shovels is increasingly involved in helping finance the gold rush.
That does not automatically mean something dodgy is happening. Nvidia chief executive Jensen Huang has argued that the company is using a strong balance sheet to help fund data centres, power and facilities expected to house Nvidia systems for decades. There is a commercial logic to that. If you have cash, a dominant product and customers growing faster than their financing capacity, helping unlock projects can be a very good move.
But it changes the risk calculation.
A normal supplier sells a product, collects the money and lets the customer worry about whether the customer’s business model works. When the supplier becomes more entwined with the financing of its customers’ expansion, the supplier has more exposure to whether the whole ecosystem produces real cash flows.
That is the question investors should be asking tonight: not just “Did Nvidia beat estimates?” but “How much of this demand is independently durable?”
There is a big difference between customers buying computing capacity because they are earning attractive returns from AI, and customers buying it because capital markets are still willing to fund the next data-centre announcement.
Vera Rubin is the next proof point
Nvidia is transitioning customers from Blackwell chips to its next-generation Vera Rubin platform, with shipments expected to begin this autumn. That handover is important because the market wants evidence that the company can keep resetting demand before customers have time to pause and ask whether they have already bought enough gear.
Morgan Stanley analysts estimate the new chips could contribute nearly $9 billion of third-quarter sales. Analysts expect Nvidia to forecast third-quarter revenue of $104.20 billion, an 82.8% year-on-year increase, while adjusted gross margin is expected to stay around 75%.
Again: ridiculous numbers. Also, numbers that explain why a beat alone may not do the job.
Nvidia’s shares had risen 11.8% so far in 2026 as of Reuters’ August 25 report, but the stock had lagged some major rivals and briefly ceded its position as the world’s most valuable company to Apple last month. Bloomberg reported the shares were heading into earnings after a seven-session losing streak, their longest since 2022.
That tells you the market is already uneasy. Investors are worried about rising memory costs, the pace of the Rubin rollout, competition from AMD, Intel and custom chips built by the cloud giants, and whether AI spending creates profits for buyers rather than simply more spending.
This is where people make the amateur mistake: they see a temporary share-price wobble and assume it means the risk has disappeared. It hasn’t. A lower share price is not the same thing as a cheap asset. Price only means something relative to the cash flows that turn up later.
The overlooked angle: Nvidia may be brilliant and your portfolio may still be lazy
Here’s the uncomfortable bit. Lots of investors saying they own “a diversified portfolio” actually own the same AI bet in six different wrappers.
They own Nvidia directly. They own the S&P 500, where Nvidia is a huge influence. They own a Nasdaq fund. They own a technology ETF. Their super or 401(k) has a US growth allocation. Then they own Microsoft, Amazon, Alphabet and Meta — companies committing eye-watering amounts to AI infrastructure.
That is not diversification. It is repetition.
If AI investment remains productive and the earnings follow, terrific. You will do well. But if long-term bond yields rise, data-centre financing tightens, or corporate boards start demanding actual returns on their AI budgets, those holdings can all fall together. Different ticker codes do not make different economic exposures.
The broader backdrop is not especially forgiving. The 30-year US Treasury yield recently hit its highest level since 2007, raising the cost of capital for households and businesses. The Treasury Department’s decision to expand buybacks of long-dated debt gave markets only temporary relief. Investors are also heading into the Federal Reserve’s Jackson Hole symposium, running August 27 to 29, looking for clues on rate policy.
Markets are pricing one 25-basis-point rate rise in 2026, while the July personal-consumption-expenditures reading due today is expected to show inflation at 3.6%. In other words: the cost of money is still very much part of the story.
High-quality growth companies can thrive in that world. But high valuations are less forgiving when money costs more and future profits are discounted harder.
Don’t confuse a great company with an instruction to chase it
The contrarian point is that Nvidia does not need to collapse for late buyers to get a disappointing result. It merely needs to become more normal.
A company can continue producing excellent revenue growth, extraordinary margins and mountains of cash — then deliver mediocre shareholder returns if investors bought it assuming perfection. This happens all the time. The business wins; the buyer who paid too much treads water.
Nvidia’s forward price-to-earnings ratio is around 21 based on the next 12 months, according to Bloomberg data. That does not scream dot-com absurdity by itself, especially against the company’s growth rate. But it does tell you the market expects growth to slow from here while remaining very strong.
That is why the conversation should not be “Is Nvidia overvalued?” as though there is one magic answer. The better question is: what needs to happen for my purchase price to make sense?
If your answer requires revenues to keep doubling, margins to remain near 75%, Rubin to ramp cleanly, AI customers to keep funding massive builds, and interest rates not to create trouble, you have not bought a share. You have bought a chain of assumptions.
Sometimes those assumptions pay handsomely. Just don’t call it conservative investing.
What this means for you
Do not trade tonight’s earnings result because a bloke on the internet says Nvidia will beat or miss. That is punting with a Bloomberg terminal aesthetic.
Instead, do four useful things tomorrow:
1. Measure your real AI exposure. Add up Nvidia, chip ETFs, Nasdaq funds, S&P 500 funds and the mega-cap tech names across every account. You may discover your “diversified” portfolio is one crowded trade.
2. Set a maximum position size before emotion takes over. For most people building long-term wealth, a single stock should not be large enough to derail the plan if it drops 30% or 40%. Decide the limit while calm, not after a green candle makes you feel like Warren Buffett’s smarter cousin.
3. Keep buying broad assets on schedule. Regular contributions into diversified, low-cost funds are boring because they work without requiring you to predict one earnings call. Use individual shares as a deliberate satellite position, not the engine of your retirement.
4. Match risk to your actual life. If you need a home deposit, have expensive debt, lack an emergency buffer or run a business with lumpy cash flow, your first job is not chasing AI upside. It is becoming hard to kill financially.
Nvidia’s earnings matter because the company has become a pressure point for markets, AI financing and investor confidence. Watch the result. Learn from it. But don’t hand one company the keys to your future.
The richest people I know are not rich because they guessed every winning stock. They are rich because they survived the moments when everybody else mistook excitement for a plan.