After Wednesday’s close in the U.S. stock market, Jensen Huang picks up the microphone.

The average expectations analysts gave were: second-quarter revenue of $92.18 billion, year over year nearly doubling, the fastest growth in seven quarters, driven by data center business growth of more than twofold. This $5 trillion company has already beaten analysts’ expectations for 14 straight quarters—its net profit in the prior quarter grew 210% year over year, while Wall Street had forecast 126%, almost a doubling.

By all accounts, this is the kind of night that should make people nervous.

But the options market pricing is telling a different story.

What are options saying?

The magnitude of the move priced in for the day after the earnings release is 5.4%, implying roughly a $280 billion change in market value—larger than the total market cap of each of 90% of the constituents in the S&P 500.

Sounds scary, but it depends on the frame of reference: before the May earnings report, the volatility priced into the market was 6.5%; over the past 12 quarters, the actual average volatility after earnings has been 7.4%. In other words, this is the calmest period expected in the last two years.

The founder of options analytics firm ORATS, Matt Amberson, put it bluntly: “This shows a kind of complacency about Nvidia—and it also means it’s becoming more predictable.”

Chris Murphy, co-head of derivatives strategy at market maker Susquehanna, put it more thoroughly: “In the early days of the AI era, Nvidia always managed to surprise everyone—so often it was 10%, 15%, 20% swings. That phase is basically over. There’s really nobody in the market who truly believes they’ll deliver a huge upside surprise to everyone and catch them off guard, and then the stock jumps.”

One indication is its stock price this year.

Year-to-date, Nvidia is up 11.7% and the S&P 500 is up 11.8%—it has only slightly lagged the broader market. The Philadelphia Semiconductor Index is up 61%. The largest sell-shovel in the entire AI buildout has, this year, underperformed the whole shovel industry. On Monday, it closed out the seventh straight trading day of declines—the longest streak since 2022.

Last month, it even briefly ceded the title of “most valuable company in the world” to Apple.

On Tuesday ahead of the open, it rebounded by about 1%, putting an end to this streak of declines. But the driver of the rebound isn’t so much Nvidia itself as Brent crude falling to around $89. Kathleen Brooks, director of research at XTB, said: “A shift in sentiment, helped by falling oil prices.”

So what exactly is the market waiting for?

Sara Araghi, portfolio manager at Franklin Equity, a unit of Franklin Templeton, made the requirement very clear on Bloomberg TV on Tuesday: Nvidia needs more than a pretty earnings report—it needs a concrete explanation of how it plans to deploy capital and how it intends to maintain its spending plan.

“More details on these investments, and the value of these investments—I think the market needs to see that.”

She pointed to a disconnect: Nvidia’s expected price-to-earnings ratio over the next 12 months is now around 21x, which doesn’t match its expected revenue and earnings growth at all. This multiple suggests the market is pricing in growth slowing down.

“(Growth) is pretty extraordinary,” she said, “but a slowdown is on the way—and unfortunately, the market is looking at next year.”

She also gave two specific checkpoints.

First is gross margin. Nvidia’s gross margin is about 75%, which is unusual for a hardware company and will be scrutinized line by line in a backdrop of rising raw material costs. To hedge higher memory costs, Nvidia has already been raising prices: for early-2027 servers based on the Vera Rubin and Grace Blackwell architectures, quotes to major customers will rise by more than 15%.

Second, it’s what they plan to use that money for. Profit growth matters because it convinces Wall Street that “the free cash flow they generate is sufficient to support the investments they need to make.” She added: “And they must use that money to buy back shares.”

Analysts at Bloomberg Intelligence issued the same warning: a routine “beat-and-raise” plus guidance might no longer be enough to boost sentiment.

Where is the money really flowing?

That’s the real issue at stake in this earnings report. Over the past month, what Nvidia has been doing has gone beyond the realm of a chip company:

It teamed up with six of Wall Street’s largest financial institutions to arrange a $50 billion AI financing plan, promising to back loans for customers who can’t afford its chips.

Last week, it agreed to provide OpenAI with a guarantee of up to $105 billion for the leasing of a large data center in Ohio—20-year lease term. This is one of its biggest AI financing commitments at its scale.

Also last week, it took a stake in Cloverleaf Infrastructure, a company specializing in lining up power for data centers.

It also struck a $6 billion deal with the startup Poolside, aiming to develop a powerful open-weight AI model.

Zacks Investment Management’s chief market strategist Brian Mulberry gave a name to this move: “It makes them play a kind of central bank role in the AI space.” He holds Nvidia shares, but also spelled out where the risk lies: “The real risk is being completely exposed to AI without any diversification. For this to work, the key is that adoption of AI tools must keep growing.”

“Circular transactions” is the core suspicion here: if I lend you money so you can buy my chips, then is my revenue demand—or chips that I’m manufacturing myself?

Huang Renxun’s defense is straightforward: the logic is simple. The company has ample cash, so it can use that money to support customers that are growing fast but still losing money. He said the Ohio deal does not constitute circular financing.

For that claim to hold, the premise must be that those customers can ultimately make money. And the signals lately don’t look good: OpenAI recently told investors that its Q2 revenue rose by only 18%, and its losses are still widening.

Huang Renxun tonight faces more than just himself

Nvidia’s earnings report arrives against a rather unfriendly backdrop.

AI’s political backlash is growing. Across the U.S., more than 500 towns and cities have restricted the construction of data centers. Bond selloffs have pushed borrowing costs to multi-year highs. Last week, the yield on the 30-year U.S. Treasury hit a 19-year high. Those former cash printers—hyperscale cloud providers—are increasingly relying on debt to fund their buildouts, and this year large tech companies’ data center spending is expected to exceed $730 billion.

Zacks’ Mulberry used a metaphor to describe how heavy this earnings report is: “It’s getting more like a World Cup final, not a Super Bowl. It’s that big.”

The answer investors really want is just one: how fast the transition from Blackwell to Vera Rubin can be, with shipments expected to begin this autumn.

AJ Bell investment director Russ Mould put it plainly: “Nvidia’s upcoming results have the power to lift the entire market or drag it down. The clue investors are looking for is—are AI demands losing momentum?”