Can Nvidia's Blowout Earnings Keep the AI Rally Alive?

In the four previous quarters, Nvidia's stock fell the day after earnings.

After Wednesday's close, Nvidia reported quarterly revenue of $96.22 billion, more than doubling year over year and setting a record. It was the 15th consecutive quarter the company beat Wall Street expectations.

But what turned the stock from down to up wasn't that number. It was another figure CFO Colette Kress gave 40 minutes later on the earnings call.

First, the beat

Revenue came in at $96.22 billion, versus the $92.27 billion analysts expected.

Data center revenue was $89 billion, up 117% year over year. Edge computing — chips for AI features in computers, game consoles, and robots — brought in $7.2 billion, up 27%.

Net income was $59.69 billion, or $2.46 per share, compared with $26.42 billion, or $1.08 per share, a year earlier. Excluding certain items, earnings were $2.22 per share, above the $2.09 consensus.

Operating expenses also rose 55% to $8.41 billion.

In a statement, Jensen Huang offered his own framing: "AI has reached an inflection point. It is doing useful work. Its tokens are productive and profitable. Now, compute is revenue."

Initially, the stock drifted lower in after-hours trading. For the first 40 minutes after the release, it barely moved.

Expectations were simply too high. The stock has rallied for three years, with its market cap soaring from $400 billion to $5 trillion. Investors weren't looking for just another good quarter.

Then the call began, and Kress delivered two things.

First, next quarter's revenue guidance: $108 billion, plus or minus 2%. Analysts had expected between $104.2 billion and $105.2 billion. If achieved, it would be Nvidia's first $100 billion quarter — a feat only nine other S&P 500 companies have accomplished.

Second, and more significant: fiscal 2028 revenue is expected to grow about 70%. Analysts surveyed by FactSet expected 45%, and LSEG's figure was 44.8%. She raised the outlook by 25 percentage points.

She added an almost boastful remark: the company remains "supply constrained," otherwise next year's revenue could double.

Also announced was an expanded partnership with Amazon Web Services. In 2027 and 2028, AWS will deploy an additional 2 million of Nvidia's latest GPUs across its global infrastructure.

The stock immediately rose 4.2%, with volume exceeding 50 million shares.

Seth Hickle, chief investment officer at Mindset Wealth Management, said: "It's hard to interpret this report as anything less than stunning."

The circular financing question

Before this report, the biggest issue hanging over Nvidia wasn't demand — it was that the company itself was helping fund that demand.

Over the past month, Nvidia has gone beyond the role of a chipmaker: it joined six of Wall Street's largest financial institutions to partially guarantee up to $500 billion in financing for customers building data centers, and it agreed to backstop OpenAI's large data center project in Ohio, potentially on the hook for billions if leasing plans fall through.

The question is blunt: if I lend you money to buy my chips, is my revenue real demand or something I manufactured myself?

Kress didn't dodge the term. On the call, she said:

"We recognize the scale of this support, and we know some call it circular financing. We see it differently."

Her defense: Nvidia expects large frontier labs like OpenAI "to become the largest technology companies in history," and their current bottleneck is access to enough compute to develop products and improve models, so they need Nvidia's financing.

Then she doubled down, previewing the next deal: the company will provide "selective credit enhancement" for nearly 2 gigawatts of compute for another frontier AI lab. She didn't name it.

On the call, Jensen Huang emphasized that demand is broadening: "This time last year, only one lab was driving this buildout; today, we have a golden age of new AI labs and startups."

Jay Hatfield, CEO of Infrastructure Capital Advisors, summed up the company's position with a joke: "It feels a lot like a boring day waiting for the Fed, but really it's waiting for Nvidia. Some people call Jensen Huang the Fed chair of AI."

The one line heading down

The report wasn't flawless. There was only one crack, but it was glaring.

Second-quarter gross margin was 75%. Third-quarter guidance is about 74%, plus or minus 50 basis points — below the roughly 74.8% analysts expected. And the company itself warned that as shortages intensify, gross margin will keep sliding, potentially to 71% early next year before improving.

The reason is memory. Data center demand is draining memory chip supply, and tech giants including Apple are absorbing that cost.

One number shows how urgent this is: Nvidia's purchase commitments with suppliers jumped from $119 billion last quarter to $279 billion — an increase of $160 billion in a single quarter — which the company said is "primarily related to memory procurement."

It's already passing costs along: servers based on the Vera Rubin and Grace Blackwell architectures, due in early 2027, will see prices rise more than 15% for major customers.

Even with this report, one trend isn't being stopped by the numbers: more and more tech companies' planned spending is flowing to their own chips, aiming to reduce reliance on Nvidia's expensive and supply-constrained processors.

Big Tech's AI infrastructure spending is expected to exceed $730 billion this year, up from $400 billion last year. How much of that ultimately lands in Nvidia's accounts depends on the progress of in-house chip development.

One more detail worth noting: Nvidia's stock is up 12.4% this year, while AMD and Intel have both more than doubled over the same period.

In the four previous quarters, Nvidia's stock fell the day after earnings.