Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Nvidia’s strong earnings and growth forecast met a muted investor response on Wednesday as anxiety about the sustainability of the AI investment boom weighed on the world’s most valuable company.
The semiconductor behemoth said it expects revenue of $78bn for the current quarter, well above Wall Street expectations of $72.1bn.
“Computing demand is growing exponentially . . . Our customers are racing to invest in AI compute,” chief executive Jensen Huang said.
But despite that nearly $6bn positive surprise, shares in the company at the heart of the AI sector ended flat in after hours-trading, reflecting a growing wariness about the huge spending by Nvidia’s largest Silicon Valley customers and a memory chip shortage squeezing the industry.
Melissa Otto, head of research at Visible Alpha, said Nvidia faces “a lot of concerns” around how the Big Tech and start-up clients it relies on are funding their hundreds of billions in AI infrastructure spending.
“These are the critical things for the stock that really get to why it isn’t moving,” she said. “We didn’t get a lot of visibility about how the rest of the year is going to play out.”
Nvidia’s earnings — considered a bellwether for the AI boom because of its chips’ crucial role in running the technology — came after weeks of stock market volatility triggered by worries about AI upending traditional industries as well as the astronomical spending spree on AI data centres.
The company reported revenue of $68.1bn for the final quarter of its financial year, which ends in late January, up 73 per cent from a year ago, and ahead of its prior guidance of roughly $65bn. Its performance beat analyst estimates of $66.2bn compiled by Visible Alpha.
Data centre revenue — which refers to Nvidia’s AI chip technology, including its current generation Blackwell hardware launched last year — was $62.3bn, better than estimates of $60.5bn.
Nvidia’s total annual revenue for its fiscal year topped $200bn for the first time in 2026, while its net income was $120bn. Net income in the final quarter was $43bn, beating expectations of $36.4bn.
It also met analyst expectations for its gross margins, at 75 per cent, which it said it expects to be sustained through the April quarter, with “mid-seventies” margins into the rest of the year.
Analysts have become increasingly concerned about Nvidia’s reliance on a small group of huge data centre players, known as hyperscalers, as well as AI start-ups such as OpenAI.
Microsoft, Google, Amazon and Meta together are projecting $660bn in capital expenditures this year, mostly focused on AI data centres. These tech giants are now facing the prospect of tapping bond and equity markets as they burn through mountains of cash.
AI start-ups — which are huge consumers of data centre capacity to train and run systems such as ChatGPT and Anthropic’s Claude — meanwhile need to continue landing huge fundraising deals to pay for computing power.
Huang dismissed concerns that the spending speed was unsustainable, saying he was “confident in [hyperscaler] cash flow growing”.
The Nvidia boss said adding computing power would lead to revenues for these groups. “In this new world of AI, compute equals revenues.”
He also predicted AI usage would continue to accelerate as advances by frontier AI model builders such as Anthropic and OpenAI have also “opened the floodgates for enterprise AI adoption”.
Nvidia is investing $30bn in OpenAI’s latest funding round, replacing an earlier $100bn multiyear strategic partnership that would have seen it invest in increments over time as its chips were used, the FT reported last week.
Huang said Nvidia was “thrilled with our partnership with OpenAI”, which he called a “once in a generation company”.
Another headwind for Nvidia is a general shortage of memory chips as data centre builders hoover up the limited supply of high-bandwidth memory from a handful of companies: SK Hynix, Samsung and Micron.
Otto at Visible Alpha said that the ongoing memory chip supply crunch meant Nvidia’s margins were now harder to predict, since Nvidia needs memory to run alongside its advanced processors.
Nvidia chief financial officer Colette Kress said the company had “inventory and supply commitments in place to address future demand including shipments extending into calendar 2027”.
“I think there needs to be more precision and transparency around, ok, exactly what is the impact [of the memory crunch] and how is it going to affect gross margin?” Otto said.
Nvidia did not include any potential AI chip revenue from China in its projection. Access to this crucial market could lift its revenues by billions of dollars if it is able to navigate the tensions between Washington and Beijing.
A deal with the White House in December gave the go-ahead to sales of Nvidia’s H200 AI chip to China. But the FT reported in January that US security reviews had delayed the issuing of licences to Nvidia’s China customers.
Kress said “small amounts” of the H200 had been “approved by the US government” but that Nvidia had “yet to generate any revenue — and we do not know whether any imports will be allowed into China”.
Unlocking the Chinese market could give a fresh boost to Nvidia, which faces ever-greater investor pressure to outperform on its financial metrics and reassure markets about the health of the AI sector.
Despite “perfect” headline numbers, Nvidia was facing “almost an unbridled scepticism that this party simply cannot continue”, said Daniel Newman, chief executive at The Futurum Group. “It has created a stream of outperformances met with lukewarm reactions.”
