Nvidia forecasts $3-$4 trillion spend

Nvidia used its May 20 earnings call to make a much larger claim than its quarterly results alone. The company said annual AI infrastructure spending could reach $3 trillion to $4 trillion by the end of the decade, extending its argument that the current buildout is still early. The forecast came as Nvidia reported record first-quarter fiscal 2027 revenue of $81.6 billion and data-center revenue of $75.2 billion. Investors, however, gave the report a muted after-hours response, even after Nvidia paired the results with an $80 billion additional share repurchase authorization and a dividend increase. ### Where did the $3 trillion to $4 trillion figure come from? Colette Kress, Nvidia's chief financial officer, gave the clearest version of the forecast on the May 20 call. "With analysts now forecasting hyperscale capex to exceed $1 trillion in 2027 and agentic AI beginning to proliferate across all industries, AI infrastructure spending is on track to reach $3 to $4 trillion annually by the end of this decade," CNBC reported her as saying. (investor.nvidia.com) Jensen Huang, Nvidia's chief executive, framed the same point in even broader terms. CNBC reported that Huang said hyperscaler capital spending was already around $1 trillion and was "growing toward the three to four" trillion-dollar range, a comment he said referred to hyperscalers such as Alphabet and Amazon and excluded some other parts of the supercomputing market. (cnbc.com) ### What did Nvidia actually report in the quarter? Nvidia said first-quarter fiscal 2027 revenue reached $81.6 billion for the period ended April 26, 2026, up 20% from the previous quarter and 85% from a year earlier. The company said data-center revenue was $75.2 billion, up 92% from a year earlier. The May 20 earnings release also announced an additional $80.0 billion share repurchase authorization. (cnbc.com) Nvidia said it was raising its quarterly cash dividend from $0.01 per share to $0.25 per share. ### Why are investors focusing on infrastructure spending instead of just revenue? Wall Street has been treating Nvidia's results as a readout on whether the largest cloud and internet companies will keep expanding AI capacity at the same pace. (investor.nvidia.com) CNBC said Huang's spending forecast ran well ahead of consensus estimates, citing Needham analysis that put hyperscaler capital expenditures at $1.03 trillion in 2028, far below Nvidia's implied path. That gap helps explain the market reaction. Nvidia's quarter was a record by its own numbers, but the investor question has shifted toward whether the rest of the industry will spend enough to support forecasts already embedded in valuations, according to CNBC's account of the call and analyst response. (cnbc.com) ### What does Nvidia say is driving that scale of spending? Huang tied the forecast to a wider buildout of AI systems and agents. CNBC reported that he said the world would have "billions of agents" and that those agents would create further demand for computing capacity. Nvidia's own results support the case that demand remains concentrated in large-scale computing. (cnbc.com) The company said data center accounted for the vast majority of quarterly revenue, with $75.2 billion of the $81.6 billion total. ### What comes next for investors trying to test the forecast? Nvidia said on April 29 that the May 20 first-quarter webcast would remain available until the company's conference call for second-quarter fiscal 2027 results. (cnbc.com) The company did not give a date in that notice for the next earnings call, but said the replay would stay up until then on its investor relations site. (investor.nvidia.com) The next test will come in Nvidia's second-quarter fiscal 2027 report, when investors will look for updated revenue guidance, data-center demand figures and any new commentary from Huang or Kress on hyperscaler capital spending. (investor.nvidia.com 1) (investor.nvidia.com 2)