Nvidia posted record earnings as corporations and sovereign states continue to deploy agentic and physical AI tools. 

The chip designer reported data center revenue of $62 billion in the fourth quarter, an increase of 75% year over year and 22% sequentially, according to its earnings report, published Feb. 25. 

“We have now scaled our data center [revenue] by nearly 13x since the emergence of ChatGPT in [November 2022],” Chief Financial Officer Colette Kress said during the company’s Feb. 25 earnings call.

Agentic AI tools have reached an “inflection point,” Kress said, adding that “adoption is skyrocketing and tokens are profitable, driving extreme urgency to scale up compute.” 

To meet the rising demand of AI hardware, Nvidia is increasing its research and development spend to $20 billion, Kress said, without specifying what the spend was previously. She added that agentic AI and physical AI, or robots, increasingly are being built on multimodel systems, and Nvidia plans to serve all client needs. 

Many financial services companies like GFT Technologies are also developing physical AI tools to help FIs deploy consumer facing AI teller machines, according to FinAi News’ prior reporting. 

AI CapEx and hyperscalers 

Expectations for 2026 AI capital expenditure across the top cloud providers and hyperscalers, which account for just more than 50% of Nvidia’s data center revenue, “are up nearly $120 billion since the start of the year and approaching $700 billion,” Kress said on the call. 

Hyperscalers are companies that build enormous cloud facilities that deliver computing power at industrial scale, according to Britannica.  

Aggregate AI CapEx by hyperscalers in 2026 is forecast to be $660 billion to $680 billion, up roughly 70% from 2025 levels, according to an analysis by FinAi News. 

The hyperscalers reported the following CapEx projections in their latest earnings: 

Hyperscalers, including Meta, Amazon, Google and Microsoft, saw their share prices drop this month after announcing increased AI capital expenditures for 2026. 

Short-term credit and stock pressures reflect the intense early-stage buildout of the AI supercycle — massive upfront CapEx for infrastructure that will deliver exponential returns as models scale and monetization accelerates, Ben T. Smith IV, lead of communications, media and technology practice at global consultancy Kearney, told FinAi News. 

“Markets are pricing execution risks, but proven demand from hyperscalers and enterprises points to long-term winners emerging stronger,” Smith said. 

It’s not just hyperscalers that are pouring money into the data center buildout, Kress said, adding that sovereign states are also investing in AI infrastructure. 

“Every country will build and operate some parts of its AI infrastructure, just like with electricity and internet today,” Kress said. “In fiscal year 2026, our sovereign AI business more than tripled year over year and over $30 billion, driven primarily by customers based in Canada, France, the Netherlands, Singapore and the U.K.” 

In Q4, Santa Clara, Calif.-based Nvidia reported: 

• Revenue of $68.1 billion, up 73% YoY; 

• Gross margin of 75%, up 2 percentage points YoY; and 

• Net income of $43 billion, up 94% YoY. 

Register here for the inaugural FinAi Banking Summit, taking place March 2-3 in Denver. View the full event agenda here.