Nvidia now expects to generate $1 trillion in revenue from its Blackwell and Vera Rubin platforms by 2027, CEO Jensen Huang said at the annual GPU Technology Conference in California. Revenue expectations are up 200% from the $500 billion in revenue by 2026 that the company previously forecast at last year’s GTC.
“A trillion dollars is a huge piece of infrastructure; you have to be absolutely sure that the trillion dollars you put in will be used, will be efficient, will be incredibly cost-effective, and will have a useful life as long as you see it,” Huang said.
Huang argued that this “complete confidence” stems from the belief that demand for AI will grow exponentially thanks to rapid developments in agent AI. This belief is based on a new belief token driven AI economy he previously detailed in the company’s latest earnings report. This new revenue model is based on the growing importance of inference with the proliferation of agent AI models.
As agency AI hype takes over Silicon Valley and models mature, the volume of data handled by an AI system expands, and inference becomes more important than training. The turning point for this new order came with Anthropic’s Claude Code AI agent, Huang claimed.
“Claude Code revolutionized software engineering,” Huang said. “There isn’t a software engineer (at Nvidia) today without one or more AI agents helping them code.”
At many points in his more than two-and-a-half-hour keynote, Huang compared agent AI to other fundamental technological advances, at one point calling it “the new computer.”
“Every SaaS company will be an AgaaS company, an agent as a service company,” Huang said.
It’s not the first time Huang has praised agent artificial intelligence and how it’s changing the tech industry, and it likely won’t be the last. In January, agent AI and the importance of inference took center stage at the launch Nvidia’s Rubin platform. Just a few weeks ago, the company made its biggest acquisition ever, buying Grog (not Grok), a chip maker that specializes in inference.
The fruits of Nvidia’s renewed mission were three announcements: a big push into CPUs, the first Grog chips Nvidia will ship in the second half of the year, and a collaboration. OpenClawan open-source AI agent software that made waves with its viral success earlier this year after launching under a different name a few months ago.
“OpenClaw Open source essentially has an operating system for agent computers,” Huang said. “It’s not unlike how Windows allows us to build personal computers.”
He then talked about how every company needs an “HTML-strategy” in the rise of the internet, and now, in the age of AI agents, every company will need an “OpenClaw strategy”.
But OpenClaw is a bit of a wild card. Creating an agent requires full access to your computer and files a cyber security minefield. Top tech companies and even Chinese government employees are reportedly advised not to rely on OpenClaw and similar agent AI platforms due to security fears. Giving an artificial intelligence too much control over your computer can have some serious risks, such as OpenClaw’s AI agent. deletes your entire inboxactually happened to a Meta executive last month.
To address at least some of these concerns, Huang revealed NemoClawNvidia’s attempt to make OpenClaw more secure and private for enterprises to use.
Nvidia’s push into the OpenClaw world is also symbolic of the company’s desire become more competitive in the open source spacesees its role as an emerging open-source model provider as a means to further global trust in the company’s hardware.
In addition to the results-based announcements, Huang also revealed that the company is working on a new Vera Rubin computer. space-based AI data centersand Hyundai will partner with Nissan and top Chinese automakers BYD and Geely to build 18 million robot-axis annually.
The world of finance has grown tired the multibillion-dollar AI investments and spending commitments and rapid growth trajectory it once loved I asked He once believed with confidence. You can see how much this skepticism has grown over the past few months by looking at how investors have made it significantly harder to buy into each announcement.
Investors are concerned that Nvidia’s revenue growth may have peaked Bloomberg, shares fell 5.5% the day after a stellar earnings report last month. Huang’s statements on Monday and his belief in agent AI may not have changed that outlook, with the company’s shares down slightly less than 1% after the market close, despite the stock’s typically key performance.
