SAN JOSE, Calif. — Nvidia Corp. CEO Jensen Huang took the stage at the company's flagship GPU Technology Conference on Monday before a packed audience at the SAP Center, delivering what analysts are calling the most consequential keynote in the chipmaker's history. Standing in his signature black leather jacket, Huang announced that Nvidia anticipates generating at least $1 trillion in revenue from its current-generation Blackwell and next-generation Vera Rubin chip platforms through the end of 2027 — nearly doubling the $500 billion forecast he provided just five months ago.
"I see through 2027 at least $1 trillion," Huang told the crowd to sustained applause. "And I am certain computing demand will be much higher than that." The remarks crystallized the accelerating pace of AI infrastructure investment worldwide, with Nvidia positioned at the center of a global buildout that now spans cloud hyperscalers, sovereign governments, and a new class of AI-native startups.
Among the hardware highlights, Huang unveiled the Vera Rubin NVLink 72, a rack-scale system delivering what the company claims is 10 times more performance per watt compared to its predecessor, Grace Blackwell. He also announced the Kyber architecture, a next-generation rack system integrating 144 GPUs in vertically oriented compute trays to improve density and reduce latency, scheduled to ship in 2027 as part of the Vera Rubin Ultra platform.
In a significant strategic pivot, Huang devoted considerable time to Nvidia's CPU ambitions. The company's Vera CPU, now in production, is being positioned as the critical orchestration layer for agentic AI workloads — applications where AI systems autonomously execute tasks, make decisions, and coordinate complex multi-step processes. "CPUs are becoming the bottleneck in terms of growing out this AI and agentic workflow," said Dion Harris, Nvidia's head of AI infrastructure, ahead of the event.
Huang also introduced the Nvidia Groq 3 Language Processing Unit, a high-efficiency inference chip stemming from Nvidia's $20 billion acquisition of Groq in December 2025 — the company's largest deal on record. The chip is expected to ship in the third quarter and targets the fast-growing inference market, where rivals including Google and Amazon have made inroads with custom silicon.
Shares of Nvidia rose approximately 2.3% on Monday, closing near session highs. The VanEck Semiconductor ETF gained 1.7%, while broader semiconductor names including AMD, Broadcom, and Marvell Technology also rallied in sympathy. Analysts widely view Huang's $1 trillion projection as a credibility-enhancing signal for the entire AI ecosystem, and several investment banks indicated they would be revising price targets upward in the coming days.
The GTC conference continues through March 19 and is expected to feature additional announcements around autonomous vehicles — Nvidia has added four new carmaker partnerships — as well as updates to its Omniverse digital twin platform and expanded enterprise agentic AI tooling. For investors, the week-long event represents the most concentrated period of AI-related catalysts in the first half of 2026.