Nvidia CEO Predicts $1 Trillion in AI Chip Orders Through 2027

Nvidia Doubles Down on AI Chip Demand

Nvidia CEO Jensen Huang delivered his annual GTC conference keynote in San Jose, California, on March 16, 2026, and unveiled some jaw-dropping projections for the companys AI chip business. During his address, Huang revealed that he expects purchase orders for Nvidia Blackwell and Vera Rubin chips to reach an astonishing $1 trillion through 2027, representing a massive doubling from the $500 billion projection through 2026 that was announced just last year. This dramatic increase in demand forecasts underscores the explosive growth of the AI sector and Nvidias dominant position in the market.

The company, currently the worlds most valuable publicly traded company, continues to ride the wave of unprecedented AI infrastructure spending from tech giants, cloud providers, and enterprises worldwide. According to The New York Times, for three years Jensen Huang has described his companys chips as the Swiss Army knife of artificial intelligence, serving as an all-purpose tool ideal for building and running AI applications. This analogy resonates because Nvidias chips have become the foundational infrastructure powering virtually every major AI breakthrough in recent years.

New Inference Chip and Space Announcements

Beyond the headline-grabbing financial projections, Huang also unveiled several new products during his keynote. Most notably, Nvidia announced the Groq 3 Language Processing Unit (LPU), the companys first chip from the startup that Nvidia acquired through a massive $20 billion asset purchase in December 2025 its largest deal ever. The acquisition, reported by CNBC, marked Nvidias entry into the inference chip market.

The new Groq 3 LPU is designed for inference workloads, which are becoming increasingly important as AI models transition from training to deployment. Huang claimed the new chip can speed up inference workloads by up to 35 times compared to previous generations. According to Business Insider, this represents Nvidias most decisive move yet to defend its dominance as inference becomes AIs next battleground.

In a forward-thinking announcement, Nvidia also revealed plans for the Vera Rubin Space-1 chip system, designed for orbital AI data centers. As Huang put it during his keynote, As we deploy satellite constellations and explore deeper into space, intelligence must live wherever data is generated. This ambitious project highlights Nvidias vision for distributed computing across the solar system.

Market Dominance and Future Outlook

Despite rising competition from chipmakers like AMD, Intel, and custom silicon from Google and Amazon, Nvidia is expected to maintain its dominant 90 percent share of chips used for AI development. Huang indicated that the company will claim approximately one-third of chips used to run AI inference workloads. This balanced approach ensures Nvidia remains the go-to choice for both AI training and inference applications.

According to Emarketer analyst Jacob Bourne, Huang mapping out a $1 trillion opportunity through 2027 underscores the durable demand for Nvidias AI infrastructure despite investor concerns. This sentiment was echoed by investors, with Nvidias stock price surging following the announcement. TechCrunch reported that Huang put Nvidias Blackwell and Vera Rubin sales projections into the $1 trillion stratosphere.

The GTC 2026 conference, Nvidias annual developer conference, brought together over 10,000 developers, researchers, and industry leaders to explore the future of AI, autonomous vehicles, robotics, and more. Huang also emphasized the importance of agentic AI and enterprise adoption of AI agents, describing it as the missing infrastructure layer that every company needs.

What This Means for the AI Industry

The $1 trillion projection signals that the AI revolution is far from over, with demand for AI infrastructure continuing to accelerate across all industries. From healthcare to finance, autonomous vehicles to robotics, every sector is investing heavily in AI capabilities. Nvidias announcements at GTC 2026 make it clear that the company is positioning itself to capture every aspect of the AI computing stack, from training to inference, from data centers to space-based orbital computing.

For investors and tech enthusiasts, Nvidias GTC 2026 keynote demonstrated that the company is not content to rest on its laurels. By expanding into inference chips, space computing, and enterprise AI agents, Jensen Huang is charting a course for Nvidia to remain at the forefront of the AI revolution for years to come. The doubling of demand forecasts from $500 billion to $1 trillion in just one year speaks to the insatiable appetite for AI processing power across the global technology landscape.