Nvidia CEO Jensen Huang announced a massive surge in market demand during the company’s annual GTC Conference in San Jose, California, projecting that orders for its Blackwell and Vera Rubin AI chips will reach a staggering $1 trillion by 2027. This milestone signals an unprecedented acceleration in the global AI hardware market, doubling previous demand estimates.
Doubling Down on AI Growth
Just one year after initial projections, Huang revealed that the outlook for Nvidia’s next-generation architecture has shifted dramatically. While the company previously estimated $500 billion in demand for these chips through 2026, the current market trajectory has forced that figure to $1 trillion.
“$500 billion is an enormous amount of revenue,” Huang stated during his keynote. “I am here to tell you that right now, where I stand—a few short months after GTC DC—I see at least $1 trillion through 2027.”
The Technological Leap: Blackwell vs. Rubin
The Vera Rubin architecture represents the next frontier in AI computing. Positioned as the state-of-the-art successor to the Blackwell platform, Rubin is engineered to handle increasingly complex model training and inference workloads.
According to technical specifications released in January, the Rubin architecture delivers significant performance gains over its predecessor:
- Model Training: 3.5x faster than Blackwell.
- Inference Tasks: 5x faster than Blackwell.
- Peak Performance: Reaching as high as 50 petaflops.
Scaling Production for 2025
As the demand for high-performance computing continues to outpace supply, Nvidia is moving quickly to solidify its manufacturing capacity. The company confirmed it expects to ramp up production of its latest chip architectures in the second half of the year to meet the massive order backlog and support the aggressive expansion of global data center infrastructure.