Jensen Huang just put Nvidia’s Blackwell and Vera Rubin sales projections into the $1 trillion stratosphere
Jensen Huang reveals demand for next-gen AI chips has doubled to a staggering $1 trillion in just months.
At Nvidia's GTC conference, CEO Jensen Huang made a staggering financial projection: the company now sees $1 trillion in demand for its Blackwell and upcoming Vera Rubin AI accelerator chips through 2027. This figure represents a doubling from the $500 billion in projected demand Huang cited just months earlier at GTC DC, reflecting the explosive and accelerating investment in AI compute infrastructure. The announcement underscores Nvidia's dominant position in supplying the hardware backbone for the global AI boom, with cloud providers and enterprises placing enormous forward orders.
Huang also provided new technical details on the Vera Rubin architecture, first announced in 2024 and now entering production. Nvidia claims Vera Rubin will operate 3.5 times faster than the current Blackwell architecture on AI model training tasks and deliver a 5x speedup on inference tasks, reaching performance levels up to 50 petaflops. Production is expected to ramp up in the second half of this year. The $1 trillion projection is not a revenue forecast but an estimate of total market demand, indicating the scale of investment required to build out next-generation AI data centers.
- Nvidia CEO Jensen Huang projects $1 trillion in demand for Blackwell and Vera Rubin AI chips through 2027, doubling previous estimates.
- The new Vera Rubin architecture, entering production this year, promises 3.5x faster training and 5x faster inference than current Blackwell chips.
- The projection reflects massive, sustained capital expenditure by cloud providers and enterprises to build out AI infrastructure capacity.
Why It Matters
This signals the immense scale and long-term capital commitment required for the AI era, solidifying Nvidia's central role.