At Nvidia's GTC conference in San Jose, CEO Jensen Huang delivered a keynote dense with technical specifications. Yet one number cut through the engineering details for investors: a projected $1 trillion in orders for the company's Blackwell and upcoming Vera Rubin chip architectures through 2027.

Huang revealed that demand for these systems had already reached an estimated $500 billion just a year ago. "$500 billion is an enormous amount of revenue," he told the audience. "Right here where I stand, I see through 2027, at least $1 trillion." This forecast underscores the staggering scale of investment flowing into artificial intelligence infrastructure.

The Rubin architecture, first announced in 2024 and now in production, is positioned as a significant leap beyond the current Blackwell platform. Nvidia claims Rubin will perform 3.5 times faster in training AI models and up to five times faster in running them, achieving speeds of 50 petaflops. The company plans to increase production volume in the latter half of this year.

This financial projection, far exceeding typical industry figures, translates the technical race for AI supremacy into a concrete market expectation. It signals Nvidia's confidence that the demand for advanced computing power, driven by global enterprises and government initiatives, has only begun to accelerate.

Source: TechCrunch