Nvidia CEO Jensen Huang unveiled a new inference system at the company’s annual GTC conference, signaling a major strategic shift as the AI industry moves beyond model training and into large-scale deployment.
The announcement represents one of Nvidia’s most decisive efforts to defend its dominance, as inference—running AI models in real-world applications—rapidly becomes the next key battleground in the sector.
Huang emphasized the scale of upcoming demand, stating that Nvidia expects at least $1 trillion in demand for its Blackwell and Rubin AI systems through 2027.
This marks a sharp increase from the roughly $500 billion in demand the company had previously projected through 2026, underscoring how quickly adoption is accelerating.