News

Nvidia CEO Jensen Huang adopts the term GPU-as-a-Service to describe AI infrastructure delivery

Monday, March 16, 2026 at 08:11 PM

Nvidia CEO Jensen Huang is reportedly using the term GPU-as-a-Service (GaaS) instead of AI-as-a-Service (AaaS) to describe the company's delivery model for AI infrastructure.

Context

At the GTC 2026 conference on March 16, 2026, Nvidia CEO Jensen Huang officially pivoted the industry’s vocabulary toward GPU-as-a-Service (GaaS) to describe the delivery of AI infrastructure. This shift distances the company from the broader 'AI-as-a-Service' (AIaaS) label, emphasizing the specific hardware-accelerated compute that powers the world's most complex models. The move reflects Nvidia’s strategy to position its DGX Cloud and hardware stack as the essential utility for an era of agentic and physical AI. This branding evolution coincides with a massive industrial buildout where renting Nvidia GPUs has become increasingly difficult due to high demand. By framing infrastructure as GaaS, Nvidia highlights the critical role of its silicon in the global supply chain, where spot prices for even older GPU generations are rising. For investors, this signals a transition from selling discrete chips to a more integrated, service-oriented model that ensures Nvidia remains the foundational 'AI factory' for enterprises and sovereign nations.

Related Companies

Nvidia
Nvidia
NVDA
US