Huang’s trillion‑dollar forecast
At GTC 2026 Jensen Huang projected that cumulative demand for NVIDIA’s next‑gen AI chip architectures will reach at least $1 trillion — a signal investors and datacenter builders are taking seriously. Analysts framed the comment as underlining massive, long‑term demand for accelerated compute across enterprises. ( )
Jensen Huang delivered the keynote at NVIDIA’s GTC on March 16, 2026 and framed the company’s next-gen initiatives around its Blackwell and the newly unveiled Vera Rubin architectures. (cnbc.com) Huang’s announcement doubled the company’s prior outlook from about $500 billion in expected demand and reset revenue visibility for the next two years. (cnbc.com) NVIDIA said the Vera Rubin platform entered production with seven new chips and rack designs — including NVL72 GPU racks and integrated Groq-derived LPUs after last year’s roughly $20 billion Groq deal. (investor.nvidia.com) The company also highlighted expanded enterprise partnerships and cloud tie‑ups, naming new or deeper pacts with IBM and Adobe while hyperscalers detailed NVIDIA-based cloud offerings at GTC. (bloomberg.com) Wall Street reaction mixed: Wedbush’s Dan Ives called the keynote a “confidence boost,” while some analysts flagged a “wall of worry” about capex sustainability; shares ran higher intra‑day—touching roughly a 4% pop—before paring gains. (thestreet.com) Industry research firms said the event underscored NVIDIA’s intent to own more of the AI value chain by packaging silicon, systems and software into reference AI factories for large-scale inference deployments. (forrester.com)
