Jensen Huang: AI doom predictions lack scientific basis

At the All-In Summit 2026 on September 15, 2026, Nvidia founder and CEO Jensen Huang publicly dismissed "AI doom" predictions as "unscientific sensationalism." He said that the safety-related parts of Anthropic CEO Dario Amodei's recent essay deserve serious consideration, but the "scientific predictions" about the future lack evidence.

Regarding recursive self-improvement (RSI), Huang described it as a combination of reasonable ideas including in-context learning, skill accumulation, reinforcement learning, and synthetic data generation. He explained that low-rank adaptation methods such as LoRA can improve model capabilities without retraining the base model, making the entire process controllable from an engineering perspective. He also said the term "RSI" is being weaponized by some to create an impression of runaway AI.

Huang further revealed that Nvidia has frontier open-source models in five verticals including autonomous driving and biopharma, such as Alpamayo, described as the world's first "thinking" self-driving car, and Proteina Complexa, a protein synthesis model. On financing, Nvidia is working with Blackstone and Goldman Sachs on the Cloverleaf project to support AI infrastructure investment, while expanding regional cloud partner networks in Australia and Southeast Asia.

During the summit, U.S. President Donald Trump joined by phone and called AI doom talk a "scam," while voicing support for data center construction.