Jensen Huang on B2G

In this podcast transcript, NVIDIA CEO Jensen Huang discusses the explosive growth of AI with hosts, framing it as a new industrial revolution. He highlights NVIDIA's massive partnerships, like investing $100 billion in OpenAI's Stargate project to build 10 gigawatts of AI data centers—potentially generating $400 billion in NVIDIA revenue—positioning OpenAI as the next multi-trillion-dollar "hyperscaler" (like Google or Meta, but focused on AI services for consumers and businesses).

Huang explains AI's three scaling laws: pre-training (teaching AI basics), post-training (practicing skills via reinforcement learning), and inference (thinking before answering, which boosts compute needs by a billion times). For example, instead of a quick AI response like "8x8=64," modern AI researches, reasons, and verifies, multiplying processing power like upgrading from a bicycle to a jet engine.

He addresses market skepticism, predicting NVIDIA's revenue could hit trillions by replacing outdated computing with AI-accelerated systems, augmenting global GDP (e.g., a $100,000 employee boosted by a $10,000 AI becomes 2-3x more productive, like giving every worker a super-smart assistant). No glut or bubble, he says, as demand from hyperscalers, sovereign nations, and everyday apps (like personalized Meta content) outpaces supply.

On competition, Huang touts NVIDIA's annual chip releases (e.g., Blackwell's 30x Hopper improvement) via "extreme co-design" (optimizing chips, software, and data centers together, like tuning a race car from engine to tires). ASICs from rivals like Google may work for niche tasks but can't match NVIDIA's flexible "AI factories." Even free rival chips lose out due to higher energy costs—imagine two farms with the same land; one yields twice the crops per acre, so you'd pick it despite upfront costs.

Globally, he stresses sovereign AI (countries building their own infrastructure for cultural/national models) and urges U.S.-China competition without decoupling, warning U.S. export bans boosted Huawei. He praises the Trump administration's pro-growth stance, energy focus, and immigration tweaks (e.g., $100k H1B fees as a start to curb abuse while attracting talent).

Looking ahead, AI will create abundance: personal robots like R2-D2 companions, digital health twins predicting diseases, and job shifts (tasks automated, but more ideas/jobs created, like how electricity displaced lanterns but sparked new industries). Huang ends optimistically: get on the AI train now to ride the exponential growth.

Link: https://youtu.be/pE6sw_E9Gh0?si=W0MG4Z3SxzUAmPGd