Jensen Huang's Latest 30,000-Word Interview: I'm 63, I Hope to Die Peacefully While Working, Learning Every Day, I Don't Believe in Succession Plans

Big, provocative claims from a CEO at the center of the AI stack

Jensen Huang, founder and CEO of NVIDIA (英伟达), used a marathon conversation with Lex Fridman to make sweeping, sometimes controversial, pronouncements — chief among them that, from an economic perspective, "AGI" has already arrived. He framed the debate not as a philosophical exercise about human-like cognition but as a question of economic agency: AI systems can reportedly run businesses and generate large-scale value without human micro-management. It has been reported that Huang cited examples of AI agents operating online businesses that produced millions in revenue in short order.

Technical roadmap and strategic shifts

Huang laid out what he called four stages of scale — from pretraining expansion to fine-tuning, inference-time scaling and the current wave he dubs "agent expansion" — arguing that agents are the token-era "iPhone moment." He predicted a dramatic rise in people who can "program" — from tens of millions today to a potential billion — and described a fundamental shift in design thinking: the unit of computation is no longer a single GPU but an entire co-designed rack, full stack and data-centre system. That extreme systems-level approach explains NVIDIA's pivot from selling discrete cards to offering integrated PODs and infrastructure.

Management, trust and geopolitics

Huang was equally blunt on management: at 63 he says he keeps learning every day and "doesn't believe in succession plans," a stance he links to the distraction of anxiety rather than immortality. He also described a decades-long relationship with TSMC (台积电) — reportedly involving hundreds of billions of dollars of work over 30 years that grew on trust rather than formal contracts. These remarks land amid rising geopolitical friction: U.S. export controls and sanctions around advanced AI chips have real consequences for Chinese AI firms and for global supply chains that depend on NVIDIA GPUs and Taiwan's foundry capacity. How companies navigate trust-based relationships when policy risk is rising is now a central question for both Western and Chinese tech players.

Why Western and Chinese audiences should pay attention

Huang is not merely bullish; he is shaping infrastructure, markets and the narratives that policymakers respond to. If AGI is framed as economic agency rather than a human-like mind, how should regulators, investors and major cloud customers react? For China — where domestic AI ambitions run hot and access to high-end accelerators is politically sensitive — Huang's ideas and NVIDIA's technical direction matter as much for strategy as they do for chip shipments. Whether one agrees with his conclusions or not, the interview is a reminder that the industry's technical architecture, corporate cultures and geopolitics are now tightly intertwined.