Scalable AI requires more intelligence for every single watt used and every single dollar spent. On the Lex Fridman podcast, NVIDIA CEO Jensen Huang explains why extreme co-design is essential for AI efficiency. Watch the full episode now:

@lexfridman Focusing on the cost per dollar spent is exactly what will allow AI to scale globally. This deep integration between hardware and software is a game changer for the industry.

@lexfridman This episode crystallized something I've been thinking about. Jensen's relentless focus on efficiency per dollar spent is the same mindset we need for agent transactions.

@lexfridman Efficiency isn’t just about raw compute. It’s about how smartly you use it. Co-design like this is what makes AI not just powerful, but sustainable and practical at scale

@lexfridman Hey @Grok Explain Please

@lexfridman

Le truc qui me frappe c'est que Jensen il dit pas juste "achetez nos GPU". Il explique que la course à la puissance brute c'est fini, faut être malin avec chaque watt. Et c'est exactement pour ça que personne arrive à rattraper Nvidia, ils pensent pas hardware OU software, ils pensent tout en même temps. Ça paraît évident dit comme ça mais y'a quasi personne d'autre qui le fait vraiment

@lexfridman Scaling is shifting from bigger models to better systems

@lexfridman @NVIDIAAI explain why your artificial intelligence is causing prices on digital devices to go up? Stop the price hikes on digital devices.

Excellent point from Jensen on the Lex Fridman podcast. True scalable AI isn’t just about more compute, it’s about getting far more intelligence per dollar. That’s exactly why capital efficiency matters so much when building out GPU clusters for pre-training, post-training, test-time scaling, or agentic AI. We help teams solve that part at GPU financing: flexible GPU Financing, preserve runway, quick access to hardware, no big capex hits. What’s the toughest part of scaling for you right now, cost, speed, or availability?

@lexfridman Can’t wait to listen to the full podcast!

@lexfridman Agree, cost and energy efficiency drive AI scale, and hardware software co design is essential

@lexfridman Efficiency isn’t optional anymore—AI has to be smart and lean. Jensen breaks down how co-design makes every watt and dollar count

@lexfridman $QMEOW $PVS $LAIR

@lexfridman @NVIDIAAI Cancel artificial intelligence, explain why your artificial intelligence is causing prices on digital devices to go up? Stop the price hikes on digital devices.

@lexfridman The per watt part is the real constraint. I keep seeing model quality matter less than whether a team can actually afford to run it.