Nvidia CEO Jensen Huang announced that the company’s AI chips are advancing faster than Moore’s Law, during a keynote at CES 2025. He stated that Nvidia's latest datacenter superchip is over 30 times faster than its predecessor for AI inference workloads. Huang believes this performance leap will reduce costs associated with AI inference, countering concerns about the affordability of advanced AI models. He highlighted three active AI scaling laws shaping the future of AI. Nvidia’s chips, he claims, are now 1,000 times better than a decade ago, indicating significant ongoing innovation.