newsfilter
← Interviews
20VC with Harry Stebbings1h 18m

Steeve Morin: Why Google Will Win the AI Arms Race & OpenAI Will Not | E1262

Steeve Morin, Harry Stebbings

Interview, Fireside Chat

Speaker ZML projects a near-term market correction where 95% of AI demand shifts to inference, rendering current NVIDIA hardware inefficient due to high memory transfer costs and inflated pricing. To overcome these bottlenecks, the industry will pivot toward hardware-agnostic architectures, compute-in-memory silicon, and specialized chips designed for low-latency reasoning rather than throughput. This transition promises to dismantle NVIDIA's dominance by enabling hyperscalers like Google and emerging competitors to deploy diverse, cost-effective solutions that prioritize reliability over aggregate compute power.