OpenAI's CFO Explains Shift Away from Exclusive Reliance on Nvidia
Sarah Friar, Chief Financial Officer of OpenAI, confirmed during an interview with CNBC’s Jim Cramer that the company is diversifying its chip suppliers. This marks a shift in strategy where NVIDIA (NASDAQ:NVDA), once considered the default provider for AI training chips, now competes among several vendors. While OpenAI remains committed to purchasing large amounts of NVIDIA hardware—up to 12 gigawatts through 2030—the change in negotiation dynamics could pressure Nvidia’s pricing power and margins. Friar highlighted a growing distinction between two types of chip workloads: training, where general-purpose silicon like NVIDIA's excels, and inference, which benefits from specialized chips tailored for specific models. OpenAI has already developed its own purpose-built inference chip called Jalapeno to improve efficiency in this area. The company’s massive $122 billion cash reserve further reduces the urgency or dependency that typically allows suppliers to maintain premium pricing. The market reacted swiftly, with Broadcom (AVGO) dropping 7.95% and AMD (NASDAQ:AMD) remaining relatively flat as investors assessed their exposure to OpenAI's evolving procurement strategy.
Topics: CFO commentary chip supply chain diversification NVIDIA OpenAI