Custom AI chips emerge as threat to Nvidia's market dominance

Nvidia's stranglehold on AI infrastructure is loosening. The semiconductor giant has long dominated the market for chips powering machine learning, but a wave of major tech companies are now charting their own course with custom silicon.

OpenAI recently announced Jalapeño, a custom inference chip built with Broadcom, joining Google, Apple, and SpaceX in developing proprietary hardware. The strategy isn't about abandoning Nvidia entirely—it's about reducing single-supplier risk and unlocking performance gains. Custom chips enable companies to optimize hardware for their specific workloads while gaining greater control over their supply chains.

The shift echoes Apple's transition away from Intel, demonstrating how vertical integration and bespoke silicon can deliver meaningful competitive advantages. As these companies invest in custom chips, the AI hardware landscape is becoming increasingly fragmented, challenging Nvidia's near-monopoly position in a market it has dominated for years.

TechCrunch's Equity podcast explores this trend and its implications for the industry. Subscribe on YouTube, Apple Podcasts, Spotify, and other platforms, or follow @EquityPod on X and Threads.