Custom chips shift OpenAI and SpaceX off Nvidia
OpenAI's Jalapeño chip with Broadcom targets inference latency as firms reduce single-supplier risk in the high-end AI market.
Priya Nair
11 min read