OpenAI and Broadcom unveil 'Jalapeño' Intelligence Processor for LLM inference
AI disclosure
Summary
<p>"A blank-slate design for modern LLM inference, not a general-purpose accelerator adapted from earlier AI workloads"</p>