OpenAI and Broadcom Unveil 'Jalapeño' AI Inference Chip for LLM Deployment
Source: https://www.engadget.com/2201045/openai-broadcom-jalapeno-inference-processor-ai-accelerator/
Executive Briefing
- Reveals Jalapeño as OpenAI's first custom 'Intelligence Processor,' designed specifically to run large language model inference workloads.
- Claims performance per watt substantially exceeds current state-of-the-art chips, though final testing and technical specs remain pending.
- Developed from initial design to manufacturing tape-out in just nine months, accelerated by OpenAI's own AI models during the design process.
- Plans initial data center deployment for late 2026 as part of a broader multi-generation compute platform with Broadcom.
- Marks OpenAI's strategic push to control more of its own hardware stack, aiming to improve efficiency and expand ChatGPT capacity.
Sponsored
