AI BriefingOpenAI and Broadcom Announce Strategic Collaboration to Deploy 10GW of Custom-Designed AI Accelerators
Key point
OpenAI and Broadcom have partnered to jointly develop and deploy 10GW of custom-designed AI accelerators and network systems.
Details
OpenAI and Broadcom have signed a multi-year partnership to jointly develop and deploy 10GW of AI accelerators and network systems. OpenAI will design its own AI accelerators and systems, directly incorporating insights gained from developing frontier models into the hardware to maximize performance.
Broadcom will provide Ethernet solutions and network systems supporting both scale-up and scale-out. The infrastructure built through this collaboration will be deployed at OpenAI's own facilities and partner data centers to meet surging global AI demand.
Key plans are as follows:
- Deployment schedule: Kicks off in the second half of 2026, targeting completion by the end of 2029
- Technical composition: Rack systems based on OpenAI-designed accelerators and Broadcom's Ethernet, PCIe, and optical connectivity solutions
- Core goal: Building open, scalable, and power-efficient infrastructure for next-generation AI clusters
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.
Related
OpenAI Reveals AI Accelerator Chip 'Jalapeño' Designed Using Its Own LLMOpenAI Discloses AI Agents Accessed US Government Websites During Training
OpenAI Prepares New $500/Month ChatGPT Pro Max Plan with 'Fastest' Inference