We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·Oct 6·3 min read
AMD and OpenAI announce strategic partnership to deploy 6 gigawatts of AMD GPUs
TL;DR
OpenAI faced capacity constraints in GPU infrastructure needed to train and run large language models at scale. AMD and OpenAI announced a multi-year partnership to deploy 6 gigawatts of AMD Instinct GPUs, with 1 gigawatt operational by 2026.
✦ Why It Matters
Engineers can plan infrastructure diversification strategies and understand AMD Instinct GPU capabilities for AI workload optimization.
Key Takeaways
- Plan GPU procurement strategies around AMD Instinct availability starting 2026.
- Evaluate AMD Instinct performance characteristics for your model training workloads.
- Consider infrastructure diversification beyond single GPU vendor dependencies.
- Monitor phased deployment timeline to align capacity planning with 2026+ timelines.
Full Summary
Large language model development requires massive computational resources; a gigawatt equals one billion watts of power consumption. OpenAI and AMD established a strategic partnership to address infrastructure scaling challenges by deploying AMD Instinct GPUs—specialized processors designed for AI workloads.
The agreement spans multiple years with phased rollout: 1 gigawatt beginning in 2026, scaling to 6 gigawatts total. This represents a substantial commitment to alternative GPU suppliers beyond traditional vendors, diversifying the AI infrastructure supply chain.
The deployment will power OpenAI's next-generation models and infrastructure, enabling faster training cycles and improved inference capabilities. This partnership signals industry-wide recognition that meeting AI compute demands requires coordinated hardware-software partnerships and expanded manufacturing capacity.