OpenAI is actively working to reduce its reliance on Nvidia hardware by diversifying its compute infrastructure. The company has announced a significant compute deal with AMD, aiming to secure 6 gigawatts of capacity, which is substantially larger than its current Texas data center. As part of this strategy, OpenAI has revealed TensorWave as its first neocloud partner associated with the AMD deal, which will provide 457 megawatts of capacity. This move comes as other AI leaders like Anthropic have already established a more diversified chip base, utilizing Google TPUs, AMD GPUs, Amazon Trainium GPUs, and Nvidia GPUs. AI
IMPACT Diversifying compute infrastructure reduces reliance on single vendors, potentially stabilizing costs and ensuring continued model development.
RANK_REASON Major compute infrastructure deal between leading AI companies and hardware providers. [lever_c_demoted from significant: ic=1 ai=0.7]
- AMD
- Andrew Ng
- Anthropic
- Demis Hassabis
- Fei-Fei Li
- Greg Brockman
- Isomorphic Labs
- Jensen Huang
- Microsoft
- Mustafa Süleyman
- Nikesh Arora
- Nvidia
- OpenAI
- Palo Alto Networks
- Reid Hoffman
- Sam Altman
- TensorWave
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →