ADVERTISEMENT
Nvidia has announced plans to invest as much as $100 billion in OpenAI, signing a letter of intent to supply the ChatGPT maker with vast computing infrastructure for its next wave of artificial intelligence models.
The deal, revealed on Monday, would see Nvidia deliver systems capable of deploying 10 gigawatts of computing power - enough energy to serve millions of households — across new data centres dedicated to training and running OpenAI’s models.
The move comes as OpenAI seeks to diversify its infrastructure base and reduce its dependence on Microsoft, which remains its largest backer and cloud provider. In January, Microsoft restructured its agreement with OpenAI, giving the startup scope to develop AI infrastructure with other partners. Since then, OpenAI has expanded its roster of collaborators, including Oracle, SoftBank and the $500 billion Stargate project.
Nvidia said the deal is intended to complement these arrangements, positioning the chipmaker as OpenAI’s “preferred strategic compute and networking partner” as it builds out its so-called AI factories.
It remains unclear whether Nvidia’s commitment will be delivered in cash, chips, cloud credits, or a combination thereof, but the partnership underscores the increasingly intertwined fortunes of the world’s most valuable chipmaker and one of the leading AI companies.