According to a report from The Information, Microsoft has been developing its own artificial intelligence (AI) chips to handle the growing expenses of in-house and OpenAI development projects. The hardware venture has reportedly been in progress since 2019, with the aim of reducing Microsoft’s reliance on Nvidia’s GPUs. The Google search engine reveals that the Nvidia H100, one of the more popular GPUs used for machine learning system training, is selling for up to $40,000 on reseller services such as eBay due to increasing market scarcity.
The high cost of GPUs has prompted several big tech firms to design their own hardware, with Google, Amazon, and Meta all developing machine learning chips in recent years. Microsoft’s chips are rumored to be named “Athena,” according to The Information, although the company has not officially commented on the project. The generative AI arms race continues to heat up, and the development of AI chips is crucial for companies to meet the increasing computational demands of their AI applications.