Microsoft took the lead earlier this year in the race to lead the new era of artificial intelligence with the million-dollar investment it made in OpenAI, the company that created ChatGPT. A little less than a year later, the Redmond technology giant has delivered another blow to the industry with the announcement of the manufacture of its own chips, specifically designed to respond to the high cost of providing artificial intelligence services. And what is more important if possible, to stop depending on Nvidia, which currently controls 80% of the world chip market for this key technology in the future.
The Azure Maia 100 chip and Cobalt 100 CPU will be available in 2024, the company announced at its Ignite developer conference, held this week in Seattle. Microsoft says it does not plan to put them up for sale, but rather use them to power its own subscription software offerings and as part of its Azure cloud computing service. With this duo of chips, Microsoft aims to “rethink cloud infrastructure for the age of AI and optimize literally every layer of that infrastructure,” according to Rani Borkar, head of Azure hardware systems and infrastructure.
Microsoft's Azure Maia 100 chip is designed to run AI workloads in the cloud, such as training and inference of large language models. It will be used to accelerate AI computing tasks and provide a foundation for its $30 per month "Copilot" service for enterprise software users as well as developers who want to create custom AI services. Also to power some of the company's largest AI workloads on Azure, such as the partnership with OpenAI.
"We believe this will help us deliver better solutions to our customers: faster, lower cost and higher quality," said Scott Guthrie, executive vice president of Microsoft's cloud and artificial intelligence group.
For its part, the Azure Cobalt CPU is a 128-core chip based on an Arm Neoverse CSS design and customized for Microsoft designed to power general services in the Azure cloud and in which “high performance” and “data management” are combined. energy,” according to the company. It is designed to save internal costs and be a response to Microsoft's main rival in the cloud, Amazon Web Services.
“Our initial tests show that our performance is up to 40% higher than our data centers that use commercial Arm servers,” Borkar explained about the first impressions of these chips, although the specifications are unknown at the moment.
In addition, Microsoft has already hinted that it plans to bring new chips to the market and that these are only the first generation of products of the AI era: “It is a series,” Borkar stated, although the outcome remains to be seen. impact of these chips on the market.
During the Microsoft Ignite event, Nvidia CEO Jensen Huang offered his vision of the moment in the technology sector: “Generative AI is the biggest revolution in computing in the last 40 years, bigger than the PC, mobile and Internet". Huang is considered one of the world's gurus in the field of artificial intelligence.
