Microsoft ( MSFT ) is taking aim at cloud rivals Amazon ( AMZN ) and Google ( GOOG , GOOGL ) with the debut of its next-generation custom AI chip.
Called Maia 200, the chip will run in Microsoft’s own data centers before the company eventually makes it available to its wider customer base.
Like Google’s TPUs and Amazon’s Trainium processors, Microsoft’s second AI chip is meant to give the Windows maker more flexibility when it comes to how to power its AI services. By using its own internally developed chips, the company ensures that it does not have to rely solely on processors developed by Nvidia (NVDA) or AMD (AMD).
Google and Amazon have been using their own custom chips for years, while Microsoft has been slower to adopt in-house AI silicon.
According to Microsoft, the Maia 200 will be built using TSMC’s 3-nanometer process and is designed to run large-scale AI workloads, while “delivering efficient performance for every dollar.”
The Maia 200 will be built into large server racks with trays that house four chips each. Microsoft is also touting how soon it can deploy the new chips in data centers, saying the chips are installed and running AI models within days of the parts arriving.
Getting AI servers up and running quickly is an important aspect of the broader data center business. It’s not just a matter of keeping construction costs down, either. The longer a chip sits idle, the less cash it can generate for the company by running AI apps.
The Maia 200 adds to the growing competition Nvidia is facing from both AMD and its own customers. Microsoft’s Maia 100 already runs both the company’s and OpenAI’s (OPAI.PVT) AI models, while Google and Amazon both run their respective models and the Anthropic (ANTH.PVT) models.
And in November, The Information reported that Meta was talking to Google about using the search giant’s TPUs in its own data centers to run its AI services. That sent Nvidia stock down at the time, as Wall Street raised fears that the company was in danger of losing market share.
Nvidia’s stock price has risen less than 1% since the beginning of the year.
Despite incursions into Nvidia’s lane, Google, Amazon, and Microsoft are unlikely to pose a serious threat to the AI leader. Experts say that while the cloud company’s AI chips may work well for their own services, that’s unlikely to translate to smaller third-party clients as easily.
Nvidia’s chips are also highly desirable because they are designed to be multipurpose, allowing companies to use them for a litany of applications and services.
In terms of performance, the Maia 200 won’t dethrone Nvidia, but Microsoft claims it beats both Google’s latest TPU and Amazon’s newest Trainium chip in a number of categories. The Maia 200 also packs more high-bandwidth memory than Google or Amazon’s offerings, which is essential for running high-powered AI applications.