Microsoft will launch more than 1 million new AI chips in 2027. Why is this a threat to Nvidia?

IT giant Microsoft has unveiled a new AI processor, the Azure Maia 300. The company plans to deploy more than 1 million of these chips by 2027 to support its AI services, according to Reuters. In this way, the company is reducing its dependence on third-party processors—primarily those from Nvidia and AMD. Other major cloud service providers, including Google and Amazon, are also developing their own chips. What does this mean for Nvidia’s business?

Microsoft plans to make its own Maia chips the “primary inference system (for generating AI responses)” for its own products and services. Maia 300 delivers 40% more tokens per megawatt compared to “leading GPU-based systems,” Microsoft claims. Yahoo Finance sees this as a direct jab at Nvidia.

A token is the basic unit of measurement in AI systems, which can represent parts of words, images, and videos. The more tokens per megawatt, the more efficient the AI rack is, Yahoo Finance explained.

A rack of 96 Maia 300 chips delivers seven times the token throughput of a similar configuration based on the previous-generation Maia 200 chips, Microsoft said.

Microsoft shares rose 1.1% at the close of trading on October 12. The company’s market capitalization once again exceeded $4 trillion—for the first time in nearly a year. The company was last valued at this level on October 29, 2025, and it reached its all-time high of $4.029 trillion the day before, MarketWatch noted. Microsoft ranks fourth in terms of market capitalization in the U.S., trailing only Nvidia, Apple, and Amazon.

Microsoft is one of the companies that are developing or already using their own AI chips. Others include cloud giants Amazon, Google, and Meta, as well as AI labs OpenAI and Anthropic. Having their own processors allows them to reduce their dependence on Nvidia and AMD and tailor the hardware to specific tasks.

This is a potential problem for Nvidia: the company regularly derives more than 50% of its quarterly revenue from cloud giants, according to Yahoo Finance. Moreover, Amazon and Google are already leasing their chips to third-party customers, which directly threatens the business of the world’s most valuable company. Amazon is also considering selling chips to customers for installation in their own data centers, rather than just leasing them out at its own facilities, Yahoo Finance noted.

On the other hand, this does not mean that cloud companies will completely stop buying from Nvidia, the publication added. Customers still need different types of chips for different workloads, which means Amazon, Microsoft, and other hyperscalers will likely continue to purchase the most in-demand processors, it concluded.

Maia 200, the second-generation chip, was unveiled by Microsoft in January 2026, with the company calling it “the most powerful in-house processor” among hyperscale solutions—it compared it to Amazon’s third-generation Trainium chips and Google’s seventh-generation TPUs.

As The Information reported in August, Microsoft was in talks with TSMC to reserve production capacity for more than 300,000 Maia 300 chips, with delivery scheduled for 2027, and plans to reserve capacity for more than 1 million of these processors in the future. According to Bloomberg, Microsoft lags behind competitors such as Alphabet and Amazon, whose proprietary chips are already widely used to power AI services.

This article was AI-translated and verified by a human editor