Microsoft CTO: Hopes to primarily use self-developed AI data center chips and independently design data center systems in the future.

```

Microsoft’s Chief Technology Officer said on Wednesday that the company hopes its data centers will primarily use its own chips in the future, a move that could reduce its reliance on major chipmakers such as Nvidia and AMD.

Semiconductors and the servers inside data centers are the core foundation supporting the development of artificial intelligence models and applications. Nvidia currently dominates this sector, holding a commanding lead in GPU market share, while its competitor AMD holds a smaller portion.

However, major cloud computing companies, including Microsoft, have also started designing custom chips for their data centers.

Microsoft Chief Technology Officer Kevin Scott discussed the company’s AI chip strategy in a fireside chat with the media at Italian Tech Week. Currently, Microsoft mainly uses Nvidia’s and AMD’s chips in its own data centers. He told the media that the company has always focused on choosing "the best price-to-performance ratio offered by each chip."

"We are not committed to using any particular chip... This means that for many years, the best price-to-performance solution has always been Nvidia."

"As long as we can ensure that we have enough capacity to meet demand, we are willing to consider any solution."

At the same time, Microsoft has already begun using its self-developed chips.

In 2023, Microsoft launched the Azure Maia AI accelerator chip, designed for AI workloads, as well as the Cobalt CPU. In addition, it was reported that the company is also developing next-generation semiconductor products. Last week, Microsoft also announced a new cooling technology called “microfluids” aimed at solving the problem of chip overheating.

When asked whether Microsoft plans for its data centers to primarily use Microsoft chips in the future, Scott replied: "Absolutely." He added that the company is already using Microsoft’s chips extensively.

Scott said that Microsoft’s focus on chip development is to eventually be able to design a complete set of data center systems.

"It’s not just about the chip, but the entire system design—including networking and cooling systems. You must have enough freedom to make the decisions you need to truly match computing resources to workloads."

Microsoft, as well as competitors Google and Amazon, are all developing their own chips—not only to reduce reliance on Nvidia and AMD, but also to make their products more efficient in meeting their own specific needs.

Computing Power Still in Short Supply

Tech giants including Meta, Amazon, Alphabet, and Microsoft have pledged over $300 billion in capital expenditures this year, with most of it invested in artificial intelligence to address the surging AI demand.

Scott noted that there is still a shortage of computing power.

"Calling this a massive crunch is still an understatement. I think ever since the launch of ChatGPT, we’ve been in a situation where it’s almost impossible to rapidly scale up capacity."

Microsoft has been expanding capacity by building data centers, but Scott warned that this is still not enough to meet demand.

"Even our most aggressive projections often prove not enough... Over the past year, we’ve deployed an astonishing amount of computing resources, and there will be much more in the coming years."

Risk Warning and DisclaimerThe market entails risks and investments should be made with caution. This article does not constitute personal investment advice and has not taken into account the special investment objectives, financial situation, or needs of any individual user. Users should consider whether any opinions, views, or conclusions expressed herein are suitable for their particular situation. Any investment accordingly is at your own risk. ```