Tech

AI Arms Race Heats Up: Microsoft and OpenAI Plan Massive Data Center “Stargate”

Artificial intelligence, there is super news again.

Foreign media reported that executives from Microsoft and OpenAI are planning to spend hundreds of billions of dollars to build an unprecedented data center, Stargate. “Stargate”

The cost of building this special data center is 100 times that of existing large-scale data centers, and it requires millions of dedicated server chips: the purpose is very clear, to provide OpenAI with powerful computing power support, and the funds are still paid by Microsoft.

I took a look at the relevant information. On March 27, Microsoft engineers revealed that when deploying 100,000 H100 training clusters for GPT-6, the power grid collapsed. Now it seems that these 100,000 card clusters are very Possibly an early prototype of Stargate.

This stargate will become OpenAI’s most powerful engine.

Microsoft has previously invested more than 13 billion US dollars in openAI, and also allowed openAI to use Microsoft data centers to run ChatGPT. In return, openAI opened up access to the core technology to Microsoft.

The reason why Microsoft dares to spend so much money is that it is very confident in openAI and the prospects of artificial intelligence.
According to Sam Altman, the main obstacle restricting the development of GPT-5 is the lack of sufficient computing power.

He has complained more than once about the insufficient number of server chips, and has been secretly comparing the number of chips between himself and Google. He boasted that as long as all the resources he wants are met, he will definitely achieve the ultimate goal, artificial general intelligence (AGI), which is everything. Competent advanced artificial intelligence.

What Microsoft and OpenAI are doing is to firmly occupy a dominant position in this market that will subvert the future.

In fact, Xiaomi’s car sales have been so popular in the past two days, but I am not so excited to check out Xiaomi Group’s stocks. There is a reason.

For a car company, success can only be achieved by building good cars and selling them well.

But for Xiaomi, making cars well and selling them well is not enough. Because Xiaomi is not a car company, but a technology company.
AI is a more important theme of the times for technology companies.

Xiaomi can create a complete ecology of people, homes, and cars. Other technology companies can build cars. Why haven’t they taken this path?

In the era of new energy vehicles, domestic parts manufacturers are extremely powerful. Anyone can save a car as long as they spend money.
Apple stopped making cars, Huawei’s car BU expanded its camp again and again, and Baidu invested 35% in Jiyue, but they did not bet on the manufacturing side.

The United States leads the world not with Detroit, but with Silicon Valley.

When Lei Jun started building cars three years ago, Tesla was at its peak. In just three months this year, Tesla’s stock price has fallen by 30%, while Nvidia’s stock price has risen by more than 80%. The change of times comes so unprepared.

In the era of ALL in AI, Xiaomi’s success in building cars may not necessarily be the success of Xiaomi’s stock price.

error: Content is protected !!