Microsoft higher-ups are taking a page from DeepSeek's overnight success' book
Microsoft execs are impressed
2 min. read
Published on
Share this article
Improve this guide
Read our disclosure page to find out how can you help MSPoweruser sustain the editorial team Read more
Key notes
- DeepSeek’s AI success, built for just $6 million, has caught Microsoft’s attention for its cost-effectiveness and efficiency.
- Microsoft executives, including Satya Nadella and Mustafa Suleyman, see the potential for rapid technology adoption.
- DeepSeek’s rise has shaken the market, with questions over its development process and the possible disruption of the AI industry.
DeepSeek stole the spotlight in the AI race as an AI product that’s built at a relatively low cost of just $6 million. And now, Microsoft executives are ready to take notes and even take a page from its book.
Satya Nadella, Microsoft’s CEO, commented on DeepSeek’s growing success and how impressive the models’ cost-effectiveness and efficiency are. The Chinese company said that it used Nvidia H100 GPUs to build the models.
Nadella is comparing the phenomenon to the Jevons paradox, an idea that as technology makes the use of a resource more efficient, the overall consumption of that resource can actually increase instead of decrease.
Mustafa Suleyman, Microsoft AI’s leader, weighed in, saying that humankind is “learning the same lesson” on technology.
“Everything of value gets cheaper and easier to use, so it spreads far and wide. It’s one thing to say this, and another to see it unfold at warp speed and epic scale, week after week,” the former Inflection AI’s higher-up says in a post on X.
DeepSeek’s launch creates some sense of urgency in the US market on how an AI is made. It shakes the market up as Nvidia, a leading AI player, reported a stock price drop by nearly 17%, alongside declines in Microsoft, Google, and other tech firms.
But, its $6 million claim is not without eyebrows raising.
There are at least three possible scenarios that happened behind DeepSeek’s development, as analyst Charles Archer says in his blog post (it’s a good read, by the way). The Chinese AI company either reverse-engineered open-source models from Meta, used low-cost Nvidia chips funded by China or found an innovative way to scale AI with older hardware.
The first two do seem more likely, but the third scenario, which could disrupt the AI boom like never before, would be something that’s unheard of yet revolutionary. Because, if a Chinese company can create an AI model with that cost, imagine what $500 million or $1 billion funding can do in the American market.
But, which one is it? Only time will tell.
User forum
0 messages