Top Vole chief executive Satya Nadella told Bloomberg Businessweek. “OpenAI has been so far ahead that no one's really come close. DeepSeek, and R1 in particular, was the first model I've seen post some points.”

The Chinese startup turned heads in January when its free chatbot app shot to the top of the US App Store rankings, sending the cocaine nose jobs of Wall Street into a full-blown AI panic. Investors dumped AI stocks, jittery that DeepSeek's cheaper and surprisingly capable models might gut demand for GPU-heavy hardware from firms like Nvidia.

OpenAI boss Sam Altman admitted the competition was getting spicy, calling R1 “an impressive model, particularly around what they're able to deliver for the price.” He said the challenge from DeepSeek was “invigorating” and enough to make OpenAI “pull up some releases.”

Chipzilla and the rest of the hardware crowd likely weren’t thrilled with the news, but Microsoft saw an opportunity. In January, it began offering versions of R1 through its Azure AI Foundry cloud platform, slotting DeepSeek in alongside offerings from OpenAI, Meta, and Mistral.

To dodge any hand-wringing over data sovereignty, Microsoft assured users that running R1 on Azure meant their information wouldn't end up on DeepSeek's servers in China. Microsoft corporate vice president Asha Sharma claimed R1 underwent “rigorous red teaming and safety evaluations” before its launch.

But not everyone's convinced the fuss is justified. Ben Buchanan, a former US government AI advisor, told The Ezra Klein Show in March that DeepSeek's engineers were top notch but not magicians.

“Where do you think they got their performance increases from? We read their papers. They're smart people who are doing exactly the same kind of algorithmic efficiency work that companies like Google and Anthropic and OpenAI are doing,” Buchanan said.