Microsoft CEO Satya Nadella has declared DeepSeek’s R1 artificial intelligence model as the first to truly compete with OpenAI’s models—a statement that, though seemingly focused on a single startup, has inadvertently belittled the progress made by other prominent players in the AI race, including Google’s Gemini, Meta’s Llama, and Elon Musk’s Grok.
In an interview published Thursday by Bloomberg Businessweek, Nadella said: “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 phrasing leaves little ambiguity. Despite the billions of dollars invested in the AI models built by major U.S. tech firms, Nadella has reserved praise only for a Chinese challenger, suggesting that other contenders have yet to meet the bar set by OpenAI.
Register for Tekedia Mini-MBA edition 17 (June 9 – Sept 6, 2025) today for early bird discounts. Do annual for access to Blucera.com.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register to become a better CEO or Director with Tekedia CEO & Director Program.
The R1 model, developed by Beijing-based DeepSeek, has stirred significant attention since its release, particularly after its free chatbot unexpectedly shot to the top of the U.S. Apple App Store in January. That surge sent a ripple through the AI investment ecosystem, triggering a sell-off in shares of companies tied to AI infrastructure, especially chipmakers like Nvidia, amid fears that more affordable models could weaken demand for high-end hardware.
Nadella’s statement, though intended as an endorsement of DeepSeek, is being interpreted across the industry as a blunt demotion of the progress made by rivals like Google’s Gemini, Meta’s Llama, and xAI’s Grok.
Microsoft is a major investor in OpenAI and has been seen as a close strategic ally. But its decision to not only host DeepSeek’s R1 on its Azure AI Foundry but also publicly elevate it above all others, except OpenAI, adds a new dimension to the competitive narrative.
DeepSeek’s rise also has geopolitical undertones. The company is based in China, and while Microsoft offers R1 through its cloud platform, it has taken precautions to avoid sending data to DeepSeek’s domestic servers.
“Using R1 on Microsoft’s platform means that data would not be sent to DeepSeek’s servers in China,” the company emphasized in its communications.
Microsoft’s AI platform head, Asha Sharma, noted in January that R1 underwent “rigorous red teaming and safety evaluations” before being offered to customers on Azure. This assurance was likely aimed at addressing concerns about data security and model safety, especially given the sensitive nature of deploying foreign-developed AI models within enterprise environments.
OpenAI CEO Sam Altman, while acknowledging the rising competition, also commended DeepSeek’s achievement. In January, Altman called R1 “an impressive model, particularly around what they’re able to deliver for the price,” and added that DeepSeek’s progress was “invigorating,” prompting OpenAI to accelerate some of its planned releases.
But others have been more cautious in their appraisal. Ben Buchanan, a former special advisor for AI in the Biden administration, said during a March episode of The Ezra Klein Show that R1’s development wasn’t as groundbreaking as headlines might suggest. “R1 is actually not that unusual,” Buchanan said, explaining that while DeepSeek’s engineers are “extremely talented,” their breakthroughs were largely the result of the same algorithmic efficiency efforts already being pursued by other AI labs, such as those at Google, Anthropic, and OpenAI.
Still, it’s Nadella’s comment that has now drawn the clearest line in the sand.
By placing DeepSeek’s R1 in a class of its own—just beneath OpenAI’s flagship models and above the likes of Gemini and Llama—the Microsoft boss has effectively reshuffled the perceived hierarchy of AI powerhouses. Google’s Gemini, launched with fanfare as the successor to Bard, was billed as a “GPT-4 killer” in some early tech coverage. Meta’s Llama models, particularly Llama 2 and the more recent Llama 3, have been positioned as open-source alternatives intended to democratize access to advanced AI. And Elon Musk’s Grok, developed under xAI and integrated into X (formerly Twitter), has been marketed as an edgy, real-time AI personality capable of competing with leading models.
Yet none of these have received the kind of affirmation Nadella has now offered DeepSeek. His remark subtly implies that, from the vantage point of performance and potential, none of the other major players have managed to stand out.
R1’s rise also presents a complicated narrative for U.S. tech firms that have emphasized AI sovereignty and control. While many Western nations have expressed concern over China’s push to dominate AI, Microsoft’s own endorsement—and decision to integrate R1—underscores the reality that innovation is not confined by geography.
DeepSeek’s R1 is now available via Microsoft’s Azure AI Foundry and GitHub, giving developers and enterprise clients direct access to the model within Microsoft’s cloud infrastructure. That also means DeepSeek can scale globally without the geopolitical friction associated with routing data back to China, at least not through Microsoft’s ecosystem.
For Microsoft, the move is pragmatic. The company is hedging its position: doubling down on OpenAI, while also opening its platform to competitive models that offer value. For the rest of the industry, particularly Alphabet, Meta, and xAI, Nadella’s public endorsement of a Chinese model may sting more than expected.