Microsoft CEO Satya Nadella highlighted the Chinese AI startup DeepSeek’s R1 model, noting it as the first AI competitor he has encountered that rivals OpenAI’s offerings. In an interview with Bloomberg Businessweek, Nadella stated that OpenAI has historically been in a league of its own, and he described DeepSeek’s R1 as having made impressive advancements.
DeepSeek gained attention in January when its free chatbot app topped the US App Store rankings. This success caused concern among investors, as DeepSeek’s competitively priced models led to a decline in AI-related stocks, raising questions about the future demand for AI hardware, particularly Nvidia chips.
Microsoft has begun offering versions of DeepSeek’s R1 model through its Azure AI Foundry platform, which also hosts AI models from other companies like OpenAI and Meta. Using R1 on Microsoft’s platform ensures data is not transmitted to DeepSeek’s servers in China.
Asha Sharma, a corporate vice president at Microsoft, mentioned in a blog post that R1 underwent extensive safety evaluations before its release. OpenAI CEO Sam Altman also praised R1, commenting on its value and indicating that OpenAI plans to respond with enhanced releases due to the invigorating competition from DeepSeek.
However, opinions on DeepSeek’s merits are mixed. Ben Buchanan, a former AI advisor in the Biden administration, remarked that while DeepSeek’s team is talented, the hype surrounding it might be overstated, as their performance improvements rely on algorithms similar to those being developed by established companies.
Microsoft, OpenAI, and DeepSeek have not responded to inquiries from Business Insider regarding these developments.