In a recent interview with Bloomberg Businessweek, Microsoft CEO Satya Nadella expressed his surprise at the capabilities of DeepSeek's R1 model, noting it as the first serious challenger to OpenAI's ChatGPT. Nadella stated, "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."
DeepSeek, a Chinese AI startup founded in July 2023 by Liang Wenfeng, has rapidly emerged as a formidable player in the AI landscape. The company's flagship model, DeepSeek-R1, launched in January 2025, delivers performance comparable to leading models like OpenAI's GPT-4o and o1 but at a fraction of the cost. Reports indicate that DeepSeek trained its V3 model, a predecessor to R1, for approximately $6 million, significantly less than the estimated $100 million cost for OpenAI's GPT-4 in 2023. (en.wikipedia.org)
The cost-effectiveness of DeepSeek's models is attributed to their innovative architecture. DeepSeek-R1 employs a Mixture-of-Experts (MoE) approach, utilizing 671 billion parameters but activating only 37 billion per query, optimizing computational efficiency. This design enables faster response times and reduced energy consumption compared to traditional dense models. (geeksforgeeks.org)
The affordability and performance of DeepSeek-R1 have led to its rapid adoption. Shortly after its release, the model surpassed ChatGPT as the most downloaded free app on the U.S. Apple App Store. (reuters.com) This success prompted Microsoft to integrate R1 into its Azure cloud services, offering customers a cost-effective alternative to existing AI models. Nadella emphasized the significance of this development, stating, "When you have something that good and that cheap, you don’t ignore it. You adapt."
The rise of DeepSeek has also influenced other tech leaders. OpenAI CEO Sam Altman acknowledged R1's impact, calling it "an impressive model" that motivated his team to accelerate their releases. (indianexpress.com) Similarly, Meta's chief AI scientist, Yann LeCun, highlighted the benefits of open research, noting that DeepSeek's advancements demonstrate the power of open-source models. (indianexpress.com)
DeepSeek's emergence underscores a shift in the AI industry towards more efficient and accessible models. Nadella drew parallels to historical technological advancements, suggesting that as AI becomes more efficient, its use will skyrocket, turning it into a ubiquitous commodity. (windowscentral.com) This development challenges established AI leaders to innovate continually and adapt to a rapidly evolving landscape.
In summary, DeepSeek's R1 model has not only surprised industry leaders like Satya Nadella but has also set a new benchmark for cost-effective AI performance, prompting a reevaluation of strategies among major tech companies.
Source: India Today Satya Nadella says DeepSeek R1 surprised him by coming close to ChatGPT
