DeepSeek AI’s recent rise has dramatically impacted global tech stocks, especially as the company launched its flagship AI model, R1, which rivals established leaders like OpenAI and Google.
US tech stocks suffered significant sell-offs with reports of DeepSeek’s remarkable efficiency and affordability coming to light. Just days after its debut, DeepSeek’s R1 model quickly climbed the charts, dethroning ChatGPT as the most downloaded app on Apple’s Top Free Apps. This unprecedented success of the Chinese startup, founded by quant hedge fund manager Liang Wenfeng, sent shockwaves through the tech market.
On January 29, 2025, Nvidia saw its shares plummet by over 17%, resulting in losses of hundreds of billions from its market capitalization. This massive decline raised eyebrows and concerns about the future demand for advanced hardware traditionally required for training large AI models.
Microsoft CEO Satya Nadella commented on the situation, stating on LinkedIn, “Jevons paradox is at play again,” referring to the economic theory where increased efficiency typically leads to greater demand. He emphasized, “as AI becomes more efficient and accessible, its adoption will soar, transforming it to become indispensable.” During the World Economic Forum, Nadella warned other tech leaders to take DeepSeek's developments seriously, highlighting the potential shifts within the industry.
Not everyone shared Nadella’s concern, though. Pat Gelsinger, the former CEO of Intel, challenged the market’s reaction, arguing, “the market is getting it wrong.” He suggested instead of fearing competition, the advent of cheaper AI technologies could expand the market, enabling broader applications. Gelsinger credited DeepSeek’s innovative engineers, noting their ability to achieve such advancements with limited resources.
Marc Andreessen, co-founder of Andreessen Horowitz, praised DeepSeek’s R1 model, calling it “one of the most amazing and impressive breakthroughs” and referring to it as “AI’s Sputnik moment”. His remarks underscored the competitive ingenuity of DeepSeek and its potential to change the dynamics of the AI industry.
Yann LeCun, Chief AI Scientist at Meta, offered a different perspective, arguing the narrative surrounding DeepSeek’s advancements could be misleading. He contended, “open source models are surpassing proprietary ones,” and the hype around cheap models might be overstated. LeCun highlighted the increasing costs of inference, cautioning against the oversimplification of DeepSeek’s technological achievements, stating, “once you put video-understanding, reasoning, large-scale memory, and other capabilities, inference costs are going to increase.”
On the other hand, Mark Zuckerberg hinted at Meta’s upcoming advancements. While he didn’t directly comment on DeepSeek, he announced plans for a new version of Facebook’s Llama AI model, pledging not just to match but to lead the state of AI tech. With over 1.3 million GPUs slated to be allocated by year’s end and projected investments nearing $65 billion, Zuckerberg emphasized Meta's commitment to lead the charge.
Nvidia also acknowledged the impact of DeepSeek’s innovations. A company spokesperson classified it as “an excellent AI advancement” showcasing how to effectively utilize existing technology within legal frameworks. Jensen Huang, Nvidia’s CEO, has yet to publicly address the developments, as market worries build around Nvidia’s competitive position.
Sam Altman, head of OpenAI, described DeepSeek’s model as “impressive,” particularly noting its advantageous pricing. He added, “more compute is more important now than ever,” hinting at the sustained need for innovation within the rapidly changing AI sector.
Altman’s outlook remained optimistic, stating, “we will obviously deliver much more superior models” and expressing excitement about the emergence of DeepSeek as competition. This sentiment reflects the broader industry acknowledgment of the shifting dynamics and the necessity for constant innovation amid growing external pressures.
DeepSeek's rise signals significant changes for the global tech industry, making traditional competitive techniques and business models increasingly untenable. With numerous key players expressing diverse viewpoints, it remains to be seen how the overall tech market adapts to this new paradigm, where affordability and efficiency may become the defining features of AI advancement.