DeepSeek launches, shaking AI industry

DeepSeek, a Chinese artificial intelligence company, has shaken the AI industry with claims that it developed a ChatGPT rival for just $5.6 million. This news sent shockwaves through the stock market, with Nvidia, a major player in the AI revolution, experiencing its worst daily percentage loss since March 2020.

deepseek's r1 is an impressive model, particularly around what they're able to deliver for the price.

we will obviously deliver much better models and also it's legit invigorating to have a new competitor! we will pull up some releases.

— Sam Altman (@sama) January 28, 2025

Nvidia’s stock plunged 17% by the close on Monday, losing $589 billion in market capitalization. This is the single largest, one-day value wipeout of any company, surpassing all previous records.

People don’t talk enough about a giant DeepSeek achievement over most US models – it actually has a reasonable name.

— Ethan Mollick (@emollick) January 30, 2025

The slide knocked Nvidia from its position as the world’s most valuable company, sending its valuation from $3.5 trillion to $2.9 trillion. The release of DeepSeek’s large-language model has shaken confidence in U.S. dominance in generative AI.

Another odd aspect of the sudden attention to DeepSeek is a lot of people assuming it is the first open weights* models. One of the biggest companies in the US has spent billions making open models & intends to keep doing so.

*not open source, there is no access to training data

— Ethan Mollick (@emollick) January 29, 2025

If big U.S. tech companies can learn from DeepSeek to design AI systems with cheaper GPUs, it could impact Nvidia’s profits, which have soared in recent years due to demand for its GPUs from American tech giants. Snowflake CEO Ramaswamy expressed his astonishment at DeepSeek’s ability to train a world-class AI model for just $6 million, challenging the belief that only a few companies could achieve such feats.

DeepSeek’s disruptive debut shakes Nvidia

Ramaswamy emphasized that DeepSeek’s breakthrough demonstrates the ongoing potential for innovation in AI and questioned whether the billions of dollars typically spent on developing cutting-edge models are truly necessary. Despite the enthusiasm surrounding DeepSeek’s claims, the industry is considerably skeptical. Questions linger about the full cost of developing DeepSeek beyond the training expenses. DeepSeek’s models, V3 and R1, have captivated the AI community with their performance and power efficiency, but some experts doubt the exact figures reported by the company. One contentious issue has been DeepSeek’s use of AI chips amid U.S. export controls restricting advanced AI chips to China. DeepSeek claims its advancements were achieved using less advanced Nvidia chips, but this has been met with skepticism from some industry leaders. As the debate continues, DeepSeek remains a company to watch in the evolving field of artificial intelligence. They are challenging the accepted paths toward achieving high-capability AI. Their achievements have had an immediate impact on the market, with AI-exposed semiconductor, hardware, and power stocks experiencing significant turbulence.