DeepSeek Shakes the Market: Chinese AI Startup Challenges NVidia and Tech Giants
The unexpected advancement of a Chinese artificial intelligence company called DeepSeek has challenged the perceived invincibility of the U.S. tech industry.
In just one year of operation, DeepSeek unveiled a remarkable capability: A ChatGPT-like AI model called R1, which delivers comparable performance at a fraction of the cost of popular models from OpenAI, Google and Meta. The company stated that training its latest AI model cost only $5.6 million, compared to the hundreds of millions or billions of dollars spent by U.S. companies on similar technologies. The 'Wall Street Journal' was the first journal to report these ultra-low costs.
This sent shockwaves through the tech sector on Monday. Last week, Meta announced plans to spend over $65 billion this year on AI development. Meanwhile, Sam Altman, CEO of OpenAI, said last year that the AI industry would require trillions of dollars in investment to support the production of the high-demand chips needed to power energy-intensive data centers that run complex models.
How did DeepSeek achieve this breakthrough?
The extraordinary success of a relatively unknown AI startup becomes even more astonishing when considering that the U.S. has spent years restricting the supply of high-performance AI chips to China due to national security concerns. This means that DeepSeek managed to develop its cost-effective model using lower-powered AI chips.
DeepSeek's interface, similar to that of its competitors. Photo: Xataka.
U.S. tech stocks took a significant hit on Monday morning. NVidia (NVDA), the leading supplier of AI chips, whose stock had more than doubled in each of the last two years, dropped 12% in premarket trading. Meta (META) and Alphabet (GOOGL), Google’s parent company, also experienced sharp declines, as did Marvell, Broadcom, Palantir, Oracle, and other tech giants.
This dragged down the broader market as tech stocks account for a substantial portion, approximately 45%, of the S&P 500, according to Keith Lerner, an analyst at Truist. S&P 500 futures pointed to a 2.4% drop at market open, while the tech-heavy Nasdaq was set to open 4.2% lower. The Nasdaq hasn’t closed down by 4% or more since September 2022. The Dow Jones was projected to open roughly 400 points, or 0.9%, lower.
“The bottom line is that U.S. outperformance has been driven by tech and the leadership U.S. companies have in AI,” Lerner said. “The rollout of DeepSeek’s model is leading investors to question the extent of U.S. companies’ lead, how much they’re spending, and whether that spending will result in profits or overspending.”
What’s next?
This week marks the start of a wave of earnings reports from tech companies, and their reactions to DeepSeek’s breakthrough could lead to volatile market movements in the coming days and weeks. However, one achievement, albeit remarkable, may not be enough to counteract years of progress in American AI leadership. Additionally, a massive customer shift toward a Chinese startup is unlikely, suggesting that the market sell-off may be somewhat overblown. Alternatively, investors may have been looking for a reason to sell.
“Time will tell if the DeepSeek threat is real. The race is on to determine which technologies work and how major Western players will respond and evolve,” said Michael Block, market strategist at Third Seven Capital. “Markets had grown too complacent at the start of this Trump second era and might have been seeking an excuse to pull back, and they found a compelling one here.”