A Chinese artificial intelligence company called DeepSeek is grabbing America’s attention — and sending a shock wave through Wall Street — due to its new tech, which some experts say rivals that of OpenAI’s ChatGPT.
DeepSeek is also catching investors off guard because of the low development costs for its AI app, which Wedbush Securities analyst Dan Ives pegged at only $6 million. By comparison, OpenAI, Google and other major U.S. companies are on track to invest a total of roughly $1 trillion in AI over the coming years, according to Goldman Sachs.
On Monday, DeepSeek’s rollout roiled shares of AI stalwarts such as Nvidia, the high-flying manufacturer of advanced chips engineered for AI development, and Dutch company ASML, another chipmaker. The Chinese company’s tech is raising questions about whether demand for Nvidia’s chips could take a hit, as well as whether investors are overvaluing tech stocks that have been buoyed by the promise of AI, from Meta to Microsoft, experts said.
“DeepSeek has taken the market by storm by doing more with less,” said Giuseppe Sette, president at AI market research firm Reflexivity, in an email. “This shows that with AI the surprises will keep on coming in the next few years.”
DeepSeek’s latest app comes just days after President Trump announced a new $500 billion venture with ChatGPT maker OpenAI, Softbank and Oracle, dubbed Stargate, which he touted as ensuring “the future of technology” in the U.S.
AI-related stocks took a hit on Monday, with Nvidia shares tumbling 17%, shedding $600 billion in value and marking the single-biggest one-day loss for a company in stock market history, according to CNBC. ASML sank 6%, while Broadcom, another semiconductor stock, also slumped 17%.
Some energy-related stocks also plunged on Monday on investor worries that the new tech could require less energy to run, translating into lower demand from the tech sector. GE Vernova, which makes wind and gas turbines, plunged 21%, while electricity generator Vistra slumped 28%.
The tech-heavy Nasdaq index slumped 3%, or 612 points, while the S&P 500 declined 1.5%. The blue-chip Dow Jones Industrial Average added 0.7%.
Despite the sharp drop on the Nasdaq, it’s far from the worst day for the index during the past five years. The worst one-day decline since Jan. 27, 2020, came on March 16, 2020, when the index plunged more than 12% as COVID-19 pandemic hit the economy.
What is DeepSeek?
DeepSeek is a private Chinese company founded in July 2023 by Liang Wenfeng, a graduate of Zhejiang University, one of China’s top universities, who funded the startup via his hedge fund, according to the MIT Technology Review. Liang has about $8 billion in assets, Ives wrote in a Jan. 27 research note.
Liang, who had previously focused on applying AI to investing, had bought a “stockpile of Nvidia A100 chips,” a type of tech that is now banned from export to China. Those chips became the basis of DeepSeek, the MIT publication reported.
Ben Reitzes, head of technology research at Melius, told investors in a note that DeepSeek makes legitimate breakthroughs as an AI tool, including better learning and more efficient use of memory, although he expressed skepticism about the “amount of chips used.”
Is DeepSeek available in the U.S.?
The company’s AI app is available in Apple’s App store, as well as online at its website. The service is free and as of Monday morning was the top download on Apple’s store, although some people were having trouble signing up for the app.
On its Chinese site, DeepSeek blamed “large-scale malicious attacks” on its service, requiring it to temporarily limit new registrations. “Existing users can log in as usual,” the company said in the post, which was dated shortly after midnight Jan. 28 in China’s local time.
The company released its latest AI model on Jan. 20, which is causing Wall Street to reappraise the AI sector.
“Last week DeepSeek launched a model that rivals OpenAI’s ChatGPT and Meta’s Llama 3.1 and was #1 on Apple’s App Store over the weekend,” Wedbush’s Ives wrote. “DeepSeek built the model using reduced capability chips from Nvidia. which is impressive and thus has caused major agita for U.S. tech stocks with massive pressure on Nasdaq this morning.”
How is DeepSeek different than other AI apps?
DeepSeek is an open-source large language model that relies on what is known as “inference-time computing,” which Sette said in layman’s terms means “they activate only the most relevant portions of their model for each query, and that saves money and computation power.”
Some experts praised DeepSeek’s performance, with noted tech investor Marc Andreessen writing on X on Jan. 24, “DeepSeek R1 is one of the most amazing and impressive breakthroughs I’ve ever seen — and as open source, a profound gift to the world.”
David Sacks, a venture capitalist named by Mr. Trump to help oversee AI and cryptocurrency policy, said on social media Monday that DeepSee’s app “shows that the AI race will be very competitive.”
However, Ives said he’s skeptical the service will gain ground with major U.S. businesses.
“No U.S. Global 2000 is going to use a Chinese startup DeepSeek to launch their AI infrastructure and use cases,” Ives wrote. “At the end of the day there is only one chip company in the world launching autonomous, robotics, and broader AI use cases and that is Nvidia.”
Are there privacy issues with DeepSeek?
Some experts are raising concerns about the personal data that DeepSeek is collecting, given that the company stores data from users — including their date of birth, keystrokes, text or audio inputs, uploaded files, chat history and other data — on servers located in China, according to its privacy policy.
That echoes some of the issues raised with TikTok, another company with Chinese ownership that sparked worries about the supposed risks its ties to China posed to national security. Last year, Congress passed a law banning TikTok in the U.S. as long as it is under Chinese ownership, although that is now in flux after President Trump signed an executive order directing the Justice Department to not enforce the ban for 75 days.
What does DeepSeek mean for Nvidia and other tech companies?
Wall Street is trying to assess the long-term impact of a low-cost AI tool from China that rivals ChatGPT and other so-called generative AI apps. It also raises questions about whether Silicon Valley is overspending on tech advancements in the AI sector, noted Angelo Zino, senior equity analyst at CFRA Research, in an email.
“The fact that this technology is supposed to take less energy and is more cost-effective than U.S.-based models has U.S. technology investors very concerned,” Jay Woods, chief global strategist at Freedom Capital Markets, said.
It’s also unclear what type of pushback or reaction could come from the White House, given that Mr. Trump has raised the possibility of placing new tariffs on Chinese imports, although he also gave the Chinese-owned TikTok a reprieve by ordering the Justice Department not to enforce a looming ban.
Some Wall Street analysts think Monday’s stock selloff is an overreaction, noting that the enormous demand for AI will continue lifting key players in the sector.
“It’s one thing to train a [large language] model for less money, but accommodating the huge demand for the consumption of all this AI technology is still going to require massive amounts of infrastructure,” Adam Crisafulli of VitalKnowledge said in a report.
Analysts at Bernstein Research also noted that while DeepSeek’s technology looks promising, it likely isn’t revolutionary, suggesting that Monday’s investment rout is overblown.
What is Nvidia saying about DeepSeek?
In a statement to CBS News, Nvidia offered praise for DeepSeek.
“DeepSeek is an excellent AI advancement and a perfect example of test-time scaling,” the company said in an email. “DeepSeek’s work illustrates how new models can be created using that technique, leveraging widely available models and compute that is fully export-control compliant.”
But, Nvidia added, AI inference, or using AI models to make decisions or predictions, “requires significant numbers of NVIDIA GPUs and high-performance networking. We now have three scaling laws: pre-training and post-training, which continue, and new test-time scaling.”
AI: Artificial Intelligence
More
