There’s a new player in AI on the world stage: DeepSeek, a Chinese startup that’s throwing tech valuations into chaos and challenging U.S. dominance in the field with an open-source model that they say they developed for a fraction of the cost of competitors.
DeepSeek’s free AI assistant — which by Monday had overtaken rival ChatGPT to become the top-rated free application on Apple’s App Store in the United States — offers the prospect of a viable, cheaper AI alternative, raising questions on the heavy spending by U.S. companies such as Apple and Microsoft, amid a growing investor push for returns.
U.S. stocks dropped sharply on Monday, as the surging popularity of DeepSeek sparked a selloff in chipmaker Nvidia and other companies that stand to benefit from investments in the technology.
Nvidia, whose chips are the top choice for powering AI applications, saw shares fall by at least 17 per cent on Monday.
“[DeepSeek] performs as well as the leading models in Silicon Valley and in some cases, according to their claims, even better,” Sheldon Fernandez, co-founder of DarwinAI, told CBC News. “But they did it with a fractional amount of the resources is really what is turning heads in our industry.”
What is DeepSeek?
Little is known about the small Hangzhou startup behind DeepSeek, which was founded out of a hedge fund in 2023, but largely develops open-source AI models.
Its researchers wrote in a paper last month that the DeepSeek-V3 model, launched on Jan. 10, cost less than $6 million US to develop and uses less data than competitors, running counter to the assumption that AI development will eat up increasing amounts of money and energy.
Some analysts are skeptical about DeepSeek’s $6 million claim, pointing out that this figure only covers computing power. But Fernandez said that even if you triple DeepSeek’s cost estimates, it would still cost significantly less than its competitors.
The open source release of DeepSeek-R1, which came out on Jan. 20 and uses DeepSeek-V3 as its base, also means that developers and researchers can look at its inner workings, run it on their own infrastructure and build on it, although its training data has not been made available.
It “upends” the business model of competitors who charge monthly fees to companies to access their AI models’ more advanced features, Fernandez said.
DeepSeek has priced its models at up to 40 times lower than Open AI’s comparable models, he added.
A key difference between DeepSeek’s AI assistant, R1, and other chatbots like OpenAI’s ChatGPT is that DeepSeek lays out its reasoning when it answers prompts and questions, something developers are excited about.
“The dealbreaker is the access to the raw thinking steps,” Elvis Saravia, an AI researcher and co-founder of the U.K.-based AI consulting firm DAIR.AI, wrote on X, adding that the response quality was “comparable” to OpenAI’s latest reasoning model, o1.
U.S. dominance in AI challenged
One of the reasons DeepSeek is making headlines is because its development occurred despite U.S. actions to keep Americans at the top of AI development. In 2022, the U.S. curbed exports of computer chips to China, hampering their advanced supercomputing development.
According to the paper on DeepSeek-V3’s development, researchers used Nvidia’s H800 chips for training, which are not top of the line. Initially developed as a reduced-capability product to get around curbs on sales to China, they were subsequently banned by U.S. sanctions.
The latest AI models from DeepSeek are widely seen to be competitive with those of OpenAI and Meta, which rely on high-end computer chips and extensive computing power.
Marc Andreessen, the Silicon Valley venture capitalist, said in a post on X on Sunday that DeepSeek’s R1 model was AI’s “Sputnik moment,” referencing the former Soviet Union’s launch of a satellite that marked the start of the space race with the U.S. in the late 1950s.
“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,” he said in a separate post.
Fernandez called it a “wake up call to the AI sector.”
“Just up until last week, the belief was to innovate in the sector you needed very, very large data centers,” he said, pointing out that large companies like Microsoft have sunk billions into AI development. “And what they have demonstrated is that if you pair ingenuity with some cleverness regarding these models, that doesn’t have to be the path.”
Nvidia drops 17%
The impact was being felt in tech markets on Monday.
Nvidia shares plummetted, putting it on track to lose roughly $600 billion US in stock market value, the deepest ever one-day loss for a company on Wall Street, according to LSEG data. Industry peers Broadcom and Marvell Technology fell about 11 per cent each.
Microsoft, Meta Platforms and Google parent Alphabet fell between 2.5 per cent and four per cent, while AI server maker Dell Technologies was down by 11.7 per cent.
Geoffrey Hinton fears an AI takeover — but says there’s a way to stop it. The British Canadian computer scientist, known as the “godfather of AI,” says governments should regulate how tech companies are developing artificial intelligence to make sure it’s done safely.
In Japan, startup investor SoftBank Group 9984.T slid more than eight per cent. Last week, it announced a $19-billion US commitment to fund Stargate, a data-centre joint venture with OpenAI.
In his announcement of the Stargate deal, U.S. President Donald Trump said the U.S. was investing up to $500 billion in the private sector to fund infrastructure for artificial intelligence.
European tech stocks also dropped on Monday in the technology sector.
“The catalyst of a foreign competitor to U.S.-led AI dominance begs other questions about trade and semiconductor chips and energy needs,” Robert Savage, head of markets strategy and insights at BNY, wrote in a note on Monday, calling the markets “unsettled.”
Discussion about this post