Hangzhou-based AI firm DeepSeek is roiling the U.S. tech sector and upending financial markets. The startup has managed to become competitive with Silicon Valley’s deep learning firms despite U.S. sanctions that prevent Chinese technology companies from buying premium chips. DeepSeek has made it into the global top 10 in terms of model performance, and as of this week had the top-ranked free AI assistant at the Apple App Store. DeepSeek’s new R1 model has drawn attention for using less computing power than competing systems, while performing comparably, despite having been developed using older Nvidia chips.

“DeepSeek-R1 is one of the most amazing and impressive breakthroughs I’ve ever seen,” Andreessen Horowitz partner Marc Andreessen wrote Friday on X social.

“Barrett Woodside, co-founder of the San Francisco AI hardware company Positron, said he and his colleagues have been abuzz about DeepSeek,” reports The Wall Street Journal, quoting Woodside calling it “very cool,” largely because “DeepSeek’s open-source models in which the software code behind the AI model is made available free,” WSJ explains.

The company’s foundation platform, DeepSeek-V3, was introduced the day after Christmas, according to The New York Times.

“DeepSeek’s development was led by a Chinese hedge-fund manager, Liang Wenfeng, who has become the face of the country’s AI push,” writes The Wall Street Journal, reporting that on January 20, the day R1 was launched, “Liang met China’s premier and discussed how homegrown companies could narrow the gap with the U.S.”

Already, the company sent ripples through the U.S. stock market, sparking a selloff of chip shares triggered by “concerns over whether the huge spending by U.S. tech giants on leading-edge semiconductors and other AI infrastructure was justified,” WSJ says, noting “the tech-focused Nasdaq fell 3 percent and Nvidia slid more than 10 percent.”

Engadget mentions that “Google parent company Alphabet and Microsoft were also down,” impacted by news of DeepSeek topping the App Store chart.

Accessible via Web, app and API, “DeepSeek is similar to AI assistants like ChatGPT with features like coding content creation and research,” Engadget explains. “DeepSeek-R1 release is available under an MIT license, so it can be used commercially and without restrictions.”

In an attention-getting twist, DeepSeek has been telling people that “training one of its latest models cost $5.6 million, compared with the $100 million to $1 billion range cited last year by Dario Amodei, chief executive of the AI developer Anthropic, as the cost of building a model,” per WSJ.

While that has been disputed, they’re clearly on to something, A DeepSeek research paper provides a peek beneath the hood.

Related:
What to Know About DeepSeek and How It Is Upending AI, The New York Times, 1/27/25
Nvidia’s $589 Billion DeepSeek Rout Is Largest in Market History, Bloomberg, 1/27/25
Nvidia Reels After China’s AI Breakthrough, The New York Times, 1/27/25
DeepSeek’s Popular AI App Is Explicitly Sending U.S. Data to China, Wired, 1/27/25
DeepSeek Won’t Sink U.S. AI Titans, The Wall Street Journal, 1/27/25