What is DeepSeek, the Chinese AI Startup Shaking Up the Stock Market and US Dominance?

Associated Press
January 28, 2025 | 3:34 pm
SHARE
People reflected in a window with a slogan about AI at a representation of a company ahead of the World Economy Forum in Davos, Switzerland, Sunday, Jan. 19, 2025. (AP Photo/Markus Schreiber)
People reflected in a window with a slogan about AI at a representation of a company ahead of the World Economy Forum in Davos, Switzerland, Sunday, Jan. 19, 2025. (AP Photo/Markus Schreiber)

A frenzy over DeepSeek, a Chinese artificial intelligence startup, disrupted stock markets Monday, sparking debates about economic and geopolitical competition between the US and China in AI development.

The company's AI chatbot became the most downloaded free app on Apple’s iPhone store, intensifying curiosity about the ChatGPT rival. Some US tech industry observers worry that DeepSeek’s rapid advancements suggest it is catching up with American AI giants at a fraction of the cost.

If true, this raises questions about the enormous investments US companies are making in data centers and chips to power further AI breakthroughs. However, the hype around DeepSeek has also created confusion about its actual technological progress.

“The models they built are fantastic, but they aren’t miracles,” said Stacy Rasgon, a semiconductor industry analyst at Bernstein, describing Wall Street’s reaction as exaggerated. “They’re not using unknown or secret innovations. These are things everybody is experimenting with.”

Advertisement

What is DeepSeek?

Founded in 2023 in Hangzhou, China, DeepSeek released its first AI large language model that year. CEO Liang Wenfeng, a co-founder of High-Flyer, one of China’s leading AI-driven hedge funds, reportedly used 10,000 Nvidia A100 chips before US restrictions limited their export to China. DeepSeek now relies on Nvidia’s H800 chips, which are compliant with US export controls.

DeepSeek gained wider attention last month after unveiling an AI model said to rival those of US companies like OpenAI, offering cost-effective solutions for training systems on vast datasets. The chatbot became available on Apple and Google app stores earlier this year, further boosting its popularity.

Last week, the release of a research paper on DeepSeek’s R1 model set off widespread panic. The paper showcased R1’s advanced reasoning skills, such as rethinking math problem approaches, at significantly lower costs compared to OpenAI’s o1 model.

“The price points freaked people out,” Rasgon noted.

AI’s ‘Sputnik Moment’?

The excitement around DeepSeek highlights the broader debate on US-China competition in AI. Venture capitalist Marc Andreessen described R1 as “AI’s Sputnik moment,” comparing it to the Soviet Union’s 1957 satellite launch that spurred the US space race.

Andreessen has argued that overregulation of US AI could stifle innovation and allow China to surpass American companies. Meanwhile, some analysts suggest DeepSeek’s timing is political, with its breakthroughs aiming to undermine US export controls.

“The technological innovation is real, but the timing of the release is political,” said Gregory Allen, director of the Wadhwani AI Center at the Center for Strategic and International Studies.

Trump’s Response

Former President Donald Trump, speaking Monday to House Republicans in Miami, called DeepSeek’s advancements “a wakeup call” and emphasized the need to stay competitive. On his first day in office last week, Trump signed an order to tighten loopholes in export controls, signaling a continuation of Biden-era policies on limiting US semiconductor exports to China.

DeepSeek’s progress could challenge the planned $500 billion AI investment by OpenAI, Oracle, and SoftBank that Trump previously lauded.

What Sets DeepSeek Apart?

DeepSeek’s open-source approach distinguishes it from competitors like OpenAI, though the company has not disclosed the data used for training its models. The R1 model, in particular, has drawn praise for its use of "Test Time Scaling," enabling the AI to improve its reasoning by "thinking out loud" without requiring additional data.

“This is the first time we’ve seen a Chinese company come this close within such a short time,” said Lennart Heim, a researcher at Rand Corp. “I used to think OpenAI was untouchable, but it turns out that’s not entirely true.”

On Monday, Nvidia’s stock fell 17 percent, but the company praised DeepSeek’s work as “an excellent AI advancement” achieved with export-compliant technology.

While questions remain about DeepSeek’s economic model, its rapid rise signals China’s growing competitiveness in AI development and its challenge to US dominance in the field.

Tags: Keywords:
SHARE

The Latest


Lifestyle 2 hours ago

Guardians of the Forest: How the Dayak Wehea Protect Their Ancestral Land

The Dayak Wehea serve as the last guardians of the Wehea Protected Forest, preserving their ancestral land amid modern challenges.
News 4 hours ago

Prabowo Eyes $44 Billion Budget Cuts by 2025 to Fund Danantara, Free Meals Program

Prabowo announces Rp 750T budget savings plan to fund Danantara and free meal (MBG) program by 2025.
Lifestyle 5 hours ago

Green Day vs. Linkin Park: Back-to-Back Jakarta Shows, Which Side Are You On?

Green Day and Linkin Park take over Jakarta with back-to-back shows! Punk vs. nu-metal—who will you choose?
News 6 hours ago

Prabowo Responds to Criticism Over 'Bloated' Cabinet at Gerindra's 17th Anniversary

President Prabowo Subianto addressed criticism of his 109-member "Red and White Cabinet," which has been labeled as "bloated".
Special Updates 6 hours ago

Energia Prima Nusantara Launches New Strategy to Spur Growth

Energia Prima Nusantara recently launched its strategy document that laid out its plans to spur business growth.
COPYRIGHT © 2025 JAKARTA GLOBE. ALL RIGHTS RESERVED