China’s DeepSeek AI Causes Market to Tumble – but Analysts Are Skeptical

Chinese AI company DeepSeek has emerged as a potential challenger to U.S. tech giants, demonstrating breakthrough AI models that claim to offer performance comparable to leading offerings at a fraction of the cost. The entire U.S. market is taking a tumble on Monday morning as a result, but many analysts are skeptical of the company’s claims.

TechCrunch reports that DeepSeek, a Chinese artificial intelligence firm founded in 2023 by Liang Wenfeng, former chief of AI-driven quant hedge fund High-Flyer, has recently made waves in the tech industry with its breakthrough AI models. The company’s mobile app, released in early January, has topped the App Store charts across major markets, including the U.S., UK, and China. However, the company’s claims have not escaped doubts about their veracity.

What sets DeepSeek apart is its open-source models — which are free to use — incorporate a reasoning feature, articulating the AI’s thought process before providing responses. The company claims to deliver performance comparable to leading AI offerings, such as GPT-4o, while using a fraction of the computing power.

This development — along with the fact that DeepSeek is open source and completely free for use — has significant implications for the tech industry, potentially reshaping competition between established tech giants and startups by lowering barriers to entry.

CNBC reports that the stock market is down heavily due to DeepSeek’s implications for high-flying tech stocks like Nvidia.  The Nasdaq Composite fell by 3.4 percent at the opening bell. The Dow fell 180 points and the S&P 500 cratered by two percent.

Immediately, social media users began exploring the extent of Chinese censorship present in DeepSeek.  One user posted a video demonstrating that the AI would begin to answer a question about the famous “Tank Man” image from the Tiananmen Square Massacre, only to censor itself.

Wall Street analysts were quick to post notes skeptical about DeepSeek’s claims. Citi questions whether DeepSeek’s achievements were made without the use of advanced GPUs to fine-tune the models or build the underlying large language models (LLMs) through the distillation technique. While the dominance of U.S. companies in the most advanced AI models could potentially be challenged, Citi estimates that in an inevitably more restrictive environment, the U.S.’ access to more advanced chips is an advantage.

Bernstein believes that the resulting panic in the tech industry seems overblown, acknowledging that DeepSeek may have reduced costs of achieving equivalent model performance by about 10 times. However, they note that current model cost trajectories are increasing by about that much every year, which cannot continue indefinitely. Bernstein argues that innovations like those employed by DeepSeek are necessary for AI to continue progressing and that any new compute capacity unlocked is more likely to be absorbed due to increased usage and demand rather than impacting long-term spending outlook.

Goldman Sachs sees potential competition between capital-rich internet giants and startups, given the lowering barriers to entry. They also expect a shift from training to more inferencing, with increased emphasis on post-training capabilities that require significantly lower computational resources compared to pre-training. Furthermore, Goldman Sachs anticipates the potential for further global expansion for Chinese players, given their performance and cost/price competitiveness.

DeepSeek’s progress suggests Chinese AI engineers have worked their way around U.S. restrictions on high-end technologies like GPU semiconductors, focusing on greater efficiency with limited resources. While it remains unclear how much advanced AI-training hardware DeepSeek has had access to, the company has demonstrated enough to suggest the trade restrictions have not been entirely effective in stymieing China’s progress.

Breitbart News will continue to report on this developing story.

Lucas Nolan is a reporter for Breitbart News covering issues of free speech and online censorship.