DeepSeek: How China Influences US Tech Stocks
The Chinese startup DeepSeek is putting significant pressure on AI providers from the U.S., including OpenAI and NVIDIA—and their stock prices are…
While Europe is still waiting for a serious response to ChatGPT, Chinese startup DeepSeek is putting immense pressure on U.S. AI players like OpenAI and NVIDIA—and dragging down their stock prices in the process.
Key Takeaways
- DeepSeek has unveiled a new AI model that reportedly delivers high performance at low cost.
- U.S. tech giants including Nvidia, AMD, Microsoft, and OpenAI lost market value following the announcement.
- Nvidia’s stock dropped 7.5%, pushing its market cap behind Apple to second place.
- Analysts doubt ChatGPT could truly be rebuilt for $5 million, calling the market panic overblown.
- DeepSeek’s MoE architecture reduces operating costs but isn’t a technical breakthrough.
Hardly had OpenAI, together with its technology partners, announced a $500 billion investment under the name Stargate to build AI data centers when the company behind ChatGPT—and other AI vendors and hardware makers—came under pressure. The reason? Chinese startup DeepSeek suddenly unveiled a seemingly superior new AI model.
DeepSeek Sends Shockwaves Through U.S. Tech
As IT Business reports, experts urge caution not to overestimate DeepSeek’s recent developments. However, market observers suggest the current debate could trigger a correction of the inflated valuations seen in many American tech companies. The pressure isn’t limited to OpenAI’s standing—it also affects Nvidia, Broadcom, AMD, and Microsoft, all of which have benefited from the AI boom.
The same applies to Japan’s SoftBank, one of the main investors in Stargate alongside OpenAI and Oracle. Its stock also stumbled over the last weekend of January. Nvidia’s shares lost 7.5% of their value, falling to $132. As a result, the chip and graphics card maker has slipped back to second place among the world’s most valuable companies, behind Apple. Microsoft, ranked third, now trails Nvidia by only a narrow margin, with a market capitalization of approximately $3.2 trillion.
According to IT Business, energy sector stocks also suffered from the news out of China, having previously enjoyed strong momentum due to existing and announced AI investments. Among the hardest hit is Siemens Energy.
“Doomsday Scenario” Overblown
At Meta, Facebook’s parent company—which has just announced AI investments of 60 billion dollars—the news about DeepSeek-R1, the name of the Chinese AI model, hasn’t been well received either, especially since the Chinese reportedly spent only 5 million dollars developing their language model, a figure some are skeptical about.
Bernstein Research, for example, says DeepSeek’s AI models are good and powerful, but OpenAI’s ChatGPT certainly wasn’t rebuilt for five million. “The doomsday scenario currently spreading across the Twitterverse seems exaggerated,” IT Business quotes analysts from the research firm led by Stacy Rasgon.
The efficiency of DeepSeek-V3 isn’t surprising given its model architecture, particularly since the Mixture-of-Experts (MoE) approach used here is specifically designed to reduce training and operational costs for AI models by activating only a portion of the model parameters at any given time. This approach, by the way, isn’t a world-first innovation. Moreover, such efficiency gains aren’t negative but instead consistently drive higher demand, according to the analysts.
Frequently Asked Questions
Which Chinese company is making waves with a new AI model?
The Chinese startup DeepSeek has attracted attention with its new AI model, DeepSeek-R1. It’s described as powerful and cost-efficient.
How much did the development of the DeepSeek model reportedly cost?
Reports suggest development costs were just five million dollars. However, this figure is being questioned by some experts.
What impact did DeepSeek’s announcement have on U.S. tech stocks?
Share prices of Nvidia, AMD, Microsoft, OpenAI, and Broadcom declined. Nvidia lost 7.5 percent in value, and SoftBank also dropped significantly.
Why is the MoE architecture significant for AI development?
The Mixture-of-Experts architecture reduces costs because only parts of the model are active at any time. It’s efficient, but not a completely new technical design.
How are analysts responding to the discussion around DeepSeek’s technological progress?
Bernstein Research considers the widespread “doomsday scenario” to be overblown. The models are good, but no reason for panic among U.S. providers.
Editor’s Reading Recommendations
- API-First: Why modern cloud architectures stand or fall on API design
- Pretext: Can a JavaScript library solve the 30-year browser problem—or is it just hype?
- AWS vs. Azure vs. Google Cloud 2026: The honest comparison for DACH companies
More from the MBF Media Network
SecurityToday | MyBusinessFuture | Digital Chiefs
Header image source: Unsplash / Anne Nygard
