DeepSeek: The Chinese AI That Shocked the AI World

DeepSeek: The Chinese AI That Shocked the AI World

Ondřej Barták
Ondřej Barták
Entrepreneur and Programmer
14. 8. 2025
4 minutes reading · 17 views
DeepSeek: The Chinese AI That Shocked the AI World

DeepSeek: The Chinese AI That Shocked the AI World Upon Release

A new app from China shot to the top of the App Store rankings overnight and caused tech giant Nvidia's market value to fall by $600 billion. This happened in January 2025, when the Chinese company DeepSeek launched its DeepSeek-R1 model. This chatbot, which rivaled ChatGPT, not only attracted millions of users but also challenged prevailing assumptions about the development of artificial intelligence (AI).

Initial Shock and Market Impact

It all began shortly after President Donald Trump took office. DeepSeek-R1 quickly became the most downloaded free app in the US. The company claimed that its chatbot outperformed ChatGPT and had been developed at a fraction of the cost. This sudden success caused Nvidia's share price to fall by 17%, representing the largest single-day loss in US stock market history—a total of $600 billion (approximately 12.6 trillion Czech korunas). Other AI-related technology stocks were also affected.

Venture capitalist Marc Andreessen called it a "Sputnik moment" for AI, referring to the Soviet satellite that sparked the space race between the US and the USSR in the 1950s. Until then, China had been seen as lagging behind the US in AI. Suddenly, the situation appeared to have reversed. Shares of Japanese semiconductor-related companies also fell following the release of China's DeepSeek model.

How DeepSeek Changed the Approach to AI

Six months after its launch, DeepSeek-R1 is no longer such a major topic in San Francisco, but its influence persists. The model challenged the idea that bigger means better, as advocated by American leaders such as Sam Altman, CEO of OpenAI (the company behind ChatGPT). Sid Sheth, CEO of AI chip startup d-Matrix, said that DeepSeek showed it was possible to create a capable model without the most powerful technology by using smarter engineering.

Interest in DeepSeek exploded over a weekend in January, when company employees downloaded it before IT departments had time to ban it. Many companies subsequently blocked the app over concerns that user data could end up in China. Nevertheless, some Americans continue to use it. Some Silicon Valley startups choose it over more expensive American models to save money—savings that then go toward employee salaries.

Users on online forums share tips on how to run DeepSeek-R1 locally on their own devices to avoid servers in China. Christopher Caen, CEO of Mill Pond Research, described this as a way to use the model without worrying about data leaks.

US-China Rivalry in AI

DeepSeek marked a turning point in the rivalry between the US and China in AI. Analyst Wendy Chang of the Mercator Institute for China Studies said that until then, China had been seen as playing catch-up with the West in large language models (LLMs—systems trained to predict words in text). DeepSeek claimed to have developed a cutting-edge model for just $5.6 million (approximately 117 million Czech korunas), while OpenAI spent $5 billion (approximately 105 billion Czech korunas) in 2024. DeepSeek-R1 outperformed OpenAI's o1 model in several tests.

This success helped American companies justify investment in AI as essential to staying ahead of China. David Sacks, an AI adviser to the Trump administration, emphasized that AI has profound implications for the economy and national security. The Trump administration unveiled an AI Action Plan to ensure that the US remains the dominant power.

Security concerns persist. The US government is investigating DeepSeek's ties to Beijing. According to Reuters reports from June 2025, a senior US State Department official said that DeepSeek likely supports the Chinese military and intelligence services. The company did not respond to the BBC's request for comment, but its privacy policy states that its servers are located in the People's Republic of China and that data may be processed there.

The Future of DeepSeek and the AI Industry

OpenAI recently released new models, its first freely available and customizable models in five years. Sid Sheth sees a direct connection to DeepSeek, which showed that smaller, more efficient models can be powerful. This shifted the industry's thinking toward models that are faster and less expensive.

Nevertheless, major American players are returning to the original approach: more data centers, chips, and energy. OpenAI recently launched GPT-5 and expanded its infrastructure. Meta, led by Mark Zuckerberg, is investing billions and attracting talent with $100 million compensation packages. Nvidia's shares recovered and reached new highs, making it the most valuable company in history.

Christopher Caen said that the original narrative surrounding DeepSeek had proved to be a dead end. According to Marina Zhang of the University of Technology Sydney, DeepSeek now faces challenges, including delays to the DeepSeek-R2 model due to a shortage of cutting-edge chips and competition from both the US and China.

DeepSeek may have shaken the AI world, but the old order appears to be returning. Nevertheless, its story shows how quickly the dynamics of this rapidly evolving field can change.

Advertisement

Content created with help from UpTier.

SEO and GEO on autopilot. UpTier’s multi-agent systems write and optimize content for search engines and AI answers.

Discover UpTier ↗

Category:AI
Did you enjoy this article?
Discover more interesting posts on our blog
Back to blog

Related posts

Google makes Tasks free for Gemini 4 Hydrogen usersGoogle makes Tasks free for Gemini 4 Hydrogen users
Tasks is becoming available for free to hundreds of millions of Gemini 4 Hydrogen users. Google says it can call businesses, book appointments and complete purchases on their behalf.
1 min read
18. 11. 2026
OpenAI introduces GPT-6.2 Astra and Sol for professional and everyday workOpenAI introduces GPT-6.2 Astra and Sol for professional and everyday work
OpenAI has introduced two models: Astra for complex professional tasks and Sol for everyday work. Both are available in ChatGPT Work, Codex and through the API.
3 min read
17. 11. 2026
Tencent releases Hy4 Flash with image support and a context window of up to 1 million tokensTencent releases Hy4 Flash with image support and a context window of up to 1 million tokens
The lighter Hy4 Flash model targets chat, documents and routine code edits. It accepts text and images, supports tool use and, according to Tencent, matched Claude Haiku 5.5 in tests of everyday tasks.
1 min read
16. 11. 2026
Přihlaste se k odběru našeho newsletteru
Zůstaňte informováni o nejnovějších příspěvcích, exkluzivních nabídkách, a aktualizacích.
CodedTrip

Operated by CodedTrip LLC, USA.

YouTube
TikTok