Chinese AI start-up DeepSeek challenges US tech giants

DeepSeek's breakthrough AI model raises concerns about US competitiveness.

A welcome message displayed on the DeepSeek artificial intelligence app, captured in Riga, Latvia, on Monday, January 27, 2025. Photo by Andrey Rudakov/Getty Images

By Alana Salsabila and Laila Azzahra

The world of technology faced a seismic shift this week as tech stocks tumbled, with giants like Meta and Nvidia confronting difficult questions about their future. At the heart of this disruption lies a little-known Chinese artificial intelligence (AI) start-up, DeepSeek, which has unveiled groundbreaking advancements that have sent shockwaves across the global tech industry.

DeepSeek's announcement of a highly efficient and cost-effective AI model has sparked concerns about whether the United States can maintain its competitive edge in AI development. This development has led to DeepSeek being labeled "AI's Sputnik moment" by Marc Andreessen, a prominent venture capitalist.

Founded by the Chinese stock trading firm High-Flyer and led by entrepreneur Liang Wenfeng, DeepSeek specializes in creating advanced AI technologies similar to OpenAI's ChatGPT and Google's Gemini. Over the years, the start-up has attracted top talent from Chinese universities by offering competitive salaries and cutting-edge research opportunities.

Since its inception, DeepSeek has released several large language models, including the newly launched DeepSeek-V3, which matches the performance of leading chatbots developed by US companies. What sets DeepSeek apart, however, is its ability to achieve such feats with significantly fewer resources.

DeepSeek's engineers revealed that their new model was trained using only 2,000 Nvidia chips, a fraction of the resources typically required by industry leaders. For context, major tech companies like Meta and OpenAI rely on supercomputers equipped with up to 16,000 chips to train their AI systems. Furthermore, DeepSeek's team managed this feat with an investment of just $6 million, compared to the billions spent by US tech giants.

This achievement stems from innovative methods in data analysis and AI training. By leveraging a "mixture of experts" technique, DeepSeek efficiently distributed data analysis across multiple AI models, minimizing resource usage while maintaining high performance.

DeepSeek's advancements have raised critical questions about the US's ability to retain its leadership in AI development. The Biden administration's restrictions on exporting advanced chips to China were intended to hinder the country's AI progress. However, these restrictions have pushed Chinese researchers to innovate with available resources, leading to breakthroughs like those demonstrated by DeepSeek.

While DeepSeek's technology is impressive, it highlights a growing concern: the commoditization of AI. Experts believe that as AI technologies become more accessible, many companies may develop similar products, leveling the playing field.

Competition heats up

DeepSeek's achievements have not gone unnoticed. Shortly before DeepSeek released its DeepSeek-V3 model, OpenAI announced its new system, OpenAI o3, which reportedly surpasses DeepSeek's capabilities in reasoning tasks involving math, science, and programming. However, OpenAI o3 has not yet been made widely available.

In response, DeepSeek launched its own reasoning model, DeepSeek R1, on January 20, further intensifying competition. This development has left US investors and tech leaders scrambling to reassess their strategies.

One of DeepSeek's most significant moves has been its decision to open-source its AI system. By sharing its underlying code, DeepSeek has enabled other researchers and companies to build upon its technology. This approach mirrors a trend initiated by Meta in 2023 with its open-source AI system, Llama.

While open-source AI accelerates innovation, it also raises concerns about misuse, including the spread of disinformation. Some US lawmakers have proposed measures to limit the practice, fearing its potential consequences. However, experts argue that stifling open-source AI in the US could give China a strategic advantage, positioning it as a global leader in AI research and development.

DeepSeek's rapid rise underscores the evolving dynamics of the global AI landscape. As Chinese start-ups demonstrate their ability to compete with US tech giants, questions about the future of AI innovation and geopolitical implications have come to the forefront.

The coming years will likely determine whether the US can adapt to these challenges or whether China's ingenuity will redefine the balance of power in the AI race. For now, DeepSeek's achievements serve as a wake-up call for the tech industry, highlighting the need for innovation, collaboration, and strategic foresight.

Share