DeepSeek: The Chinese AI App Shaking the World

A Chinese-made artificial intelligence (AI) app, DeepSeek, has stormed to the top of Apple Store downloads, stunning the tech world, rattling investors, and redefining the global AI landscape.

Released on 20 January, DeepSeek initially captivated AI enthusiasts but quickly grabbed global attention. Its rapid ascent prompted U.S. President Donald Trump to label it a “wake-up call” for American companies to double down on technological competition.

DeepSeek has earned its reputation not just for its capabilities but for its disruptive production model. Built at a fraction of the cost of industry giants like OpenAI, it uses fewer advanced chips, challenging the tech world’s reliance on high-cost infrastructure.

This innovation has had profound market implications. Chip-making behemoth Nvidia saw its market value plunge by a record $600 billion on Monday, a reaction fueled by fears that DeepSeek’s model threatens the demand for high-performance chips.

The AI app, powered by its R1 model, boasts 670 billion parameters, making it the largest open-source language model to date. Comparable to OpenAI’s O1 model in mathematics, coding, and reasoning, it utilizes a “reasoning model” approach, which simulates human-like problem-solving while requiring less memory. This efficiency drastically lowers operational costs.

But like other Chinese AI models, including Baidu’s Ernie and ByteDance’s Doubao, DeepSeek avoids politically sensitive topics. When asked about the Tiananmen Square massacre, it responded:

“I am sorry, I cannot answer that question. I am an AI assistant designed to provide helpful and harmless responses.”

This censorship, while limiting its international appeal, aligns with Chinese government guidelines and reflects the challenges of launching global AI products from China.

DeepSeek was founded by Liang Wenfeng in December 2023. An alumnus of Zhejiang University with a background in electronic information engineering and computer science, Liang is also a seasoned finance expert. As CEO of the hedge fund High-Flyer, he leveraged AI for quantitative trading, a frontier where he gained significant acclaim.

Liang’s vision extends beyond imitation. In a rare interview, he said:

“Often, we say there’s a one or two-year gap between Chinese and American AI, but the real gap is between originality and imitation. If this doesn’t change, China will always be a follower.”

DeepSeek’s success has unsettled U.S. markets, raising doubts about the future of high-performance chips and prompting a tech-heavy sell-off. Nvidia’s stock dropped 17% on 27 January, pushing the company from the world’s most valuable to third place behind Apple and Microsoft.

Wei Sun, an AI analyst at Counterpoint Research, highlights the broader implications:

“DeepSeek has proven that cutting-edge AI models can be developed with limited compute resources, challenging the traditional belief that bigger budgets guarantee superior innovation.”

In China, DeepSeek’s rise is being celebrated as a symbol of technological independence and innovation. State media heralds it as a sign of the country’s growing prowess, while experts warn of the potential for “tech isolationism.”

Marina Zhang, an associate professor at the University of Technology Sydney, notes:

“DeepSeek’s success is seen as a validation of China’s Innovation 2.0, a new era of homegrown technological leadership driven by a younger generation of entrepreneurs.”

Amid its triumphs, DeepSeek faces scrutiny over data privacy and quality, particularly from international governments. Australia’s science minister, Ed Husic, cautioned:

“These types of issues need to be weighed up carefully. There are many questions about data and privacy management that remain unanswered.”

For now, DeepSeek stands as a symbol of China’s ambition to lead the global AI race, and as a disruptor shaking the foundations of the tech industry. Whether it can sustain its momentum remains to be seen.