A Chinese-made artificial intelligence (AI) model called DeepSeek has quickly risen to the top of the Apple Store’s download list, stunning investors and dragging down some tech stocks.
Its latest version was released on January 20, quickly impressing AI experts before capturing the attention of the entire tech industry — and the world.
US President Donald Trump said it was a “wake-up call” for US companies to focus on “competing to win”.
What makes DeepSeek special is the company’s claim that it was built at a fraction of the cost of industry-leading models like OpenAI — because it uses fewer cutting-edge chips.
That possibility cost chip giant Nvidia nearly $600bn (£482bn) in market value on Monday – the biggest one-day loss in US history.
DeepSeek also raised questions about Washington’s efforts to curb Beijing’s ambitions for technological dominance, as one of its main restrictions is a ban on the export of advanced chips to China.
But Beijing has doubled down, with President Xi Jinping declaring AI a top priority. And startups like DeepSeek are vital as China shifts away from traditional manufacturing like clothing and furniture and toward cutting-edge technology — chips, electric vehicles and AI.
So what do we know about DeepSeek?
Australia warns against DeepSeek – is it safe to use?
DeepSeek vs ChatGPT – how are they different?
China’s DeepSeek AI Shakes Industry and Undermines America’s Confidence
What is artificial intelligence?
Sometimes, AI can make a computer look like a human.
Machines use technology to learn and solve problems, often by being trained with huge amounts of information and recognizing patterns.
The end result is software that can converse like humans or predict people’s shopping habits.
In recent years, the technology has become best known as the technology behind chatbots like ChatGPT — and DeepSeek — aka generative AI.
These programs again learn from huge amounts of data, including online text and images, to be able to generate new content.
But these tools can generate false positives and often reproduce biases present in their training data.
Millions of people use tools like ChatGPT to help them with everyday tasks like writing emails, summarizing text, and answering questions — and others even use them to aid in basic programming and learning.
What is DeepSeek?
DeepSeek is the name of a free AI-powered chatbot that looks, feels, and works very similar to ChatGPT.
That means it’s used for many of the same tasks, although how well it actually performs compared to its competitors remains a matter of debate.
It’s reportedly as strong as OpenAI’s o1 model — released late last year — on tasks including math and coding.
Like o1, R1 is a “reasoning” model. These models generate responses step by step, simulating a process similar to how humans reason through problems or ideas. It uses less memory than its competitors, ultimately reducing the cost of performing the task.
Like many other Chinese AI models — Baidu’s Ernie or ByteDance’s Doubao — DeepSeek is trained to avoid politically sensitive questions.
When the BBC asked the app about what happened in Tiananmen Square on June 4, 1989, DeepSeek did not provide any details about the massacre, a taboo subject in China.
The assistant replied, “I’m sorry, but I can’t answer that question. I’m an AI assistant designed to provide helpful and harmless answers.”
Chinese government censorship is a major challenge to the country’s AI ambitions on the international stage. But DeepSeek’s base model appears to have been trained on authentic sources while introducing a layer of censorship or withholding some information through an additional layer of protection.
Deepseek says it can do this cheaply – the researchers behind the tool claim it cost $6m (£4.8m) to train, a fraction of the “over $100m” that OpenAI boss Sam Altman alluded to when discussing GPT-4.
DeepSeek’s founder is said to have built up a stockpile of Nvidia A100 chips, which have been banned from export to China since September 2022.
Some experts believe that this collection — which some estimate to be as many as 50,000 — helped him build such a powerful AI model, by pairing these chips with cheaper, less sophisticated ones.
On the very day that DeepSeek’s AI assistant became the most downloaded free app on Apple’s App Store in the United States, the app was subjected to a “widespread malicious attack,” the company said, forcing it to temporarily limit sign-ups.
The company’s website also crashed on Monday.
Who is behind DeepSeek?
DeepSeek was founded by Liang Wenfeng in December 2023 and released its first large AI language model the following year.
Little is known about Liang, who graduated from Zhejiang University with a degree in electronic information engineering and computer science. But he now finds himself in the international spotlight.
He recently appeared at a meeting chaired by Chinese Premier Li Qiang, reflecting DeepSeek’s growing prominence in the AI industry.
Unlike many American AI entrepreneurs who hail from Silicon Valley, Mr. Liang also has a background in finance.
He is the CEO of High-Flyer, a hedge fund that uses artificial intelligence to analyze financial data to make investment decisions—known as quantitative trading. In 2019, High-Flyer became the first quantitative hedge fund in China to raise more than 100 billion yuan ($13 million).
In a speech last year, Liang said: “If the United States can develop its quantitative trading sector, why can’t China?”
In a rare interview last year, he said China’s AI industry “cannot lag behind forever.”
“We often say that there is a gap of one or two years between Chinese and American AI, but the real gap is between originality and imitation. If this does not change, China will always be an imitator.”
Asked why DeepSeek’s model surprised so many in Silicon Valley, he said: “Their surprise comes from the fact that they saw a Chinese company coming into the game as, ‘Be an innovator, not just a follower – which is what most Chinese companies are used to.'”
Australia’s Science Minister has expressed doubts about the app’s security.
“There are a lot of questions that need to be answered in due course, around quality, consumer preferences, data management and privacy,” Ed Husic told ABC.
“I would be very careful about that. Such matters must be carefully considered.”
How does this affect US companies like Nvidia?
DeepSeek’s achievements have challenged the belief that bigger budgets and high-end chips are the only way to advance AI, raising uncertainty about the future of high-performance chips.
“DeepSeek has shown that advanced AI models can be built with limited computing resources,” said Wei Sun, principal AI analyst at Counterpoint Research.
“In contrast, OpenAI, valued at $157 billion, is under intense scrutiny for its ability to maintain a commanding lead in innovation or justify its massive valuation and spending without generating significant profits.”
The company’s lower costs may have caused turmoil in financial markets on Jan. 27, sending the tech-heavy Nasdaq index down more than 3% in a broad sell-off that also included chipmakers and hubs. global data
The hardest hit was Nvidia, whose shares fell 17% on Monday but began to recover gradually on Tuesday, rising about 4% by midday.
The chipmaker was once the world’s most valuable company by market capitalization, but on Monday it fell to third place behind Apple and Microsoft, with its market value falling from $3.5 trillion to $2.9 trillion, Forbes reported.
DeepSeek is a private company, which means investors cannot buy its shares on any major exchanges.
Watch: AI bot DeepSeek responds to BBC question about Tiananmen Square
China celebrates DeepSeek’s influence
The development of DeepSeek technology is a huge support for the Chinese government, which wants to become independent from the West in terms of technology.
While the Communist Party has yet to comment on the matter, Chinese state media is keen to report that Silicon Valley and Wall Street giants are “losing sleep” over DeepSeek, a company that is “turning everything upside down.” US Stock Exchange
“In China, DeepSeek’s progress is seen as evidence of the country’s growing technological power and self-reliance,” said Marina Zhang, an assistant professor at the University of Technology Sydney.
“The company’s success is seen as a validation of China Innovation 2.0, a new era of national technological leadership driven by a younger generation of entrepreneurs.”
But she warned that such a mindset could also lead to “technological isolationism.”
