“We confirmed DeepSeek communicating with ByteDance,” the South Korean data protection regulator told Yonhap News Agency.

The country had already removed DeepSeek from app stores over the weekend over data protection concerns.

The Chinese app caused shockwaves in the AI world in January, wiping billions off global stock markets over claims its new model was trained at a much lower cost than US rivals such as ChatGPT.

Since then, multiple countries have warned that user data may not be properly protected, and in February a US cybersecurity company alleged potential data sharing between DeepSeek and ByteDance.

DeepSeek’s apparent overnight impact saw it shoot to the top of App Store charts in the UK, US and many other countries around the world – although it now sits far below ChatGPT in UK rankings.

In South Korea, it had been downloaded over a million times before being pulled from Apple and Google’s App Stores on Saturday evening.

Existing users can still access the app and use it on a web browser.

The data regulator, the Personal Information Protection Commission (PIPC), told South Korea’s Yonhap News Agency that despite finding a link between DeepSeek and ByteDance, it was “yet to confirm what data was transferred and to what extent”.

Critics of the Chinese state have long argued its National Intelligence Law allows the government to access any data it wants from Chinese companies.

However, ByteDance, headquartered in Beijing, is owned by a number of global investors – and others say the same law allows for the protection of private companies and personal data.

Fears over user data being sent to China was one of the reasons the US Supreme Court upheld a ban on TikTok, which is owned by ByteDance.

The US ban is on hold until 5 April as President Donald Trump attempts to broker a resolution.

Cybersecurity company Security Scorecard published a blog on DeepSeek on 10 February which suggested “multiple direct references to ByteDance-owned” services.

“These references suggest deep integration with ByteDance’s analytics and performance monitoring infrastructure,” it said in its review of DeepSeek’s Android app.

Security Scorecard expressed concern that along with privacy risks, DeepSeek “user behaviour and device metadata [are] likely sent to ByteDance servers”.

It also found data “being transmitted to domains linked to Chinese state-owned entities”.

On Monday, South Korea’s PIPC said it “found out traffic generated by third-party data transfers and insufficient transparency in DeepSeek’s privacy policy”.

It said DeepSeek was cooperating with the regulator, and acknowledged it had failed to to take into account South Korean privacy laws.

But the regulator advised users “exercise caution and avoid entering personal information into the chatbot”.

South Korea has already followed a number of countries such as Australia and Taiwan in banning DeepSeek from government devices.

The government agency said the AI model will become available again to South Korean users when “improvements and remedies” are made to ensure it complies with the country’s personal data protection laws.

In the week after it made global headlines, DeepSeek became hugely popular in South Korea leaping to the top of app stores with over a million weekly users.

But its rise in popularity also attracted scrutiny from countries around the world which have imposed restrictions on the app over privacy and national security concerns.

ADVERTISEMENT

South Korea’s Personal Information Protection Commission said the DeepSeek app became unavailable on Apple’s App Store and Google Play on Saturday evening.

It came after several South Korean government agencies banned their employees from downloading the chatbot to their work devices.

South Korea’s acting president Choi Sang-mok has described Deepseek as a “shock”, that could impact the country’s industries, beyond AI.

Despite the suspension of new downloads, people who already have it on their phones will be able to continue using it or they may just access it via DeepSeek’s website.

‘DeepSeek moved me to tears’: How young Chinese find therapy in AI

DeepSeek: The Chinese AI app that has the world talking

DeepSeek: How China’s ‘AI heroes’ overcame US curbs to stun Silicon Valley

ADVERTISEMENT

China’s Deepseek rocked the technology industry, the markets and America’s confidence in its AI leadership, when it released its latest app at the end of last month.

Its rapid rise as one of the world’s favourite AI chatbots sparked concerns in different jurisdictions.

Aside from South Korea, Taiwan and Australia have also banned it from all government devices.

Italy’s regulator, which briefly banned ChatGPT in 2023, has done the same with DeepSeek, which has been asked to address concerns over its privacy policy before it becomes available again on app stores.

Meanwhile, lawmakers in the US have proposed a bill banning DeepSeek from federal devices, citing surveillance concerns.

At the state-government level, Texas, Virginia and New York, have already introduced such rules for their employees.

DeepSeek’s “large language model” (LLM) has reasoning capabilities that are comparable to US models such as OpenAI’s o1, but reportedly requires a fraction of the cost to train and run.

That has raised questions about the billions of dollars being invested into AI infrastructure in the US and elsewhere.

In January, it released its latest model, DeepSeek R1, which it said rivalled technology developed by ChatGPT-maker OpenAI in its capabilities, while costing far less to create.

Its popularity and potential rattled investors, wiping billions of dollars off the market value of chip giant Nvidia – and called into question whether American firms would dominate the booming artificial intelligence (AI) market, as many assumed they would.

President Donald Trump described it as a “wake-up call” for US companies.

What is artificial intelligence?

To understand why DeepSeek has made such a stir, it helps to start with AI and its capability to make a computer seem like a person.

A machine uses the technology to learn and solve problems, typically by being trained on massive amounts of information and recognising patterns.

The end result is software that can have conversations like a person or predict people’s shopping habits.

In recent years, it has become best known as the tech behind chatbots such as ChatGPT – and DeepSeek – also known as generative AI.

These programs again learn from huge swathes of data, including online text and images, to be able to make new content.

But these tools can also create falsehoods and often repeat the biases contained within their training data.

Millions of people use tools such as ChatGPT to help them with everyday tasks like writing emails, summarising text, and answering questions – and others even use them to help with basic coding and studying.

What is DeepSeek?

DeepSeek is the name of a free AI-powered chatbot, which looks, feels and works very much like ChatGPT.

That means it’s used for many of the same tasks, though exactly how well it works compared to its rivals is up for debate.

It is reportedly as powerful as OpenAI’s o1 model – released at the end of last year – in tasks including mathematics and coding.

Like o1, R1 is a “reasoning” model. These models produce responses incrementally, simulating how humans reason through problems or ideas.

Deepseek says it has been able to do this cheaply – researchers behind it claim it cost $6m (£4.8m) to train, a fraction of the “over $100m” alluded to by OpenAI boss Sam Altman when discussing GPT-4.

It has also seemingly be able to minimise the impact of US restrictions on the most powerful chips reaching China.

DeepSeek’s founder reportedly built up a store of Nvidia A100 chips, which have been banned from export to China since September 2022. Some experts believe he paired these chips with cheaper, less sophisticated ones – ending up with a much more efficient process.

DeepSeek also uses less memory than its rivals, ultimately reducing the cost to perform tasks for users.

That combination of performance and lower cost helped DeepSeek’s AI assistant become the most-downloaded free app on Apple’s App Store when it was released in the US.

The same day, it was hit with “large-scale malicious attacks”, the company said, causing the company to temporary limit registrations.

Its website also experienced outages.

Like many other Chinese AI models – Baidu’s Ernie or Doubao by ByteDance – DeepSeek is trained to avoid politically sensitive questions.

When the BBC asked the app what happened at Tiananmen Square on 4 June 1989, DeepSeek did not give any details about the massacre, a taboo topic in China, which is subject to government censorship

Who is behind DeepSeek?

DeepSeek was founded in December 2023 by Liang Wenfeng, and released its first AI large language model the following year.

Not much is known about Mr Liang, who graduated from Zhejiang University with degrees in electronic information engineering and computer science. But he now finds himself in the international spotlight.

He was recently seen at a meeting hosted by China’s premier Li Qiang, reflecting DeepSeek’s growing prominence in the AI industry.

Unlike many American AI entrepreneurs who are from Silicon Valley, Mr Liang also has a background in finance.

He is the CEO of a hedge fund called High-Flyer, which uses AI to analyse financial data to make investment decisions – what is called quantitative trading. In 2019 High-Flyer became the first quant hedge fund in China to raise over 100 billion yuan ($13m).

In a speech he gave that year, Liang said, “If the US can develop its quantitative trading sector, why not China?”

In a rare interview last year, he said China’s AI sector “cannot remain a follower forever” of US AI development.

Asked why DeepSeek’s model surprised so many in Silicon Valley, Liang said: “Their surprise stems from seeing a Chinese company join their game as an innovator, not just a follower – which is what most Chinese firms are accustomed to.”

But it has drawn scrutiny from global leaders.

Australia has banned DeepSeek on government devices and systems, saying it poses a national security risk.

Several data protection authorities around the world have also asked DeepSeek to clarify how it handles personal information – which it stores on China-based servers.

Italy blocked DeepSeek’s app on 30 January and ordered the company to stop processing the personal information of its citizens over data protection concerns.

Why were US companies like Nvidia hit?

DeepSeek’s achievements undercut the belief that bigger budgets and top-tier chips are the only ways of advancing AI, a prospect which has created uncertainty about the future of high-performance chips.

“DeepSeek has proven that cutting-edge AI models can be developed with limited compute resources,” says Wei Sun, principal AI analyst at Counterpoint Research.

“In contrast, OpenAI, valued at $157 billion, faces scrutiny over its ability to maintain a dominant edge in innovation or justify its massive valuation and expenditures without delivering significant returns.”

DeepSeek’s apparently lower costs roiled financial markets on 27 January, leading the tech-heavy Nasdaq to fall more than 3% in a broad sell-off that included chip makers and data centres around the world.

Nvidia’s stock price plunged 17% on Monday before it began to recover on Tuesday.

The chip maker had been the most valuable company in the world, when measured by market capitalisation.

But it fell to third place after Apple and Microsoft on Monday, when its market value shrank to $2.9tn from $3.5tn, Forbes reported.

DeepSeek is a privately owned company, which means investors cannot buy shares of stock on any of the major exchanges.

How has China reacted to DeepSeek’s impact?

DeepSeek’s rise is a huge boost for the Chinese government, which has been seeking to build tech independent of the West.

While the Communist Party is yet to comment, Chinese state media was eager to note that Silicon Valley and Wall Street giants were “losing sleep” over DeepSeek, which was “overturning” the US stock market.

“In China, DeepSeek’s advances are being celebrated as a testament to the country’s growing technological prowess and self-reliance,” says Marina Zhang, an associate professor at the University of Technology Sydney.

“The company’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.”

But she also warned that this sentiment may also lead to “tech isolationism”.

DeepSeek stunned the world in January when it unveiled a chatbot which matched the performance level of US rivals, while claiming it had a much lower training cost.

Billions of dollars were wiped off stock markets internationally, including in Australia, where stocks tied to AI – such as chipmaker Brainchip – fell sharply overnight.

The Australian government has insisted the ban is not due to the app’s Chinese origins but because of the “unacceptable risk” it poses to national security.

DeepSeek has been approached for comment.

Australia’s move specifically requires any government entities to “prevent the use or installation of DeepSeek products, applications and web services”, as well as remove any previously installed, on any government system or device.

That means a wide range of workers will not be able to use the tools in the country, including those working in such varied areas as the Australia Electoral Commission and Bureau of Meteorology.

It is less clear whether it means DeepSeek would be banned from public sector computers in different areas of the economy, such as schools.

The ban does not extend to devices of private citizens.

Growing – and familiar – concerns

Western countries have a track record of being suspicious of Chinese tech – notably telecoms firm Huawei and the social media platform, TikTok – both of which have been restricted on national security grounds.

The initial reaction to DeepSeek – which quickly became the most downloaded free app in the UK and US – appeared to be different.

President Donald Trump described it as a “wake up call” for the US but said overall it could be a positive development, if it lowered AI costs.

Since then, though, doubts about it have started to be voiced.

An Australian science minister previously said in January that countries needed to be “very careful” about DeepSeek, citing “data and privacy” concerns.

The chatbot was removed from app stores after its privacy policy was questioned in Italy. The Italian goverment previously temporarily blocked ChatGPT over privacy concerns in March 2023.

Regulators in South Korea, Ireland and France have all begun investigations into how DeepSeek handles user data, which it stores in servers in China.

The White House press secretary Karoline Leavitt has also said the US is now looking into possible security implications.

The US Navy has reportedly banned its members from using DeepSeek.

Generally, AI tools will analyse the prompts sent to them to improve their product.

This is true of apps such as ChatGPT and Google Gemini as much as it is DeepSeek.

All of them gather and keep information, including email addresses and dates of birth.

However, security experts have previously warned that anyone working on confidential or national security areas needs to be aware of the risk of whatever they enter into chatbots being kept and analysed by the developers of those tools.

DeepSeek has also faced accusations it has unfairly used US tech.

OpenAI has complained that rivals, including in China, are using its own work to make rapid progress with their own products.

The release of China’s new DeepSeek AI-powered chatbot app has rocked the technology industry. It quickly overtook OpenAI’s ChatGPT as the most-downloaded free iOS app in the US, and caused chip-making company Nvidia to lose almost $600bn (£483bn) of its market value in one day – a new US stock market record.

The reason behind this tumult? The “large language model” (LLM) that powers the app has reasoning capabilities that are comparable to US models such as OpenAI’s o1, but reportedly requires a fraction of the cost to train and run.

DeepSeek claims to have achieved this by deploying several technical strategies that reduced both the amount of computation time required to train its model (called R1) and the amount of memory needed to store it. The reduction of these overheads resulted in a dramatic cutting of cost, says DeepSeek. R1’s base model V3 reportedly required 2.788 million hours to train (running across many graphical processing units – GPUs – at the same time), at an estimated cost of under $6m (£4.8m), compared to the more than $100m (£80m) that OpenAI boss Sam Altman says was required to train GPT-4.

Despite the hit taken to Nvidia’s market value, the DeepSeek models were trained on around 2,000 Nvidia H800 GPUs, according to one research paper released by the company. These chips are a modified version of the widely used H100 chip, built to comply with export rules to China. These were likely stockpiled before restrictions were further tightened by the Biden administration in October 2023, which effectively banned Nvidia from exporting the H800s to China. It is likely that, working within these constraints, DeepSeek has been forced to find innovative ways to make the most effective use of the resources it has at its disposal.

Reducing the computational cost of training and running models may also address concerns about the environmental impacts of AI. The data centres they run on have huge electricity and water demands, largely to keep the servers from overheating. While most technology companies do not disclose the carbon footprint involved in operating their models, a recent estimate puts ChatGPT’s monthly carbon dioxide emissions at over 260 tonnes per month – that’s the equivalent of 260 flights from London to New York. So, increasing the efficiency of AI models would be a positive direction for the industry from an environmental point of view.

Of course, whether DeepSeek’s models do deliver real-world savings in energy remains to be seen, and it’s also unclear if cheaper, more efficient AI could lead to more people using the model, and so an increase in overall energy consumption.

If nothing else, it could help to push sustainable AI up the agenda at the upcoming Paris AI Action Summit so that AI tools we use in the future are also kinder to the planet.

What has surprised many people is how quickly DeepSeek appeared on the scene with such a competitive large language model – the company was only founded by Liang Wenfeng in 2023, who is now being hailed in China as something of an “AI hero”.

The model is constructed from a group of much smaller models, each having expertise in specific domains

The latest DeepSeek model also stands out because its “weights” – the numerical parameters of the model obtained from the training process – have been openly released, along with a technical paper describing the model’s development process. This enables other groups to run the model on their own equipment and adapt it to other tasks.

This relative openness also means that researchers around the world are now able to peer beneath the model’s bonnet to find out what makes it tick, unlike OpenAI’s o1 and o3 which are effectively black boxes. But there are still some details missing, such as the datasets and code used to train the models, so groups of researchers are now trying to piece these together.

Not all of DeepSeek’s cost-cutting techniques are new either – some have been used in other LLMs. In 2023, Mistral AI openly released its Mixtral 8x7B model which was on par with the advanced models of the time. Mixtral and the DeepSeek models both leverage the “mixture of experts” technique, where the model is constructed from a group of much smaller models, each having expertise in specific domains. Given a task, the mixture model assigns it to the most qualified “expert”.

DeepSeek has even revealed its unsuccessful attempts at improving LLM reasoning through other technical approaches, such as Monte Carlo Tree Search, an approach long touted as a potential strategy to guide the reasoning process of an LLM. Researchers will be using this information to investigate how the model’s already impressive problem-solving capabilities can be even further enhanced – improvements that are likely to end up in the next generation of AI models.

So what does this all mean for the future of the AI industry?

DeepSeek is potentially demonstrating that you don’t need vast resources to build sophisticated AI models. My guess is that we’ll start to see highly capable AI models being developed with ever fewer resources, as companies figure out ways to make model training and operation more efficient.

Up until now, the AI landscape has been dominated by “Big Tech” companies in the US – Donald Trump has called the rise of DeepSeek “a wake-up call” for the US tech industry. But this development may not necessarily be bad news for the likes of Nvidia in the long term: as the financial and time cost of developing AI products reduces, businesses and governments will be able to adopt this technology more easily. That will in turn drive demand for new products, and the chips that power them – and so the cycle continues.

It seems likely that smaller companies such as DeepSeek will have a growing role to play in creating AI tools that have the potential to make our lives easier. It would be a mistake to underestimate that.

The status of OpenAI – and other US firms – as the world leaders in AI has been dramatically undermined this week by the sudden emergence of DeepSeek, a Chinese app that can emulate the performance of ChatGPT, apparently at a fraction of the cost.

Bloomberg has reported that Microsoft is investigating whether data belonging to OpenAI – which it is a major investor in – has been used in an unauthorised way.

The BBC has contacted Microsoft and DeepSeek for comment.

OpenAI’s concerns have been echoed by the recently appointed White House “AI and crypto czar”, David Sacks.

Speaking on Fox News, he suggested that DeepSeek may have used the models developed by OpenAI to get better, a process called knowledge distillation.

“There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models,” Mr Sacks said.

“I think one of the things you’re going to see over the next few months is our leading AI companies taking steps to try and prevent distillation… That would definitely slow down some of these copycat models.”

In a statement, OpenAI said Chinese and other companies were “constantly trying to distil the models of leading US AI companies”.

“As we go forward…it is critically important that we are working closely with the U.S. government to best protect the most capable models,” it added.

The accusation of disrespecting intellectual property rights is however far from a new one in tech – and has been frequently levelled at major US AI firms.

US officials are also considering the national security implications of DeepSeek’s emergence, according to White House press secretary Karoline Leavitt.

“I spoke with [the National Security Council] this morning, they are looking into what [the national security implications] may be,” said Ms Leavitt, who also restated US President Donald Trump’s remarks a day earlier that DeepSeek should be a wake-up call for the US tech industry.

The announcement comes after the US navy reportedly banned its members from using DeepSeek’s apps due to “potential security and ethical concerns”.

According to CNBC, the US navy has sent an email to its staff warning them not to use the DeepSeek app due to “potential security and ethical concerns associated with the model’s origin and usage”.

The Navy did not immediately respond to a request for comment from BBC News.

Data safety experts have warned users to be careful with the tool, given it collects large amounts of personal data and stores it in servers in China.

Meanwhile, DeepSeek says it has been the target of cyber attacks. On Monday it said it would temporarily limit registrations because of “large-scale malicious attacks” on its software.

A banner showing on the company’s website says registration may be busy as a result of the attacks.

Yuyuan Tantian, a social media channel under China’s state broadcaster CCTV, claims the firm has faced “several” cyber attacks in recent weeks, which have increased in “intensity”.

America’s AI industry has been shaken by DeepSeek’s apparent breakthrough, especially because of the prevailing view that the US was far ahead in the race.

A slew of trade restrictions banning China’s access to high-end chips was believed to have cemented this.

Although China has boosted investment in advanced tech to diversify its economy, DeepSeek is not one of the big Chinese firms that have been developing AI models to rival US-made ChatGPT.

Experts say the US still has an advantage – it is home to some of the biggest chip companies – and that it’s unclear yet exactly how DeepSeek built its model and how far it can go.

The announcement comes after the US navy reportedly banned its members from using DeepSeek’s apps due to “potential security and ethical concerns”.

Meanwhile, the maker of ChatGPT, OpenAI, has promised to work closely with the US government to prevent rivals from taking its technology.

Earlier this week, DeepSeek’s reportedly cheap yet powerful AI model caused a slump in the stocks of US technology firms as investors questioned the billions of dollars they are spending on new AI infrastructure.

“I spoke with [the National Security Council] this morning, they are looking into what [the national security implications] may be,” said Ms Leavitt, who also restated US President Donald Trump’s remarks a day earlier that DeepSeek should be a wake-up call for the US tech industry.

According to CNBC, the US navy has sent an email to its staff warning them not to use the DeepSeek app due to “potential security and ethical concerns associated with the model’s origin and usage”.

The Navy did not immediately respond to a request for comment from BBC News.

Speaking on Fox News, the recently appointed “White House AI and crypto czar”, David Sacks, also suggested that DeepSeek may have used the models developed by top US firm OpenAI to get better.

This process – which involves one AI model learning from another – is called knowledge distillation.

“There’s substantial evidence that what DeepSeek did here is they distilled the knowledge out of OpenAI’s models,” Mr Sacks said. “I think one of the things you’re going to see over the next few months is our leading AI companies taking steps to try and prevent distillation… That would definitely slow down some of these copycat models.”

OpenAI echoed this in a later statement that said Chinese and other companies are “constantly trying to distill the models of leading US AI companies.”

“As the leading builder of AI, we engage in countermeasures to protect our [intellectual property]… and believe as we go forward that it is critically important that we are working closely with the U.S. government to best protect the most capable models”.

Meanwhile, DeepSeek says it has been the target of cyber attacks. On Monday it said it would temporarily limit registrations because of “large-scale malicious attacks” on its software.

A banner currently showing on the company’s website says registration may be busy as a result of the attacks.

Yuyuan Tantian, a social media channel under China’s state broadcaster CCTV, claims the firm has faced “several” cyber attacks in recent weeks, which have increased in “intensity”.

DeepSeek shot to fame last week as AI geeks lauded its latest AI model and people began downloading its chatbot on app stores. Its rise caused a slump in US tech stocks, many of which have since recovered some ground.

But America’s AI industry was shaken by the apparent breakthrough, especially because of the prevailing view that the US was far ahead in the race. A slew of trade restrictions banning China’s access to high-end chips was believed to have cemented this.

Although China has boosted investment in advanced tech to diversify its economy, DeepSeek is not one of the big Chinese firms that have been developing AI models to rival US-made ChatGPT.

Experts say the US still has an advantage – it is home to some of the biggest chip companies – and that it’s unclear yet exactly how DeepSeek built its model and how far it can go.

As DeepSeek rattled markets this week, President Trump described it as “a wake-up call” for the US tech industry, while suggesting that it could ultimately prove to be a “positive” sign.

“If you could do it cheaper, if you could do it [for] less [and] get to the same end result. I think that’s a good thing for us,” he told reporters on board Air Force One.

He also said he was not concerned about the breakthrough, adding the US will remain a dominant player in the field.

Shares in chip giant Nvidia were up over 6% by mid-day trade having sank on Monday, as experts said the US AI sell-off may have been an over-reaction.

The market hit came as investors rapidly adjusted bets on AI, after DeepSeek’s claim that its model was made at a fraction of the cost of those of its rivals.

Analysts said the development raised questions about the future of America’s AI dominance and the scale of investments US firms are planning.

US President Donald Trump described the moment as “a wake-up call” for the US tech industry, while also suggesting that it could ultimately prove ” a positive” for the US.

“If you could do it cheaper, if you could do it [for] less [and] get to the same end result. I think that’s a good thing for us,” he told reporters on board Air Force One.

He also said he was not concerned about the breakthrough, adding the US will remain a dominant player in the field.

Optimism about AI investments has powered much of the boom in US stock markets over the last two years, raising fears of a possible bubble.

DeepSeek has become the most downloaded free app in the US just a week after it was launched.

Its emergence comes as the US has been warning of a tech race with China, and taking steps to restrict the sale of the advanced chip technology that powers AI to China.

To continue their work without steady supplies of imported advanced chips, Chinese AI developers have shared their work with each other and experimented with new approaches to the technology.

This has resulted in AI models that require far less computing power than before.

It also means that they cost a lot less than previously thought possible, which has the potential to upend the industry.

Nvidia – the company behind the high-tech chips that dominate many AI investments, that had seen its share price surge in the last two years due to growing demand – was the hardest hit on Monday.

Its share price dropped by roughly 17% on Monday, wiping roughly $600bn off its market value.

Janet Mui, head of market analysis at RBC Brewin Dolphin, said investors’ first response to something that appears groundbreaking is to sell because of the uncertainty.

But she said she expected many companies, like Apple, to benefit if the cost of AI models becomes becomes cheaper.

It could also be a boon for other tech giants, which have faced scrutiny for their high spending on AI so far.

Following the shock to markets in the US on Monday, the main indexes were steady.

As of mid-day trading in New York, the Dow Jones was roughly flat, the S&P was up about 0.8% and the tech-heavy Nasdaq had risen 1.7%.

The FTSE 100 stock index of the UK’s biggest publicly-listed companies was also steady on Tuesday, closing 0.35% higher.

Earlier shares in Japanese AI-related firms including Advantest, Softbank and Tokyo Electron fell sharply, helping to push the benchmark Nikkei 225 down by 1.4%.

Several other markets in Asia were closed for the Lunar New Year holiday. Mainland China’s financial markets will be shut from Tuesday and will reopen on 5 February.

The company was founded in 2023 by Liang Wenfeng in Hangzhou, a city in southeastern China.

The 40-year-old, an information and electronic engineering graduate, also founded the hedge fund that backed DeepSeek.

He was recently seen at a meeting between industry experts and the Chinese premier Li Qiang.

In a July 2024 interview with The China Academy, Mr Liang said he was surprised by the reaction to the previous version of his AI model.

“We didn’t expect pricing to be such a sensitive issue,” he said.

“We were simply following our own pace, calculating costs, and setting prices accordingly.”

After DeepSeek-R1 was launched earlier this month, the company boasted of “performance on par with” one of OpenAI’s latest models when used for tasks such as maths, coding and natural language reasoning.

DeepSeek’s technology has been praised by high profile figures including OpenAI chief Sam Altman who called it “an impressive model, particularly around what they’re able to deliver for the price”, though he added that OpenAI would “obviously deliver much better models” moving forward.

“DeepSeek’s ability to rival US models despite limited access to advanced hardware demonstrates that software ingenuity and data efficiency can compensate for hardware constraints,” said Marina Zhang, an associate professor at the University of Technology Sydney, who focuses on China’s high-tech industries.

Ion Stoica, co-founder and executive chair of AI software company Databricks, told the BBC the lower cost of DeepSeek could spur more companies to adopt AI in their business.

“If that happens, this reduction in cost can accelerate the progress of AI,” he said. “So overall, the market will expand faster, and the value of the market will grow faster.”

The Chinese company claims its model can be trained on 2,000 specialised chips compared to an estimated 16,000 for leading models.

But not everyone is convinced. Some have cast doubt on some of DeepSeek’s claims, including tech mogul Elon Musk.

He responded to a post which claimed that DeepSeek actually has around 50,000 Nvidia chips that have now been banned from export to China, saying: “Obviously.”

The sudden explosion in popularity has prompted some to raise cyber security concerns.

In Australia, science minister Ed Husic was among the experts urging caution, telling Australia’s national broadcaster ABC: “There are a lot of questions that will need to be answered in time on quality, consumer preferences, data and privacy management.

This dominance has been mainly down to enormous capital investment – but China’s DeepSeek was developed for a fraction of the price of its US rivals. Its sudden debut has had a huge impact, wiping $1tn off the value of US tech stocks.

The efficiency and capability of DeepSeek’s model should not be underestimated.

All of this has been achieved using lower-end technology, a consequence of US restrictions on the export to China of high-tech components – Nvidia’s H100 chip at the higher end and its H800 chip at the lower end, both of which are commonly used in AI.

The US barred its export over fears that China could challenge American AI dominance if given unfettered access to Silicon Valley technology, so a viable AI model created in this environment speaks for itself in terms of ingenuity and potential.

Despite this, it is what DeepSeek represents, rather than what it has produced, that may ultimately be its lasting legacy.

It highlights a new way of thinking about the economics of the AI industry.

It levels the playing field for governments and companies with aspirations to become AI power players.

And it demonstrates that innovation born of necessity can produce results with the power to make the money markets rethink the economic direction of travel.

Many already felt the US AI industry was rife with inflated valuations, leading to talk of an AI bubble. That bubble hasn’t quite burst, but its structural integrity is certainly now under strain.

Some may interpret DeepSeek’s impact as a sign that the seat of AI power is shifting eastward – but it’s also possible that innovators worldwide will now take inspiration and attempt to develop their own lower-cost AI technologies.

The investment plans announced in the US – worth hundreds of billions of dollars – were simply not replicable elsewhere, but that may no longer be such a problem.

The UK government has made clear its intention to use AI as an economic driver. If lower development costs become the norm, this ambition may become more attainable.

The UK has never lacked innovators, but British businesses have often struggled to scale without significant overseas investment. A cheaper, more resourceful approach to AI could help the UK and other governments realise the goal of cultivating homegrown AI powerhouses.

The US tech giants, however, are unlikely to take this lying down. They may have been given a bloody nose by the markets, but they still have enormous technical and financial resources at their disposal.

These companies are already exploring new ways of monetising their AI technologies and finding applications for AI across public life.

However, they may now face a tightening of belts and lowering of valuation expectations as a new economic reality kicks in.

Chinese tech, from Huawei to TikTok, has repeatedly been the subject of allegations the firms are linked to the Chinese state, and fears this could lead to peoples’ data being harvested for intelligence purposes.

Donald Trump has said DeepSeek is a “wake up call” for the US but did not seem to suggest it was a threat to national security – instead saying it could even be a good thing if it brought costs down.

But Husic told ABC News on Tuesday there remained a lot of unanswered questions, including over “data and privacy management.”

“I would be very careful about that, these type of issues need to be weighed up carefully,” he added.

DeepSeek has not responded to the BBC’s request for comment – but users in the UK and US have so far shown no such caution.

DeepSeek has rocketed to the top of the app stores in both countries, with market analysts Sensor Tower saying it has seen 3 million downloads since launch.

As much as 80% of these have come in the past week – meaning it has been downloaded at three times the rate of rivals such as Perplexity.

According to DeepSeek’s own privacy policy, it collects large amounts of personal information collected from users, which is then stored “in secure servers” in China.

This may include:

• Your email address, phone number and date of birth, entered when creating an account
• Any user input including text and audio, as well as chat histories
• So-called “technical information” – ranging from your phone’s model and operating system to your IP address and “keystroke patterns”.

It says it uses this information to improve DeepSeek by enhancing its “safety, security and stability”.

It will then share this information with others, such as service providers, advertising partners, and its corporate group, which will be kept “for as long as necessary”.

“There are genuine concerns around the technological potential of DeepSeek, specifically around the terms of its privacy policy,” said ExpressVPN’s digital privacy advocate Lauren Hendry Parsons.

She specifically highlighted the part of the policy which says data can be used “to help match you and your actions outside of the service” – which she said “should immediately ring an alarm bell for anyone concerned with their privacy”.

But while the app harvests a lot of data, experts point out it’s very similar to privacy policies users may have already agreed to for rival services like ChatGPT and Gemini, or even social media platforms.

“For any openly available AI model, with a web or app interface – including but not limited to DeepSeek – the prompts, or questions that are asked of the AI, then become available to the makers of that model, as are the answers,” said Emily Taylor, chief executive of Oxford Information Labs

“So, anyone working on confidential or national security areas needs to be aware of those risks,” she told the BBC.

Dr Richard Whittle from University of Salford said he had “various concerns about data and privacy” with the app, but said there were “plenty of concerns” with the models used in the US too.

“Consumers should always be wary, especially in the hype and fear of missing out on a new, highly popular, app,” he said.

The UK data regulator, the Information Commissioner’s Office has urged the public to be aware of their rights around their information being used to train AI models.

Asked by BBC News if it shared the Australian government’s concerns, it said in a statement: “Generative AI developers and deployers need to make sure people have meaningful, concise and easily accessible information about the use of their personal data and have clear and effective processes for enabling people to exercise their information rights.

“We will continue to engage with stakeholders on promoting effective transparency measures, without shying away from taking action when our regulatory expectations are ignored.”