Mitchell Labiak
BBC Business reporter

It's just gone past 13:00 in New York (18:00 GMT) and the US stock market seems to have calmed after yesterday's sell off.

Some investors clearly feel the reaction to the sudden rise in popularity of Chinese AI app DeepSeek was overdone, so they have been buying shares in the firms that sank.

But the rebound has not been felt equally by everyone.

Microsoft has climbed over 2% today to higher than it was before Monday's sell off.

Meanwhile, though Nvidia climbed 6% today, it is still far off from recovering the value it lost yesterday - which was the biggest single-day loss in Wall Street history.

Trading on the US stock market should return to normal from tomorrow after two days of reaction to DeepSeek.

This is partly because markets are quick to react and speedily re-value stocks based on new information but also because DeepSeek has "gone quiet".

It has entered "holiday mode", according to the South China Morning Post, for the Lunar New Year holiday.

That's all for our live coverage, thank you for joining us.

Image source, EPA

As US tech stocks steady following a groundbreaking Monday for Chinese AI app DeepSeek, we'll be ending our live coverage soon.

Tech giant Nvidia is showing slight signs of recovery after yesterday's monumental plummet - up over 6% today after its 17% loss in value on Monday.

Less than 24 hours after praising DeepSeek's performance, competitor OpenAI launched a new version of ChatGPT for government agencies.

And Siemens Energy - one of the biggest losers from Monday's market sell-off - is up over 7% on Tuesday.

For more on DeepSeek, we've plenty of material to sink your teeth into:

• How China's 'AI heroes' overcame US curbs to stun Silicon Valley
• China's DeepSeek AI shakes industry and dents America's swagger
• DeepSeek vs ChatGPT - how do they compare?
• DeepSeek shows AI's centre of power could shift away from US

Did it really cost $6m to ‘make’ DeepSeek?published at 17:51 Greenwich Mean Time 28 January

17:51 GMT 28 January

Tom Gerken
Technology reporter

That widely-reported figure is at the heart of what has caused such consternation in the stock market - the idea that this impressive AI toolkit was developed for little compared to its big budget US-based rivals.

The true figure is much harder to say for sure.

Firstly, it all comes from the first table in DeepSeek’s publicly available report , externalon how the tool works. It estimates the total cost of training DeepSeek V3 at just under $5.6m (£4.5m).

Some people have challenged this figure, but if we take it at face value, that would still only cover the cost of training this version of its AI - DeepSeek-V3 came as a result of various previous iterations.

Its researchers state in the report that the $5.6m figure excludes the costs of its previous research in the space (which has been ongoing since 2023), as well as the other experiments they tried before landing on this methodology.

But more broadly, the number also ignores all the other day to day costs that add up, from renting or buying an office space to paying for staff - Chinese media say DeepSeek’s team comprises of just under 140 people, who all presumably earn a salary.

And that doesn’t consider the cost of actually deploying DeepSeek into the sleek, easy-to-use app that has exploded in popularity on app stores in the UK and US.

So while researchers say it cost less than $6m to train this particular version of DeepSeek, the true cost will have been much higher.

Tom Gerken
Technology reporter

As we reported earlier, OpenAI has offered up its latest product in the face of a new rival - a version of ChatGPT built specifically to be used by the US government.

Government agencies will be able to submit “non-public, sensitive information” securely to be used by GPT, in order to help speed up their work.

It will formally enable government workers to use AI to translate documents, summarise them, write emails and so on.

It’s unlikely to be directly linked to DeepSeek - these sorts of products can take months and years to create and rollout - but the timing of the announcement is no less suspicious.

And this isn’t just something being worked on in the US - the UK government announced in January it would give civil servants a set of AI tools to perform similar tasks, named "Humphrey" after the scheming official from the classic sitcom Yes, Minister.

What DeepSeek AI means for the US' relationship with Chinapublished at 17:40 Greenwich Mean Time 28 January

17:40 GMT 28 January

Michelle Fleury
US business correspondent

Image source, Reuters

DeepSeek's impressive artificial intelligence model, reportedly made on the cheap, has shaken America’s confidence in its AI lead over China, not to mention the $600bn (£482) hole it blew in Nvidia’s market value on Monday.

The DeepSeek news came two weeks after the outgoing Biden administration tightened controls on computer chip exports, which were meant to contain China's AI advance.

The Trump administration hasn’t said if it will implement the new controls.

Although export controls undoubtedly forced China to be more creative and efficient, experts argue that DeepSeek is proof that America needs to double down on export controls.

"They will continue to have a big impact on China's overall AI ecosystem," Matt Sheehan, an AI and tech expert focusing on China at the Carnegie Endowment for International Peace, tells me.

"Especially when it comes to how frequently these models can be used and how many people can be using the model at the same time."

If export controls are tightened or new ones added, he believes they will continue to be effective.

DeepSeek shows AI's centre of power could shift away from USpublished at 17:34 Greenwich Mean Time 28 January

17:34 GMT 28 January

Marc Cieslak
AI correspondent

Image source, Getty Images

DeepSeek's arrival at the top of the Apple App Store charts has placed it firmly in the public consciousness, shaking the belief that the US would continue as the largely unchallenged global superpower of AI.

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.

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.

• Read more about the possibility that innovators worldwide will now take inspiration and attempt to develop their own lower-cost AI technologies

Mitchell Labiak
Business reporter

That's how analyst Alex Rudolph at investment platform IG has described today's activity in the US stock market.

That "wall of worry", of course, is the fear that DeepSeek's sudden popularity means that US tech stocks related to its AI industry are overvalued.

Today, however, Rudolph says investors were "bargain hunting" after yesterday's sell off.

For example, a share in Nvidia cost $142 on Friday, slumped to a low of around $118 midway through Monday, and climbed to $121 by midway through trading on Tuesday - having risen over 2% today as investors consider whether the slump yesterday was overdone.

Meanwhile, Microsoft has almost recovered most of its lost value. It plunged from $443 on Friday to a low of around $424 on Monday but had crept back up to close to $442 by midway through trading on Tuesday.

Faisal Islam
Economics editor

Image source, Getty Images

China's success in technology has not come out of thin air, even given the unlikely origins of the DeepSeek deep shock.

The obscure Hangzhou hedge fund that coded a ChatGPT competitor as a side project it claims cost just $6m emerges from a concerted effort to invest in future generations of technology.

This is not an accident. This is policy.

The raw materials of artificial intelligence (AI) are microchips, science PhDs and data. On the latter two, China might be ahead already.

There are on average more than 6,000 PhDs in STEM subjects (science, technology, engineering and mathematics) coming out of Chinese universities every month. In the US it is more like 2,000-3,000, in the UK it is 1,500.

Delegates who accompanied the UK Chancellor Rachel Reeves to China earlier this month marvelled at how the Beijing air had been cleaned up, and indigenous electric cars were everywhere. Another UK CEO told me of a visit to Huawei's Oxbridge-style campus complete with spires and bridges, and its own Tube line, purely for its scientists.

Clearly, however, there are concerns about censorship, democracy and security.

• Read more about how there is still an obvious balancing act for the UK government

Not such an overnight success story for Deepseekpublished at 17:11 Greenwich Mean Time 28 January

17:11 GMT 28 January

Tom Gerken
Technology reporter

With the sudden impact it has had on stock markets around the world, you would be forgiven for thinking that DeepSeek was an overnight success.

In reality, it is the result of multiple iterations, with versions of DeepSeek having been publicly released since 2023.

Looking at how these models have developed tells a story of a team which has repeatedly looked for (and found) significant ways to improve the efficiency of AI models in general.

Their solutions are quite complex, but the end result is a highly powerful large language model that is open source.

The most surprising part of this new model - as you have no doubt read a hundred times over the past two days - is that it runs comparatively to some of the big boys in various tasks, and even wins out in some tests.

Even that leap forward wasn’t necessarily the thing that shook the markets - the shockwaves came after DeepSeek released its own chatbot app to rival ChatGPT.

Janet Mui, head of market analysis at RBC Brewin Dolphin, says yesterday's sell-off of Nvidia was "likely an overreaction".

When something groundbreaking happens, she explains, investors' first instinct is to sell because of the uncertainty.

But with the popularisation of artificial intelligence, there will be more demand for AI infrastructure in the future, she tells BBC News.

Some companies, like Apple, will benefit if the cost of AI models is cheaper - as they'll be able to integrate more AI into their phones, Mui says.

Tech giants have also been "scrutinised" for their "high spending" on AI so far, so AI users will benefit if the cost becomes cheaper, she adds.

Nvidia were the "darlings of AI development" in 2024, she says, but this year investors will focus more on the "use case" of AI.

Image source, Getty Images

Yesterday, DeepSeek took Wall Street by a storm, causing tech-focused shares in the US to fall - but not because of what you might expect.

What's key about DeepSeek isn't its capabilities compared to other AI models. It's that DeepSeek claims to do the same sorts of things as its competitors (like OpenAI, or Meta AI) for much cheaper.

The researchers behind the Chinese AI start-up claim it was trained for around $6m (£4.2m) - significantly less than the billions spent by rivals.

What DeepSeek isn't currently doing is reasoning like a human being. This is known as Artificial General Intelligence (AGI), which is a type of AI.

AGI is a term for a human-level intelligence that no technology firm has yet achieved which DeepSeek hopes to achieve. Others in Silicon Valley, like Open AI, are also pursuing this.

So, what's captured the attention of investors and disrupted the stock market isn't necessarily the capabilities of DeepSeek. Rather, its claim to have "performance on par with" one of OpenAI's latest models at a fraction of the price.

Mitchell Labiak
Business reporter

One of the biggest losers from Monday’s market sell-off has arguably turned into the biggest winner today.

Siemens Energy, a European tech giant which makes hardware related to AI, lost over a fifth of its value after news of DeepSeek’s sudden popularity spread to European markets, but it is up over 7% today.

That's a much better recovery than Nvidia, external, Microsoft, or Alphabet, Google's owner, had today.

Siemens Energy's share price bump comes after it published better-than-expected financial results this morning. This happens a lot in stock markets.

Investors usually base their valuation of a company on predictions of their results. So, when firms beat those expectations, the share price tends to go up.

But the news is particularly good timing for Siemens Energy after yesterday’s slump.

Graham Fraser
Technology reporter

Image source, Reuters

The emergence of Chinese AI app DeepSeek has shocked financial markets and prompted US President Donald Trump to describe it as "a wake-up call" for the US tech industry.

DeepSeek's claim that its R1 artificial intelligence model was made at a fraction of the cost of its rivals has raised questions about the future about of the whole industry.

So how does it compare to its much more established and apparently much more expensive US rivals, such as OpenAI's ChatGPT and Google's Gemini?

I had a look into four areas - writing assistance, coding, brainstorming, and research. On simple tasks and prompts, DeepSeek was certainly comparable to its US competitors.

Prof Neil Lawrence, DeepMind Professor of Machine Learning at Department of Computer Science and Technology, at the University of Cambridge, said this was just the start.

"I think it's just the tip of the iceberg in terms of the type of innovation we can expect in these models.”

• Read more in this summary about their comparisons

The US stock market hasn't been open for very long, but tech giant Nvidia is showing slight signs of recovery from yesterday's plummet.

The graph below shows how the share price in the AI chip company is already seeing an incline today, reaching $120 again.

It is early in the trading day in the US, however. Stay with us, as we continue monitoring how prices may change.

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

But while DeepSeek 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.

According to DeepSeek's own privacy policy, external 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".

Is DeepSeek a national security threat? The TikTok factor looms largepublished at 16:22 Greenwich Mean Time 28 January

16:22 GMT 28 January

Joe Tidy
Cyber correspondent

Amidst the small but growing concerns that DeepSeek might pose some sort of data protection risk to Western users, one company keeps getting brought up - TikTok.

It’s fascinating that DeepSeek has risen to prominence whilst TikTok’s future hangs in the balance over national security concerns in the US.

To critics, DeepSeek should be viewed like TikTok - with huge suspicion because it is based in China.

If anything, DeepSeek is even more Chinese than TikTok. TikTok at least has some distance from Beijing with its international CEO and data security protocols.

But it’s notable that so far only Australia has raised any alarm about DeepSeek as a risk.

Whether or not these concerns spread to other countries largely depends on how the US treats the hit new AI app.

And how the US handles national security concerns about TikTok in the next couple of months could determine DeepSeek’s future.

Rachel Flynn
Live reporter

Image source, EPA

Let's take stock of what's happened over the past 24 hours.

As the US stock market opened on Monday, Chinese-made artificial intelligence (AI) model DeepSeek stunned investors and sank some tech stocks.

Chip-making giant Nvidia shed almost $600bn (£482bn) of its market value - the biggest one-day loss in US history. Other tech stocks such as Microsoft, Google-owner Alphabet and Dell Technologies also sank - albeit by smaller margins.

A day later, it seems slightly calmer. Nvidia shares had a mixed start as the US stock market opened, business reporter Nick Edser tells us. After flip-flopping slightly in the first few minutes, its value recovered - standing 1.3% higher.

In Asia, the markets in Taiwan, China and South Korea have closed for the Lunar New Year. But by the end of Tuesday's trading, Tokyo's Nikkei 225 index had dropped 1.39%, while Hong Kon's Hang Seng made a small 0.14% gain.

For now, any shockwaves sent by DeepSeek seemed to land on Monday.

But we'll be keeping a close eye on its continued impact across the global markets - so stay with us.

Fan Wang and João da Silva
BBC News

US stock markets seem to have stabilised this morning after DeepSeek caused stocks in major tech giants to plummet yesterday - but the app's success hasn't come without challenges.

When the US barred the world's leading chip-makers such as Nvidia from selling advanced tech to China, it was certainly a blow.

Those chips are essential for building powerful AI models that can perform a range of human tasks, from answering basic queries to solving complex maths problems.

Long before the ban, DeepSeek acquired a "substantial stockpile" of Nvidia A100 chips - estimates range from 10,000 to 50,000 - according to the MIT Technology Review.

Leading AI models in the West use an estimated 16,000 specialised chips. But DeepSeek says it trained its AI model using 2,000 such chips, and thousands of lower-grade chips - which is what makes its product cheaper.

• Read more about how DeepSeek's launch has astonished Silicon Valley and the world

Lily Jamali
North America Technology Correspondent

While ChatGPT-maker OpenAI has been haemorrhaging money - spending $5bn (£4bn) last year alone - DeepSeek's developers say it built this latest model for a mere $5.6m (£4.5m).

That is a tiny fraction of the cost that AI giants like OpenAI, Google, and Anthropic have relied on to develop their own models.

Many observers, investors, and analysts appeared stunned.

Some wondered if this marked a buying opportunity. Others questioned the information DeepSeek was providing.

Veteran analyst Gene Munster, speaking to me on Monday, questioned the financials DeepSeek is citing, and wondered if the startup was being subsidised or whether its numbers were correct.

Regardless, its sudden arrival is a "flex" by China and a "black eye for US tech" to use his own words.

Nick Edser
Business reporter

While tech stocks appear - so far - to have arrested yesterday’s dramatic decline, the “DeepSeek shock” has left “US tech sector stocks battered and bruised compared to before the weekend”, according to Thierry Wizman, an analyst at Macquarie Group.

However, he points out that while some tech stocks saw steep falls on Monday, the majority of stocks on the major indexes ended the day higher.

While the arrival of DeepSeek might be bad news for some firms, it could benefit others.

“Whatever its ultimate impact on AI hardware revenue might be, cheaper access to AI services can help reduce costs for AI users and promote earlier adoption,” he says.