DeepSeek AI live: Chatbot vanishes in Italy amid claims OpenAI’s model was used to train Chinese AI

TechDeepSeek says its AI model is similar to US giants like OpenAI, despite fears of censorship around issues sensitive to Beijing

Trump views Chinese DeepSeek AI as ‘wake-up call’ for US

Sign up to our free weekly IndyTech newsletter delivered straight to your inboxSign up to our free IndyTech newsletterSign up to our free IndyTech newsletter

I would like to be emailed about offers, events and updates from The Independent. Read our privacy policy

OpenAI has reportedly claimed it has evidence that Chinese competitor DeepSeek used the company’s own model to train its rival chatbot.

The release of Deepseek’s open source R1 model has shocked Silicon Valley and caused tech shares to plunge, with the Chinese startup’s supposedly low cost model prompting investors to question huge spending plans by leading AI firms in the United States.

OpenAI and its major backer, Microsoft, have been investigating whether DeepSeek obtained data in an unauthorised manner, after observing some individuals exporting large amounts of data from OpenAI’s products, Bloomberg News reported.

The Financial Times also reported that OpenAI, led by Sam Altman, said it had seen some evidence of “distillation”, which it suspects to be from DeepSeek. Distillation refers to the process of transferring knowledge from a large model to a smaller one.

Global share prices steadied on Wednesday. Japan’s Topix index rose by 0.7%, while Australia’s ASX rose by 2.9%. The FTSE 100 was roughly flat at the opening bell.

Attention now turns to Wall Street, with markets opening later in the day.

Why does DeepSeek keep going down?Much has been made about how cheaply DeepSeek has been made, and the fact it is as good as far more expensive models. But much of the cost of an AI model like DeepSeek or ChatGPT is in keeping it online – which might explain the lower cost, and the fact that the system keeps breaking. That’s according to Yann LeCun, one of the godfathers of AI who now works at Meta (one of the companies reportedly at panic stations over the release).

In this Threads post, he lays out that theory.

(Threads/Yann LeCun)Andrew Griffin29 January 2025 13:36

‘People don’t know what they own’: Experts explain DeepSeek-related record sell-off after Nvidia shares riseWhy did all that happen? Maybe because people aren’t entirely sure what’s going on. Here’s the latest from our business desk.

Andrew Griffin29 January 2025 13:20

AI crosses ‘red line’ after learning to replicate itselfAway from DeepSeek, AI is continuing to advance. Here’s Anthony Cuthbertson on the artificial intelligence’s ability to “self-replicate”.

Andrew Griffin29 January 2025 13:18

European shares hit record high as ASML lifts tech stocksEuropean shares hit a record high on Wednesday, as chip equipment maker ASML led technology stocks higher after reporting strong quarterly results.

The pan-European STOXX 600 .STOXX was up 0.5%, as of 0923 GMT, rising past Tuesday’s all-time high close.

ASML’s shares ASML.AS jumped 11.2% after the Dutch company reported better-than-expected fourth-quarter bookings of 7.09 billion euros ($7.39 billion).

ASML’s earnings also reassured investors that AI chip prospects were still healthy, as they recovered from a selloff earlier this week sparked by the release of China’s DeepSeek’s model, which uses less computing power than those of rivals.

STMicroelectronics STMPA.PA, BE Semiconductor BESI.AS and ASM International ASMI.AS gained between 2.1% and 7.2%.

Technology .SX8P was the top winning sector, soaring 4.5%, and was set for its best single day in a year.

“Today you see that at least ASML is still profiting from demand from the chip sector… It seems that for now, the fears for the whole AI and supply chain for AI have been overblown,” said Teeuwe Mevissen, Senior Market Economist at Rabobank.

Other AI-exposed stocks such as Schneider Electric SCHN.PA and Siemens Energy ENR1n.DE, which took a hit on Monday, were trading 5.8% and 1.7% higher, respectively.

Jabed Ahmed29 January 2025 10:59

DeepSeek down amid huge interestDeepSeek doesn’t seem to be working properly this morning. The service is intermittent and does seem to function occasionally.

The app’s official status page says that it has been suffering from “degraded performance”, but says that a fix was rolled out overnight.

Andrew Griffin29 January 2025 10:46

DeepSeek says it built its chatbot cheap. What does that mean for AI’s energy needs and the climate?AI uses vast amounts of energy, much of which comes from burning fossil fuels, which causes climate change.

Tech companies have said their electricity use is going up, when it was supposed to be ramping down, ruining their carefully-laid plans to address climate change.

Making AI more efficient could be less taxing on the environment, experts say, even if its huge electricity needs are not going away.

When there’s an innovative technology that’s useful to the general population and it’s affordable, people will use it, said Vic Shao, founder of DC Grid, which delivers off-grid, direct current power to data centers and electric vehicle charging stations.

That means data centers will still be built, though they may be able to operate more efficiently, said Travis Miller, an energy and utilities strategist at Morningstar Securities Research.

“We think that the growth in electricity demand will end up at the lower end of most of the ranges out there,” he said.

Jabed Ahmed29 January 2025 10:29

‘DeepSeek has the potential to reshape the future of AI,’ expert says DeepSeek’s arrival could herald a new era of intense competition in the space that stretches beyond the US tech industry. The company’s biggest markets include the Middle East and Russia, where download rates of its app have been more than 300 per cent higher than its rivals, according to data shared with The Independent by app tracking firm Sensor Tower.

“DeepSeek has the potential to reshape the future of AI. This new rivalry will drive faster progress, spark fresh ideas, and deliver technologies that benefit people, businesses, and economies worldwide,” said Russ Shaw, founder of Global Tech Advocates.

“More importantly, it could fuel economic growth and unlock opportunities to develop talent and skills on a massive scale. It’s clear we’re at a pivotal moment that could shape innovation and economic progress for years to come.”

Jabed Ahmed29 January 2025 09:58

Tech stocks rebound in Asia as DeepSeek worries easeTechnology stocks led gains in Asia-Pacific markets on Wednesday, tracking advances on Wall Street as investor angst ebbed over the emergence of a low-cost Chinese AI model that some see rivalling U.S. dominance of the industry.

The dollar remained relatively firm as traders rotated back into the currency from safe-haven peers such as the Japanese yen, while also getting a boost from fresh tariff warnings from the Donald Trump administration.

Trading was thinned in Asia by Lunar New Year holidays that shuttered exchanges in mainland China and Hong Kong, as well as Taiwan, Singapore and South Korea.

Japan’s Nikkei share average .N225 advanced 1% on the day, snapping three straight days of declines.

Australia’s stock benchmark .AXJO added 0.6%, with a subindex of tech names .AXIJ climbing 1.8%.

Futures for the U.S. S&P 500 EScv1 rose 0.2%, following a 0.9% rise for the cash index overnight. Nasdaq futures NQc1 pointed 0.4% higher following a 2% rally.

The tech-heavy Nasdaq had tumbled more than 3% in the previous session, after the spiking popularity of Chinese startup DeepSeek’s app called into question sky-high valuations for U.S. chipmaker Nvidia NVDA.O and others at the forefront of the AI revolution.

“There appeared to be a level of relief in the rally, mostly because of a forming consensus that while ostensibly impressive, DeepSeek will either lack the scalability to truly disrupt the AI space and, if anything, the company’s low-cost model will actually increase demand for GPUs,” said Kyle Rodda, senior financial market analyst at Capital.com.

Attention now turns to mega-cap tech company earnings coming up on Wall Street later in the day from Facebook owner Meta Platforms META.O, Microsoft MSFT.O and Tesla TSLA.O.

Jabed Ahmed29 January 2025 09:21

DeepSeek goes quiet for lunar new year despite worldwide buzzJabed Ahmed29 January 2025 07:59

Chinese firm Alibaba releases AI update it claims surpasses DeepSeek Chinese tech company Alibaba released a new version of its Qwen 2.5 artificial intelligence model on Wednesday that it claimed surpassed DeepSeek-V3.

The unusual timing of the Qwen 2.5-Max’s release, on the first day of the Lunar New Year when most Chinese people are off work and with their families, points to the pressure Chinese AI startup DeepSeek‘s meteoric rise in the past three weeks has placed on not just overseas rivals, but also its domestic competition.

“Qwen 2.5-Max outperforms … almost across the board GPT-4o, DeepSeek-V3 and Llama-3.1-405B,” Alibaba’s cloud unit said in an announcement posted on its official WeChat account, referring to OpenAI and Meta’s most advanced open-source AI models.

The 10 January release of DeepSeek‘s AI assistant, powered by the DeepSeek-V3 model, as well as the later release of its R1 model, has shocked Silicon Valley and caused tech shares to plunge, with the Chinese startup’s purportedly low development and usage costs prompting investors to question huge spending plans by leading AI firms in the United States.

But DeepSeek‘s success has also led to a scramble among its domestic competitors to upgrade their own AI models.

Two days after the release of DeepSeek-R1, TikTok owner ByteDance released an update to its flagship AI model, which it claimed outperformed Microsoft-backed OpenAI’s o1 in AIME, a benchmark test that measures how well AI models understand and respond to complex instructions.

Angus Thompson29 January 2025 07:30