As ‘black swan’ events go it takes some beating.
Barely a week ago few people had heard of a Chinese technology start-up called DeepSeek.
Fewer still knew anything about its reclusive, low-profile founder Liang Wenfeng.
All that changed today.
Shares in US chip giant Nvidia crashed 13 per cent – wiping $465billion (£370billion) off its stock market valuation – on news that DeepSeek’s artificial intelligence tool had become the most downloaded iPhone free app, overtaking OpenAI’s ChatGPT chatbot.
It was the biggest one-day destruction of value in stock market history.
Nivida, whose stock had soared on the back of booming demand for its most advanced semiconductors that power AI and underpin all digital computing, was also toppled from its perch as the world’s biggest quoted company.
Shares in other American tech titans fell in sympathy as fears grew that the fortunes they have spent on cutting-edge AI models that recognise patterns and make decisions without human intervention may have been wasted.
DeepSeek founder Liang Wenfeng (right) speaks at a Chinese government symposium last Monday
DeepSeek said it had built an AI training model using Nvidia’s cheaper chips for under $6 million (£4.8million) – a fraction of the cost of what US firms were spending.
‘Big tech firms like Microsoft, Google, Amazon and Facebook-owner Meta have deployed hundreds of billions to purchase GPUs (graphic processing units, or microchips) from Nvidia and ensure chip supply to satisfy the insatiable demand for AI,’ said Javier Correonero, equity analyst at Morningstar.
‘DeepSeek is challenging that hypothesis.’
High-flying Nvidia wasn’t meant to fall to earth like this.
Experts always feared the chip designer’s remarkable rise was most at risk from China.
They pointed to Beijing’s long-held claim to Taiwan, the neighbouring island which makes more than 60 per cent of the world’s microchips – silicon discs with billions of transistors carved into them.
Crucially, Taiwan also makes over 90 per cent of the most advanced chips used in satellites, stealth jets and, increasingly, AI.
The US government vowed to defend Taiwan with force in the event of an invasion – possibly within three years, it reckoned.
But it turned out Nvidia and other US tech giants were laid low without a shot being fired.
Instead, they were hit by a much more imminent threat that came completely left-field – from a billionaire hedge fund manager who started a tiny AI lab in eastern China as a hobby just a few years ago.
DeepSeek’s Liang made his fortune betting on financial markets.
The son of a primary school teacher, Liang, 39, is an engineering graduate from Zhejiang University who founded a successful hedge fund called High-Flyer.
It gained a reputation for its innovative use of AI trading strategies which use computer algorithms to predict market trends and make data-driven investment decisions.
In 2021, Liang started buying thousands of Nvidia’s graphic processing units (GPUs) for what was at the time an AI sideline to his day job as hedge fund manager.
‘When we first met him, he was this very nerdy guy with a terrible hairstyle talking about building a 10,000-chip cluster to train his own models,’ one of Liang’s business partners told the Financial Times.
‘We didn’t take him seriously,’ the person added. ‘He couldn’t articulate his vision other than saying: “I want to build this, and it will be a game change”.’
In 2023, Liang launched DeepSeek, with a plan to develop human-level AI by focusing on research and engineering, rather than commercial return – the company is not listed on any stock market so you can’t invest in it.
In a rare interview, Liang said he was fascinated by the potential of artificial general intelligence – a type of AI that gives a machine the power to understand or learn any intellectual task that a real person can by mimicking the cognitive abilities of the human brain.
Companies such as Google, run by Sundar Pichai, will now have to face down a brand new competitor
‘The essence of human intelligence might be language,’ he said.
‘Human thought could be a linguistic process. You think you’re thinking, but you might actually be weaving language in your mind. This implies that human-like artificial intelligence could emerge from large language models.’
Another motive of Liang’s is to challenge US hegemony in AI.
American companies like Nivida have led the way in AI chip design, while OpenAI and Google’s DeepMind have pioneered developments in reasoning models that try to make machines match – or even surpass – human cognitive capabilities.
‘Nvidia’s leadership isn’t just the result of one company’s efforts.’ Liang said. ‘It’s the collective achievement of the entire Western tech community and industry. They can see the next generation of technological trends and have a roadmap.’
Liang believes China must build its own AI ecosystem in response.
‘We often say there’s a one- or two-year gap between China and the US but the real gap is between originality and imitation. If this doesn’t change, China will always be a follower,’ he said. ‘That’s why China must have people at the forefront of technology.’
When Washington recently banned Nvidia from exporting its most powerful chips to China, Liang’s team were ready to pounce.
DeepSeek had already worked out how to find innovative ways to maximise the computing power of the Nvidia chips China already had – even though they were not state-of-the-art.
The company released its R1 model last week, explaining in detail how to build a large language model on a shoestring that can automatically learn and improve itself without human supervision.
It was the news that sent shockwaves across the tech world – and led to one of the biggest sell-offs in stock market history.
It’s also made Liang a national folk-hero at home.
He was the only AI leader picked to attend a meeting of entrepreneurs with the country’s second-most powerful leader, Li Qiang.
They were told to ‘concentrate efforts to break through key core technologies’.
Liang has certainly done that – as investors around the world are learning to their cost.
The rout in technology stocks will certainly be painful for many British investors.
The so-called Magnificent Seven – Apple, Google owner Alphabet, Amazon, Facebook owner Meta, Microsoft, Tesla and Nvidia – have proved hugely popular investments on this side of the Atlantic in recent years.
And they have been highly lucrative, with Nvidia up almost ten-fold since the start of 2023 before yesterday’s record sell-off.
Many savers who do not own these technology stocks outright are exposed through their pensions or other investments such as London-listed trusts.
For example, the Scottish Mortgage Investment Trust, a member of the FTSE 100, counts Amazon, Meta and Nvidia among its top five holdings. Its shares fell four per cent.
So what does it mean for investors?
Jochen Stanzl, chief market analyst at CMC Markets, said the rush to invest in AI ‘drove stocks like Nvidia to astronomical valuation levels’.
But he added: ‘Now a harsh reality is setting in. Never before has a company in the S&P 500 lost so much value in a single day.
‘The US can no longer assume that it will retain its dominance in the AI market. China’s low-cost AI could undermine the margins of highly valued AI companies.
‘Is this the beginning of the bursting of a massive speculative bubble?’
Ben Barringer, a technology analyst at Quilter Cheviot, said that while the sell-off may be ‘a knee-jerk reaction’, it also ‘provides a wake-up call’ to the industry in terms of how much it really needs to spend to achieve results.
‘While current leaders like Nvidia have a strong foothold, it is a reminder that AI dominance cannot be taken for granted,” added Charu Chanana, chief investment strategist at Saxo Markets.
‘The emergence of China’s DeepSeek indicates that competition is intensifying, and although it may not pose a significant threat now, future competitors will evolve faster and challenge the established companies more quickly.’
Other experts were more sanguine.
OpenAI chief executive Sam Altman. DeepSeek’s artificial intelligence tool had become the most downloaded iPhone free app, overtaking OpenAI’s ChatGPT
Susannah Streeter, head of money and markets at Hargreaves Lansdown, said: ‘While DeepSeek has clearly ruffled feathers, its spike in popularity isn’t likely to drastically damage chipmakers like Nvidia alone.
‘Big corporates are still likely to pay top dollar to access the best chips on the market, but when it comes to widespread adoption globally, it’s not going to be an open goal and the big players will still have to keep on their toes to stay dominant.’
Russ Mould, investment director at AJ Bell, said the danger with technology stocks was that another next generation of companies will emerge ‘to make the latest new order look like the next old order, and disrupt the disruptors’.
‘Obsolescence risk has always been technology stocks’ biggest threat, and it always will be,’ he added. ‘The AI super-race is seeing new challengers emerge and not everyone is going to win.’
Analysts said investors should spread the risk, particularly if they are heavily exposed to US technology stocks.
Kenneth Lamont, principal at Morningstar, said: ‘While DeepSeek itself isn’t publicly investable, one way to guard against such risks may be diversifying away from the so-called Magnificent Seven tech giants.
‘This serves as a fresh reminder for thematic investors: mega-trends rarely unfold as expected, and today’s dominant players might not be tomorrow’s winners.’
Nigel Green, chief executive of global financial advisory firm deVere Group, described the launch of the DeepSeek chatbot as ‘a historic pivot in the balance of technological power’.
He added: ‘US Big Tech, long considered untouchable, is facing credible and intensifying competition from China. For investors, this is both a warning and an opportunity – it’s time to rethink traditional tech allocations and seek out new areas of growth.
‘Traditional tech giants are no longer the guaranteed winners. The focus must now shift to sectors and regions that are driving the next wave of innovation. This includes not only AI but also the critical infrastructure needed to support and secure it.’
Some links in this article may be affiliate links. If you click on them we may earn a small commission. That helps us fund This Is Money, and keep it free to use. We do not write articles to promote products. We do not allow any commercial relationship to affect our editorial independence.
