On the first full day of President Trump’s State Visit to Britain, Nvidia, Microsoft, OpenAI and Google pledged to invest billions of pounds in investment in U.K.AI.
It comes after weeks of nervousness about the wider sector, and slumps in share prices for many AI-oriented tech companies.When OpenAI CEO Sam Altman admits the industry is over-hyped and Alibaba co-founder Joe Tsai sees the ‘beginning to see some kind of bubble,’ then it’s probably justified.
In the Swedish capital, however, investors are taking a rather typically lagom approach (neither too little nor too much).After Silicon Valley, Stockholm generates more startups than any other city – think Klarna.Here, AI-tech startups have been springing up faster than mushrooms in the forest.
Yet shades of similarity remain with the dot.com bubble.Are we here again?
“It could be a bubble,” admits Tove Larsson, General Partner at Norrsken VC, the leading Swedish impact investor.“This is one of the most intense capital flows we have seen.
There’s a clear risk that 80% of these companies will be out of business eventually.”
“Most of the companies we’ve seen in the AI space focus on productivity gains, and the rest on robots.The real wins will come from focusing on the problems that really matter and building AI solutions to them.”
Tools such as Chat GBT may have been widely embraced, but they primarily enhance individual productivity, rather than profit.
A recent study by MIT showed that despite $30-40bn of enterprise investment in Generative AI, 95% of organisations have zero return.It too concludes that tools such as Chat GBT may be popular but they primarily enhance individual productivity, rather than profit.
“A shake-out in AI feels inevitable – the question isn’t if, but when,” agrees Ali Sarrafi, a former Spotify exec and co-founder of the data science and analytics company Combient Mix, bought by Silo AI and later AMD.
His latest AI-related venture, Kovant, is still in stealth mode.
With drone warfare in Ukraine and Poland, the defence applications of AI software are starkly illustrated.AI is already critical to all kinds of data applications, with governments heavily invested in AI applications in national models and infrastructure.That’s unlikely to change.
“Unlike the dot.com bust, this isn’t the end of AI but a natural maturing cycle.It will be the moment the sector separates durable value from short-term hype,” adds Sarrafi.
How The AI Shake Out Will Look
“I believe we’ll see a clearer separation between those that stand to win and lose in this AI era,” says Emmett King, co-founder of Swedish investors J12 Ventures.”Some companies won’t live up to the hype, but that’s more a sign of market maturity rather than a meltdown.”
In his view, the key distinction from the dot-com era is that AI is already delivering substantial productivity gains in certain industries.
King believes these will keep AI on a positive trajectory, even if valuations correct and some players disappear.
“Strong categories are productivity and workflow automation, customer support and sales, decision support in healthcare, finance, and manufacturing, and vertical applications in areas like law or HR,” he argues.
While AI will attract an outsized share of investment, capital will also flow to adjacent areas such as semiconductors, energy infrastructure, and data management.Non-AI companies won’t vanish from investor focus, but the gravitational pull of AI will dominate.
AI Boom Closer to 1995
Further south in Malmö, Dominic Davies, the British-born co-founder and CEO of Lightbringer believes the shake out is necessary.A clear example of a vertical application, Lightbringer’s AI-based platform is a helps startups and innovators to apply and manage patent applications faster and more easily than through conventional legal channels.
Davies thinks where we are now is more like 1995 than 2000.
“The underlying technology is revolutionary and has enormous potential, but the market is looking for proof of viable businesses.”
“Many of the winners will be the providers of a foundational model on which other companies base their products.Companies like OpenAI, Anthropic, and Google are a long-term infrastructure play, building the core platforms of the future.”
He also expects applied AI companies – firms scaling up by automating high-value professional services to be winners.
The losers will be those who attach AI to their names but fail to solve any real problems.
AI for impact
Norrsken has nailed its colours firmly to the Big Challenge mast.It has pledged to invest €300mn in AI tech startups tackling problems such as climate change, healthcare and food scarcity.
“The first trillion dollar company will be the one that leverages AI to solve some of these challenges,” argues Larsson.“It’s also bringing about a seismic shift in technology companies.AI is really changing the way companies are built.”
Because AI tech can achieve much more with less, she expects to see companies driving huge values with far fewer resources and becoming profitable much faster.
She likes Submer, which cools data centres by placing them in a bathtub filled with nonconductive fluids taking up less space and costing less than conventional data centres.“Some companies will be able to cool their data centre right in the heart of the city,” she says.
Larsson also favours Biorce, which speeds up the absorption of massive amounts of data from clinical trials to cut the time and cost in trials.
“The tools can really help pull together all the relevant data from other studies, and help the person writing the protocol for the clinical trial to be faster and more accurate, avoiding as many errors,” adds Larsson.
“By using huge data sets, they’ll speed up access to new medicines and help bring costs down.”
In the centre of Stockholm, a massive infrastructure project has been underway for several years at Slussen, the city’s old lock.Slowly the area is transforming while the city’s transport and hydrology infrastructure improves.It’s a long-term investment in a country where long-term investment is at the heart of economic growth.Sweden came late to the industrial revolution, but it is going to be at the forefront of the AI one..
