This analysis was first published in SvD Näringsliv, in Swedish, on October 23rd, 2026. This piece was translated from Swedish by Claude. Some phrasing may differ from a human translation.
Volvo Cars’ dismal numbers are nothing but a nightmare for the company. They arrive with the worst possible timing – and now the winds are blowing from several directions in the global economy.
What does car sales have to do with AI? That’s the question you may ask after Volvo Cars plunged on the stock market. More than a quarter of its market value was wiped out in early trading on Friday morning.
The quarterly figures are so bad that the fall is justified on their own merits – revenue dropped 23 percent, and the number of cars sold in China plunged a full 43 percent compared with the same period a year earlier. But the timing is also extraordinarily bad for the carmaker. The market is shaking because of an awakening that has nothing to do with cars. It is the fear of AI that has turned the stock market deep red.
The signals that something is wrong have been there for a long time, but few wanted to acknowledge them until this past week. The market has trembled over AI before, but now the global economy is caught in crosswinds from several directions. To come out and underdeliver this badly at such a moment is simply a nightmare.
Three major factors make the market uneasy.
The first is the oil price, where the conflict in the Strait of Hormuz keeps frightening the market. Despite Donald Trump’s repeated attempts to say the war with Iran is nearly won, new threats to global oil trade keep coming. Oil passed $114 a barrel, compared with around $65 a year ago.
Nothing about this signals that the situation is under control. Oil has nothing to do with AI in itself, but it affects sentiment across the global economy. Not even the AI sector is immune.
The second factor is interest rates, and what it costs to finance the build-out of data centers around the world. You might have expected the world’s most highly valued companies – and in many cases among the most profitable – to be able to fund it themselves. But four of the five so-called hyperscalers – the name for the big cloud giants Meta, Alphabet (Google), Amazon, Oracle and Microsoft – now have negative free cash flow. They are already investing more than they bring in. So instead they have chosen another, riskier path.
The solution has been to borrow billions to keep funding the build-out. Hyperscalers borrow on their own, and also buy capacity from other companies whose business is exactly that – such as CoreWeave and Nebius.
The yield on one-year U.S. Treasuries is now around 4.5 percent. That is generally considered a risk-free return. If you are going to lend money for data centers, you naturally want a higher return than that. It has made borrowing expensive for all these players.
The third and last point concerns who is actually going to use all this expanded capacity. Here the market has sent mixed signals. On one hand, heavyweights such as Nvidia CEO Jensen Huang have said demand is enormous. On the other, many worry that it is artificial demand that doesn’t reflect how AI will be used in the long run.
The financial giant Apollo put its finger on this during the week. Its chief economist reviewed data from the payments platform Ramp. It showed that among the single percent of companies spending the most money on AI, 80 percent are venture-backed.
That points to a major concentration risk, and could make demand sensitive if venture capital stops flowing in. Right now the game is to grab market share quickly, at any price. But historically that phase passes, and the focus on profitable growth returns. Then venture capital takes a different form, and the dark-red numbers stop being accepted.
At AI giant Anthropic, which is heading for the stock market, a similar pattern can be seen. A leaked prospectus obtained by the Wall Street Journal states that a fifth of its revenue came from just two customers during the first six months of the year. The customers’ names aren’t listed, but they could be companies that compete with Anthropic in various ways – Cursor (owned by SpaceX) and Microsoft, via Copilot. Nvidia, too, got 44 percent of its revenue from just three customers in the first half of its fiscal year.
This extreme concentration in a handful of companies means the actual demand for all these AI services can be questioned. And if you do – is it reasonable to borrow enormous sums to build new data centers? When financing has at the same time become more expensive?
When the U.S. Nasdaq 100 fell on Thursday, 90 percent of the loss in value came from AI-related stocks – chip companies and the AI giants Meta, Microsoft, Amazon and Alphabet. That is how the uncertainty described above is expressing itself right now.
So what does car sales have to do with AI? In a market that has for a long time been held up essentially by the AI sector alone, the answer is: a great deal. They are not alone. The industrial group Atlas Copco, which makes technology used in the AI industry, was also dragged down on Thursday.
It isn’t Volvo Cars’ fault that the market is starting to doubt AI. But together with such weak car sales – in an industry in crisis even without AI worries in the world – it has become their problem to solve.
