Switch the Editor’s Digest free free
Roula Khalaf, editor of the FT, selects her favorite stories in this weekly newsletter.
It is difficult to talk about the economic history of the 21st century without discussing the “China shock”. This term often describes China’s admission to the world market, a change that gave the rich countries a wealth of cheaper goods, but put down entire industries and labor.
Deepseek could deliver a sequel. A little -known Chinese hedge fund has thrown a grenade into the world of artificial intelligence with a large voice model that is actually equal to the market leader Sam Altmans Openaai, to a fraction of the costs. And while Openai treats the functionality of its models as proprietary, the R1 from Deepseek carries its technical interior to the outside world, which makes it attractive for developers to use it and build it on it.
In the AI age it goes faster; Terrifyingly. Five of the largest technology shares focused on AI-the chip manufacturer Nvidia and the so-called Hyperscaler Alphabet, Amazon, Microsoft and Meta Platforms-lost almost $ 750 billion in market value before the US markets opened on Monday. For Nvidia, it could be particularly dark if it turns out that Deepseek has won without the use of his most shiny chips.
Investors in technology companies, including Europeans such as the chip manufacturer ASML, and energy companies, of which investors hoped for an upswing by firing data centers, are wondering whether their investments will be in smoke. According to Visible Alpha estimates, the hyperscalers should invest almost $ 300 billion in investments this year. Analysts assume that Meta and Microsoft will report investments for 2024 of a total of $ 94 billion on Wednesday.
In truth, the game is not over yet. The actual potential of Deepseek is still unclear and the “artificial general intelligence” still has to be achieved, the human -like state, the meta and openai strive. But the rules could have changed. At least Deepseek could win some of the US giants. In the worst case, the basic conviction questioned that more hardware is the key to better AI. This principle has underpinned the market value of Silicon-Valley companies that invest with full strength.
What is bad for the hyperscaler could still be a stroke of luck for everyone else. For most business users, the absolutely best model is less important than one that is reliable and good enough. Not every driver needs a Ferrari. Progress in thinking like R1 could be a big step for “agents” who deal with customers and perform tasks at the workplace. If these are available more cheaply, corporate reduction is likely to increase.
In this sense, this second China shock could resemble the first. It could not only have destruction, but also a restructuring – albeit a painful one for many. Researchers have estimated that for every workplace lost for everyone lost by the China shock The purchasing power of US budgets rose by more than $ 400,000. The race for AI pre-rules is on ice. The big raffle has started.