Data centres: the slowdown in the AI race wipes 4,200 billion off the stock market
Since June, I have lost almost half of the 9,400 billion earned over the previous twelve months by industrial, energy and mining companies involved in the construction of artificial intelligence infrastructure
The rush to build data centres to support the development of artificial intelligence is taking its toll on the financial markets. Since June, manufacturing, energy and mining companies most exposed to the development of new digital infrastructure have lost a total of 4,200 billion dollars in market capitalisation. This correction has wiped out almost half of the 9,400 billion gained by these same companies over the previous twelve months, when investors had bet on the benefits of growth in artificial intelligence.
An analysis by the Financial Times, based on data from FactSet, sheds light on the scale of the phenomenon. The companies hardest hit have been precisely those that provide the infrastructure necessary for the operation of large data centres: from power generators and cooling systems to the fibre-optic cables connecting the servers used to process AI models.
One of the most notable examples is that of Caterpillar. The American group, one of the leading suppliers of gas generators for data centres, has seen its market capitalisation fall by $214 billion, or 44 per cent, from the all-time high reached by its shares on 30 June. Oklo, a nuclear energy company that signed a contract with Meta in January, has also lost 49 per cent of its value. Comfort Systems, a manufacturer of air-conditioning systems, and Corning, which specialises in optical fibre, have both seen their market capitalisation fall by around half.
This shift in direction reflects a broader re-evaluation by investors of the sector’s prospects. The main factors at play are doubts about the sustainability of investments by major technology firms – which are increasingly financed through debt – and growing political opposition to the construction of new data centres in the United States. The fear is that enthusiasm for artificial intelligence infrastructure has fuelled excessive valuations, exposing companies in the sector to the risk of a sharp correction.
The industrial sector has become the second-worst-performing sector in the S&P 500 index since last summer, falling by 9 per cent between 30 June and 13 November. Only the technology sector fared worse, falling by 11 per cent. By contrast, over the same period, the energy sector gained 22 per cent, healthcare 19 per cent and consumer staples 11 per cent.
