AI Compute Costs Soar, Sparking Investor Jitters and Infrastructure Scrutiny

The burgeoning demand for artificial intelligence is driving unprecedented spending on computing infrastructure, with reports indicating AI companies like Anthropic could face staggering costs upwards of $750 billion. This surge in expenditure is creating significant challenges for AI developers, as cloud providers increasingly view them as competitors, complicating crucial partnerships. The strain on infrastructure is also evident in the energy sector, with the U.S. restarting nuclear reactors to meet data center power needs, backed by major tech players like Microsoft and Google.

These escalating costs and the complex web of infrastructure dependencies are beginning to unnerve investors. Concerns over AI valuations are mounting, leading to a significant increase in trading of credit default swaps linked to major tech groups, as investors seek protection against a potential market downturn. The recent market debut of South Korean data center builder Elice, which saw its shares tumble 32%, further highlights these growing anxieties. Even established data center hopefuls are facing scrutiny, with Microsoft reportedly reassessing a pre-IPO deal with Nscale, signaling a potential reality check for the fast-growing AI sector.