Nscale

The demand for computing power is so fierce that OpenAI’s departure barely mattered to Nscale, which has quickly become one of Europe’s hottest startups, valued at $14.6 billion. Known as a “neocloud,” so called because it is built primarily for the needs of new AI models, it is part of an emergent category of tech companies that have risen up to satiate the market’s ravenous hunger for AI computing power. Nscale is barely two years old, but business is booming, with five data centers in various stages of construction in the U.S., U.K., and Norway. Investors are lining up in advance of a possible IPO later this year. In March Nscale raised $2 billion, the largest round of its kind in European history. Josh Payne, the company’s 32-year-old CEO, tells TIME he wants to turn Nscale into a “$1 trillion hyperscaler”—a new cloud company that can compete with the likes of Amazon, Microsoft, and Google.

• Before launching Nscale in 2024 after moving from Australia to London, [Payne] worked in coal mines, where long shifts left stretches of downtime he filled by reading entrepreneurship books such as The 4-Hour Work Week, according to The Times. Eventually he launched a recruitment platform that connected construction workers with jobs across Australia. That business exposed him to large infrastructure projects and energy markets, which later pulled him into renewable energy investing and cryptocurrency mining, two industries that rely heavily on massive amounts of electricity.
  • Rather than build all of that infrastructure themselves, some technology companies have begun turning to a new generation of providers sometimes referred to as “neoclouds,” companies that construct and operate AI data centers and sell computing power to developers. Nscale is part of that emerging group. Companies like CoreWeave in New Jersey, Nebius in Amsterdam, and IREN in Sydney are pursuing similar strategies, building the physical backbone that AI developers depend on while assuming much of the financial risk of constructing the facilities, according to The Times. The model allows large technology companies to secure computing power quickly, while delaying the decision to invest billions into their own infrastructure.

Mar09'2026: But a Guardian investigation has shown the money isn’t necessarily real, the datacentres may not be new, the jobs are unaccounted for – and the supercomputer site 12 miles north of London is still a scaffolding yard.