Brussels Bets on Mega-Computing While Climate Fears Grow
China’s LineShine computer in Shenzhen has displaced the top-ranked US computer El Capitan to become the world’s most powerful supercomputer in the Top500 rankings released on Tuesday, a reminder that the race for technological supremacy is being measured in ever larger machines, ever higher electricity use, and ever more concentrated control. It is the first time since 2017 that a Chinese computer has topped the list, which is sometimes viewed as a measure of a nation’s technological prowess.
LineShine’s debut on the list put it ahead of El Capitan, which sits at the US government’s Lawrence Livermore National Laboratory in California and now ranks second. Two other US supercomputers at national laboratories in Tennessee and Illinois follow behind it, while Germany’s Jupiter dropped to fifth place. The five are the only publicly verified exascale computers in the world.
The Hardware Arms Race
Scientists involved in the Top500 project said LineShine at China’s National Supercomputing Center achieved 2.198 exaflops, meaning it can perform more than 2 quintillion calculations per second. Unlike other high-performance computers, LineShine runs entirely on conventional computer chips, or CPUs, instead of the graphics processors, or GPUs, commonly used for AI. It requires about 42.2 megawatts of electricity to operate, according to the list.
The ranking also places other states and institutions in the same contest. Italy, Switzerland and Japan each have machines in the top 10. The UK has 11 machines in the list of 500, with the University of Bristol’s Isambard-AI the highest ranked of that group at 11, down two places since the last ranking. Isambard-AI, fitted with 5,400 Nvidia “superchips”, sits inside a black metal cage topped with razor wire.
That image is hard to improve on: the future of AI, wrapped in steel and razor wire, guarded like a strategic asset rather than a public good.
Brussels’ €20bn Bet
Last year the EU revealed a €20bn plan to build sites equipped with vast supercomputers to develop the next generation of AI models, as Europe attempts to catch leaders in the US and China. The AI “gigafactories” will target “moonshot” innovations in areas such as healthcare, biotech, industry, robotics and scientific discovery.
The best-performing AI factories already use supercomputers equipped with up to 25,000 advanced AI processors, but a gigafactory would exceed 100,000 AI processors, the EU strategy document said. The scale is the point: more processors, more power, more water, more centralised infrastructure, more dependence on the institutions that can afford to build and control it.
An EU official said these power-hungry facilities, which can require huge amounts of water for cooling, should run “as much as possible” on a green energy supply, with plans for “recycling” water if it was used. Campaigners fear power-hungry datacentres could undermine Europe’s climate ambitions.
The language is familiar from the Brussels apparatus: green promises wrapped around infrastructure that devours electricity and water, while the costs are pushed outward and downward. The EU’s answer to technological competition is not restraint, but bigger machines and a larger industrial footprint.
Who Gets the Future
The Top500 list is being used as a scoreboard for national technological prowess, with the United States and China at the top and Europe trying to catch up through its €20bn plan. The result is a familiar hierarchy: state laboratories, national supercomputing centres, and corporate hardware suppliers at the top; everyone else left to absorb the environmental and social consequences.
The article does not describe any public debate over whether the race itself should be happening, only how fast each bloc can build the next machine. That is the logic of the system on display: competition first, consequences later, and climate ambitions treated as something to be “undermined” rather than a limit to be respected.
What emerges from the numbers is not just a contest over computing power, but a contest over who gets to command the infrastructure that increasingly shapes work, science, and industry. The EU’s gigafactory plan, with its 100,000-processor ambition, sits squarely inside that logic — a continental attempt to keep pace by scaling up the same energy-hungry model.
For now, the world’s fastest machine is in Shenzhen, the US still holds two of the next three spots, Germany has slipped to fifth, and Brussels is promising a future built on even larger data centres and even more electricity. The climate, as usual, is expected to make room.