Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
Supercomputing EU Hardware

Europe Hopes To Join Competitive AI Race With Supercomputer Jupiter (france24.com) 41

Europe on Friday inaugurated Jupiter, its first exascale supercomputer and the most powerful AI machine on the continent. Built in Germany with 24,000 Nvidia chips, the 500-million-euro system aims to close the AI gap with the US and China while also advancing climate modeling, neuroscience, and renewable energy research. France 24 reports: Based at Juelich Supercomputing Centre in western Germany, it is Europe's first "exascale" supercomputer -- meaning it will be able to perform at least one quintillion (or one billion billion) calculations per second. The United States already has three such computers, all operated by the Department of Energy. Jupiter is housed in a centre covering some 3,600 meters (38,000 square feet) -- about half the size of a football pitch -- containing racks of processors, and packed with about 24,000 Nvidia chips, which are favored by the AI industry.

Half the 500 million euros ($580 million) to develop and run the system over the next few years comes from the European Union and the rest from Germany. Its vast computing power can be accessed by researchers across numerous fields as well as companies for purposes such as training AI models. "Jupiter is a leap forward in the performance of computing in Europe," Thomas Lippert, head of the Juelich centre, told AFP, adding that it was 20 times more powerful than any other computer in Germany. [...]

Yes, Jupiter will require on average around 11 megawatts of power, according to estimates -- equivalent to the energy used to power thousands of homes or a small industrial plant. But its operators insist that Jupiter is the most energy-efficient among the fastest computer systems in the world. It uses the latest, most energy-efficient hardware, has water-cooling systems and the waste heat that it generates will be used to heat nearby buildings, according to the Juelich centre.

Europe Hopes To Join Competitive AI Race With Supercomputer Jupiter

Comments Filter:
  • "Europe" who? (Score:5, Insightful)

    by Misagon ( 1135 ) on Saturday September 06, 2025 @04:04AM (#65642852)

    Many Europeans are hoping that this bubble will burst already ...

    • The internet bubble burst in the late 1990s. We still have an internet and the internet is if anything more useful and ubiquitous than most people predicted in the late 1990s. There likely is an AI bubble, but if/when it bursts we'll likely see the same thing. A technology can have a lot of potential and staying power and also be part of a bubble.
      • But are NVidia and OpenAI going to be entrenched winners like google, or more like Blackberry or Sun Microsystems? Those companies saw the future well enough and fit right into it, but it was too big for them to swallow up and bigger fish came along.
  • by polyp2000 ( 444682 ) on Saturday September 06, 2025 @05:40AM (#65642926) Homepage Journal

    The energy usage of AI on banks of GPU's in datacentres is a major obstacle in efforts to combat climate change.
    Competing over who has the most powerful AI is only going to make this situation worse.

    You might worry about losing your job to an AI , but the truth is that once again we have more confirmation that we are on the wrong trajectory. With a 5% chance of changing course to one that does not end in socio economic collase, climate breakdown and eventually mass extinction.

    There is simply no chance when these are the whims and attitudes of those who are supposed to be steering this
    ship to safety.

    My plan is not to participate in this system i oppose wherever possible. I suggest others do the same - if only to buy yourself a little more time and build reslience for the unavoidable hardships that are just around the corner.

    • All anyone or a press release can talk about is AI nowadays.

      However, this machine uses Grace-Hopper superchips and a distributed-memory network. It is more, or maybe equally, suited for computational tasks involving predictive science. As an example, predictive science includes solving classic and boring partial-differential-equations (PDEs), i.e., providing foundational training data for scientific AI. Using both predictive science and AI to solve a problem has a lot of promise and this machine is we
  • TFS tells us this 3600 meters is about half the size of a football pitch. I also cover 3600 meters, if you use a line that is thin and squiggly enough, although I can't pitch a football that far. I don't any living human can.

    • They mean area not length, but forgot the square. It's 3600 m^2, 60 m × 60 m, or about 200 ft × 200 ft.

  • The Yapese economy operated on a system that almost prefigured the architecture of a modern AI chatbot: massive stone disks called rai stones, painstakingly quarried on distant islands and transported home via monumental canoe journeys, served as Yap's store of value.

    Yet these stones’ worth was not inherent, but algorithmically assigned by a kind of proto-blockchain of collective memory, with their ownership and transaction records cryptographically secured in the oral consensus of the island
  • Must be really popular with the poor unemployed serfs struggling to pay for expensive electricity that used to be generated by cheap clean natural gas. It must feel like " kick 'em while they're down ".
  • so now more than ever, "boys go to Jupiter to get more stupider."

Never underestimate the bandwidth of a station wagon full of tapes. -- Dr. Warren Jackson, Director, UTCS

Working...