Germany's new supercomputer runs entirely on renewable power

Germany just switched on a machine that can do in seconds what would take a regular computer weeks. Otus, installed at Paderborn University, has 42,656 processor cores and enough storage to hold five petabytes of data. But what makes it genuinely significant isn't just the raw power — it's that the whole thing runs on renewable electricity, with waste heat piped into university buildings to keep them warm.

This matters because supercomputing has always been an energy hog. These machines need constant cooling, they draw enormous power, and they've typically been justified as necessary despite their carbon footprint. Otus breaks that equation. Lenovo and pro-com Datensysteme built it with indirect free cooling that works year-round, meaning the system operates efficiently without the energy spike you'd normally expect.

"High-performance computing is hugely relevant to the pressing challenges of our time," said Paderborn University President Matthias Bauer. What he means: researchers can now simulate experiments that are too expensive, dangerous, or physically impossible to run in a lab. A chemist can model molecular reactions at atomic scale. An engineer can optimize shipping routes to cut fuel consumption. A materials scientist can design more efficient solar cells. All without building prototypes or running costly trials.

Otus isn't just available to Paderborn's researchers. It's a national resource. Scientists across Germany submit proposals, which go through independent review. Once approved, their simulations queue up and run automatically as computing capacity opens up. The system runs 24 hours a day, 365 days a year — which means the machine is almost always working on someone's problem.

"Experiments that would be extremely expensive, dangerous or quite simply impossible can be simulated," Bauer explained. That's not abstract. It means faster drug discovery, better climate models, more efficient renewable energy systems. It means research that would otherwise stall or require years of funding can move forward in months.

The inauguration brought together leading academics to discuss what becomes possible with this kind of computing power. Talks ranged from machine learning in chemistry to quantum photonics to language models — the kind of research that's reshaping everything from materials science to AI.

What's worth noticing: this is how infrastructure gets built now. Not as a symbol of national pride (though that's part of it), but as a tool that works better when it's shared, sustainable, and open to competitive merit. Otus sits at the intersection of three things that usually don't align — raw computing power, environmental responsibility, and equitable access. The fact that all three are present in one machine suggests what's becoming possible when those priorities stop competing and start reinforcing each other.