Germany has deployed what it says is Europe’s most powerful AI machine to date, as Chancellor Friedrich Merz on Friday oversaw the activation of a new Nvidia-powered supercomputer, called Jupiter, at the Juelich research center in western Germany, according to Reuters.
The system, built with support from French IT firm Atos and German company ParTec, now ranks as the fourth-fastest supercomputer in the world.
It’s also the first in Europe to reach Exascale class, capable of performing one quintillion operations per second, or about the combined processing strength of 10 million laptops.
Friedrich called the machine a “historic European pioneering project” and said it is Europe’s way of responding to the United States and China, both of which lead in the push toward an AI-driven economy. “We in Germany and in Europe have all the opportunities to catch up and then to hold our own,” he said at the launch.
Jupiter is designed for use in scientific fields like biotechnology and climate research, areas where supercomputing is essential for running simulations, building models, and processing massive datasets. But the installation is also seen as a political signal.
European institutions are trying to reduce dependence on foreign-controlled digital infrastructure and chips, especially from US tech giants and Chinese manufacturers.
Officials say Jupiter should stay accessible to firms and researchers
Ralf Wintergerst, who leads Germany’s digital business association Bitkom, said the new machine will push Germany to the front of the global high-performance computing field and help expand the country’s AI capabilities.
He urged that Jupiter be made easily accessible for use. “Access to it should be made as unbureaucratic as possible for start-ups and established companies,” Ralf said on Friday. The European Union has lagged behind in developing the kind of hardware needed to support large-scale AI development.
While engineers in Silicon Valley run their models on dense racks of Nvidia GPUs, and Chinese labs scale up through state-funded manufacturing, Europe has mostly stayed on the sidelines.
Nvidia pushes back against proposed US export law limiting chip sales
On the same day Jupiter went online, Nvidia issued a public warning about a proposed US law that could block machines like Jupiter from ever being upgraded again.
The law, titled the GAIN AI Act (short for Guaranteeing Access and Innovation for National Artificial Intelligence) was introduced under the National Defense Authorization Act and would require AI chipmakers to prioritize US domestic orders over international shipments.
A spokesperson for Nvidia said the bill would do more harm than good.
“We never deprive American customers in order to serve the rest of the world,” the company said. “In trying to solve a problem that does not exist, the proposed bill would restrict competition worldwide in any industry that uses mainstream computing chips.”
The bill proposes strict licensing rules. Any chip with a performance score above 4,800 would require an export license, and the US Department of Commerce would have the power to deny licenses altogether.
The draft legislation says that exports of advanced chips should be blocked if US buyers are still waiting for supply. It follows the AI Diffusion Rule implemented under former President Joe Biden, which placed limits on how much processing power US companies could export to other countries.
The goal was to prioritize US access to critical AI infrastructure and to prevent China from gaining the hardware needed to strengthen its military through artificial intelligence.
Despite these restrictions, President Donald Trump struck a deal with Nvidia in August, agreeing to let the company resume exports of banned AI chips to China, in exchange for giving the government a percentage of Nvidia’s sales tied to those exports.
The agreement raised questions about whether economic interest is now being weighed more heavily than security concerns in the regulation of AI technology.
While the US tightens control, Germany is trying to open things up… at least internally. Europe’s goal is to build infrastructure that can handle next-generation AI development without needing constant permission from Washington.
The problem is that most of the world’s best chips are still made in the US, and Nvidia remains the core supplier for nearly every major AI system globally.
If you're reading this, you’re already ahead. Stay there with our newsletter .
