AI for science is a global phenomenon, thanks to the potential for significant breakthroughs by incorporating AI technology and techniques into scientific discovery. In the United States and European Union, central governments have taken the lead in harnessing emerging AI techniques onto existing computational science and HPC systems. While the ultimate goals may be similar, the two regions have different approaches to getting there. However, there is one aspect that’s identical across both continents: the need for greater funding.
Technologists at the cutting edge of the AI for science movement recently discussed their respective governments’ approaches to pursuing breakthroughs during a session at the ISC 2026 conference in Hamburg, Germany.
Among the speakers at the session, titled “Towards a Common European AI Hub: Storing, Sharing, and Operating AI/ML Models, Applications, and Vector Data at Scale,” was Henrik Nortamo, who is a senior application specialist at the Finnish firm CSC-IT for Science and the technical lead for the EuroHPC Federation Platform, as well as Scott Atchley, who is the CTO at Oak Ridge National Laboratory.
Atchley discussed ORNL’s role in Genesis Mission, the Department of Energy effort to establish US leadership in the emerging field of AI for science and engineering. He mentioned the effort to stand up Lux, a new supercomputer to be installed at ORNL that will be dedicated to Genesis Mission workloads.
“We are trying a whole of government [approach], as well as with academia and industry partners, to unify all of our research capabilities, from experimental facilities to compute facilities, integrating with hyperscalers so that we can automate discovery,” Atchley said. “Today, it’s a very manual process, and the goal is to be able to streamline that, but also bring in AI to help us get to answers quicker.”
The DOE has delegated various national labs to different components of the Genesis Mission. For instance, ORNL is in charge of setting up the infrastructure as part of the American Science Cloud (AmSC), while Argonne National Lab is taking the lead with AI model development through the Transformational AI Models Consortium (ModCon). There are also the Lighthouse problems, or the 26 grand challenges, which are open to DOE teams as well as researchers in other institutions, universities, and companies around the world.
“There’s a strong push from headquarters to do this. They’ve delegated the work to the labs to make it happen,” Atchley continued. “It’s not something that we can do on [our] own. It’s not something the US government can do on its own. We absolutely have to have partnerships with academia and industry.”
The EuroHPC AI for Science initiative is similar in some respects to Genesis Mission, according to Nortamo. “I would say the current approach for the EU funded mission is also very top down in the sense that [we] feed the information, feed the users, feed the authorization,” Nortamo said. “But after that happened, that is fairly independent.”
As the technical lead for the EuroHPC Federation Platform, Nortamo is deeply involved in technical aspects of managing the group’s HPC resources, including authorization, access control, and storage. Today, the group allows fairly broad access to open public sources of compute. There is some cohesiveness to the various HPC resources, and users can federate access with a given context. Much of the work is done in the open, using open source software. However, the integration could be improved on to bolster AI research.
“I think the main challenge is we want to enable this kind of AI federation of artifacts across specifically this kind of infrastructure,” Nortamo said. “You do compute and then the project ends and then you apply for the next one. Now, if we’re thinking more that you start creating a lot of useful reusable and transient resources to use multiple times, this doesn’t really work well because then you have to migrate from project to project. There is no umbrella concept of yes, these are my things I have access to, and I can share those with others.”
Compared to the centralized efforts of the Genesis Mission, which is being driven by the DOE and headed up by the Undersecretary for Science Dario Gil, the European effort is less centralized. Nortamo expressed his desire for more centralization (and more funding).
“I think a lot of the AI efforts currently in Europe [are] well, typically European in the sense that you spread it out and you do something there, you do something there,” Nortamo said. “Speaking as the technical lead of the platform, if somebody were to create an AI hub, and I would put full technical capability to provide users identities, authorization to whatever interfaces, whatever granularity–that would be really cool because at that point, you could tie together the artifacts and where they’re actually used in like one cohesive way, but that’s something which people way above would need to fund again.”
While the US Government has allocated several hundred million dollars for Genesis Mission projects in fiscal year 2025 and 2026, more funding will be necessary to realize the dream. The White House is asking Congress to fund Genesis Mission, via the Office of Artificial Intelligence and Quantum (AIQ), to the tune of $1.2 billion for fiscal year 2027. But at this point, it’s anybody’s guess whether that funding will ever become available.
Henrik Nortamo (second from right) speaking at ISC 2026 in Hamburg, June 25, 2026
“There’s a, there’s a strong push from headquarters to do this. They’ve delegated the work to the labs to make it happen. They’ve given us a little bit of funding, but that was spread over a lot of different institutions,” Atchley said. “Given the minimal amount of funding that we’ve got, the focus right now is specifically on DOE. But we are working with other entities in the hyperscalers [and] academia, because we want to make sure that the interfaces can then be usable as well, so they can tie in. The intent is very much to bring the hyperscalers and academia onto the project. But this first year is very much a sprint to try to show some demonstrations.”
There are a lot of unanswered questions about how Genesis Mission will work in the long run, Atchley said. How will archiving be handled? Will it be controlled by DOE? If so, who will be given access?
“We’re still going along at a very, very small budget right now,” Atchley said. “The plans are very large. The scope that they have, the vision they’ve laid out, is enormous. We’re not going to be able to do it in a one-year sprint. We’re not going to be able to do it, even if well-funded, in just a few years. It’s something that’s going to grow organically. But it does need funding right now. We’re trying to do a lot with very little. But if the funding doesn’t come, then it will probably wither in time.”
 
As investment in Physical AI accelerates, simulation is taking on a much larger role than…
Scientists have relied on HPC facilities to help them process large volumes of data, execute…
Nvidia this week highlighted an early deployment of its Vera CPU at Los Alamos National…
The world spent more than $300 billion on AI data centers in 2025, representing a…
As previously noted, the Seventh National Research Platform (7NRP) workshop was held in La Jolla,…
The buzz of Hamburg, Germany over the past two days has been LineShine, the Chinese…
Quantum computing has become a strategic consideration for high-performance computing (HPC) centers worldwide. National laboratories,…
China’s National Supercomputing Center in Shenzhen last week introduced LineShine, a new supercomputer that it…
After several months of radio silence on the matter, Honeywell announced today that Quantinuum confidentially…
The incredible demand that AI is placing on technology is forcing changes to be made…
IBM and RIKEN today announced they have successfully used a quantum computer to simulate a…
The Chinese supercomputer LineShine surprised the field and took the number one spot on the…
Engineering has never had more computing power. Yet engineering transformation has never been more difficult….
Meta has announced that it will start manufacturing a new custom AI chip from September…
Cadence Design Systems has been delivering agentic AI capabilities with its semiconductor and system design…
The stars appear to be aligning for SiPearl, the French chip startup that was founded…
The last five days were very challenging for the 125 students participating in the ASC26…
The quantum computing market is entering a new phase as commercial deployments are beginning to…
Alex Woodie Editorial Director +
HPCwire Managing Editor
Jaime Hampton AIwire Managing Editor
Ali Azhar BigDATAwire Managing Editor
Drew Jolly QCwire Managing Editor
Douglas Eadline Contributing Editor
Addison Snell Contributing Editor
Elizabeth Leake Contributing Editor
More Off The Wire
Position: High Performance Computing Engineer
Location: Switzerland
Since 1989, HOT CHIPS has been one of the semiconductor industry’s leading conferences on high-performance microprocessors and related integrated circuits.
15julAll Day16Google Cloud Advanced Computing Community Chicago 2026(All Day)(GMT-05:00) Chicago, IL
For decades, Google Cloud has supported scientific and technical computing for researchers, developers, and engineers across industry, academia, and government labs. With recent major
For decades, Google Cloud has supported scientific and technical computing for researchers, developers, and engineers across industry, academia, and government labs. With recent major investments in HPC infrastructure and AI, Google Cloud aims to be the most comprehensive, capable, and accessible platform for scientific discovery and technical computing.
Join us at in person at Google Chicago to learn more about our commitment to this mission, explore our newest infrastructure, and see how Google Cloud can accelerate your organization’s innovation and productivity. Through a mix of presentations, hands-on labs, panels, and networking, you will gain valuable insights and forge new collaborations. This is your opportunity to connect with Google experts, HPC and AI practitioners, and peers dedicated to elevating their expertise and driving organizational success.

CalendarGoogleCal
16julAll Day18DATA 2026 Conference(All Day)(GMT+01:00) Porto, Portugal
CalendarGoogleCal
19julAll Day20SIGGRAPH 2026(All Day)(GMT-07:00) Los Angeles, CA
CalendarGoogleCal
The Information Nexus of Advanced Computing and Data systems for a High Performance World
© 2025 HPCwire. A TCI Media Publication (Formerly Tabor Communications). All Rights Reserved.
Use of this site is governed by our Terms of Use and Privacy Policy.