RICHLAND, Wash.- The Department of Energy's Pacific Northwest National Laboratory has joined collaborators across the nation to develop and deploy artificial intelligence to vastly accelerate the speed of discovery for science, energy and national security.
The Genesis Mission, led by DOE for the nation, brings together all 17 DOE national laboratories in a race to harness AI for national priorities. PNNL is key to the effort, leading and participating in projects central to the nation's security and international competitiveness. As a multi-disciplinary laboratory with a workforce addressing many types of scientific questions, PNNL brings a valued and unique wide-angle lens to the Genesis AI landscape.
"The Genesis Mission represents a pivotal moment of profound national purpose, where the brightest minds across DOE's national labs, industry and academia are coming together to redefine the boundaries of scientific discovery," said Laboratory Director Deb Gracio. "At PNNL, we are seizing this opportunity to shape the future of science with urgency and determination."
At its core, the Genesis Mission is assembling a unified scientific workhorse powered by DOE-funded supercomputers, integrated AI systems and emerging quantum technologies. Few people realize that DOE manages some of the most advanced scientific instruments in the world. But until now, those instruments operated in silos at individual laboratories. The Genesis Mission connects both the instruments and their terabytes of data to an intelligent agentic platform not only capable of interpreting the data rapidly and proposing new experiments but also interpreting data streams from multiple instruments much faster, more efficiently and at lower cost than currently possible. Working together, the instruments, robotic automation and AI-powered analysis form the integrated infrastructure expected to compress years of painstaking work into months, weeks or even hours.
It's a monumental expansion of the kind of exploratory collaboration that PNNL embarked on with Microsoft in 2024 to speed discovery of energy materials.
While the ambition embodied in the Genesis Mission is vast, three projects particularly exemplify how PNNL expertise is shaping the enterprise.
Orchestrated platform for autonomous laboratories
We often think of biology in terms of human health, but biological systems are the ultimate energy transformers and engines. Our food supply, housing needs and economy all depend on complex biological systems. While researchers have successfully harnessed biology to improve health and our food supply, we now need it to do more and do it faster.
The effort will be coordinated by the Orchestrated Platform for Autonomous Laboratories (OPAL), a four-lab collaborative research network led in part by PNNL that will give scientists access to a distributed national lab that combines automation, robotics and data analysis.
Researchers will draw on the resources from across the DOE laboratories, including the Environmental Molecular Sciences Laboratory, a DOE Office of Science national user facility located on the PNNL campus. EMSL is home to the newly commissioned Anaerobic Microbial Phenotyping Platform (AMP2), which has opened a new chapter in autonomous biological discovery by giving researchers the speed and scale needed to accelerate biotechnology discovery. Scientists will use AMP2 to explore questions about anaerobic microorganisms, which are organisms that grow in an oxygen-free environment and that play important roles in industrial processes to make chemicals, fuels and biomaterials.
As part of this effort, PNNL researchers will work toward creating the ultimate translation system for instruments and data sources that currently don't speak the same language.
"We are creating a platform that enables multi-AI workflows," said Chris Oehmen, a technical lead for the project. "What that means is we can ask a question, for example, in the language of a mass spectrometer instrument and get an answer in the language of genes. The goal is to have a platform that can speak multiple data languages, integrate those data and tell us which experiments to do next."
"That's powerful," added Oehmen. "If I'm doing an experiment all by myself in a laboratory, I might say well, I just need this microbe to grow faster. I only get to optimize for one condition. If I'm talking to an AI agent that has access to multiple conditions and datasets, I can say I want the microbe to grow faster, but I also want it to be more robust, so don't find me an optimal growth condition that's really fragile. I want an optimal growth rate that's also very, very stable. The point is not just to do more experiments, it is to do better experiments."
This is going to change the way scientists think about their work, Oehmen added. It will accelerate discovery toward rapid manufacturing of biologically based, difficult-to-make specialty chemicals and even potentially recovery of critical minerals from sea water, for example.
AI for planning and operation of the U.S. power grid
PNNL will bring its expertise in AI and electric grid modernization to the Genesis Mission. Often called the world's largest machine, the U.S. power grid is the backbone of the nation's economic engine. Its sheer complexity makes it an almost perfect test case for taking advantage of the power of AI to provide just-in-time data and analysis. As a key component of the Genesis Mission, a grid-focused multi-lab endeavor is also among the most closely tied to boots-on-the-ground commercial operators.
Today, grid operators rely on firsthand experience and limited data to make operational decisions. The proposed national AI platform for energy systems operators is designed to evolve continuously, learn securely and support operators in real time with adaptive, reliable intelligence. The massive endeavor is led by the National Laboratory of the Rockies, with Argonne National Laboratory and Oak Ridge National Laboratory serving as co-leads, and support from six additional laboratories including PNNL as a core technical contributor.
The labs have already worked together on several grid-related projects; for example, PNNL and NLR together completed the multi-year National Transmission Planning Study to help chart out the future of grid resources across the nation. Researchers say the Genesis Mission will speed up similar studies considerably.
"When operators often work with tight deadlines to complete tasks such as reviewing data as part of market operations, they can have as little as three hours to make important decisions," said Marcelo Elizondo, one of the technical leads for the PNNL team. "We're hoping that AI models can be useful to support operators for those tasks, as well as to derive important information to help operator's decisions regarding power system reliability."
The model will be trained on grid control data, past grid adverse events and stress testing from open and proprietary data sources. Then grid operators will test its ability to support planning, forecasting and decision-making, all key elements of electric grid modernization.
"A platform that realizes that kind of speed is going to be a game changer for us and for our industry partners," Elizondo said.
Autonomous Characterization of Materials Across Scales (ACMAS)
As part of the Genesis Mission, PNNL scientists will use their decades-long expertise in nuclear science to deploy AI to create new ways to analyze nuclear materials.
The work is part of an enduring effort, led by the National Nuclear Security Administration, to modernize the U.S. nuclear weapons stockpile and ensure the safety, security and reliability of the U.S. nuclear deterrent.
"In order for us to deploy modern manufacturing that is faster and more efficient, we have to make sure that the new production approaches give us materials that behave the same way as the old processes. Proving that requires state-of-the-art materials science capabilities," said Aaron Luttman, senior advisor to ACMAS.
"The pace just isn't where we need the pace to be. To have a truly agile deterrent, we need to characterize stockpile materials at orders-of-magnitude faster speeds, and that requires AI," he added.
PNNL leads the effort and brings extensive experience analyzing the properties of uranium and plutonium, a role that traces back to the Manhattan Project. For this 21st-century mission in nuclear material science, the team is designing and deploying an AI agentic framework that transforms scanning electron microscopes into autonomous materials science platforms. The system will image large-scale material samples, analyze the data, use the results to determine which data to collect next and then iterate through the process with little-to-no human intervention. This frees up PNNL researchers to focus on the fundamental science of material performance and offloads the tedious tasks of data collection to the AI system.
"Over the last three years, we've had the opportunity to process fewer than 10 samples of one of our most important uranium alloys," said Luttman. "At the end of this project, we will be analyzing 10 samples every three months. This will provide the NNSA and its design laboratories and production plants with new options for materials and manufacturing, at a pace that we've never seen before. That's what I'm excited about."
In addition to their involvement in specific projects, PNNL scientists are leading broader efforts. Court Corley, chief scientist for AI at PNNL, is co-PI of an effort known as the Transformational AI Models Consortium, with a focus on developing AI models to revolutionize scientific discovery. PNNL scientists are also part of core teams developing new AI models to address other challenges.
The Genesis Mission is supported by DOE in collaboration with the White House Office of Science and Technology Policy. The initiative also includes several industry partners and academic researchers.
"The Genesis Mission is giving us a once-in-a-generation opportunity to build the world's greatest scientific instrument and collaborative infrastructure for discovery science, energy and national security," said Corley, who is also director of the Center for AI @ PNNL. "We must deliver this capability for American innovation, and we are rising to the moment."
