PNNL's core role in the Genesis Mission: AI that moves experiments, the grid, and materials science faster
- PNNL will be a core contributor on national AI projects for the grid, materials, and security.
- Genesis Mission unites all 17 U.S. national labs to double R&D productivity.
- AI workflows will speed experiments and provide decision support for grid operators; potential lift for local jobs.
The Department of Energy's Genesis Mission is an all-labs push to pair scientists with intelligent systems that can reason, simulate, and run experiments at speed. Pacific Northwest National Laboratory (PNNL) in Richland will lead and co-lead efforts in grid operations, autonomous materials characterization, and AI-driven lab workflows.
PNNL leadership framed the effort as a high-urgency reset for how science is done across energy, discovery research, and national security. DOE described the target in plain terms: double the productivity and impact of U.S. R&D.
Context and impact in the Tri-Cities
The national lab is the Tri-Cities' largest single employer. After a drop of nearly 400 positions from 6,440 in fiscal year 2024 due to uncertain federal funding, Genesis could open a path to restore roles aligned to AI, automation, and data-intensive research.
Project 1: Real-time AI support for the U.S. power grid
The electric grid is a live, high-dimensional system-ideal for AI that learns from control data, historical events, and stress tests. PNNL will be the core technical contributor among nine labs building a continuously evolving AI platform that supports grid operators with reliable, real-time intelligence.
Operators often have as little as three hours to make market and reliability decisions. The team is training models to assist with planning, forecasting, and operational decision support so staff can review more scenarios, surface edge cases, and act with clearer confidence.
For more on the national labs network that's collaborating on this effort, see the DOE overview of the system of labs here.
Project 2: Autonomous analysis of nuclear materials
PNNL will apply AI with scanning electron microscopy to analyze uranium and plutonium alloys tied to the nation's stockpile stewardship mission. The workflow images samples, interprets data, and selects next experiments with little to no human intervention, moving researchers from repetitive collection to higher-order materials questions.
Throughput is the key shift. The team processed fewer than 10 priority uranium-alloy samples in the last three years; the goal is to analyze 10 samples every quarter by the end of the project-providing NNSA, design labs, and production plants with faster, data-rich feedback on materials and manufacturing options. Learn more about the mission context at the National Nuclear Security Administration.
Project 3: AMP2 and AI-guided biology at scale
DOE leadership recently commissioned PNNL's prototype Anaerobic Microbial Phenotyping Platform (AMP2). It combines robotics with AI to run microbe experiments, analyze outcomes, and suggest the next set of tests-accelerating work on organisms used to produce chemicals, fuels, and biomaterials.
Construction is planned for an EMSL addition to house a larger system. PNNL is developing a "translation" layer so instruments and datasets across modalities-mass spec, genomics, microscopy-can interoperate and route clear instructions for the next experiment.
The goal is better experiments and more of them. Instead of tuning one growth condition by hand, the system can scan many conditions and datasets to find settings that increase rate and hold up when variables shift.
What researchers should watch
- Data and model standards: Expect momentum on shared vocabularies, metadata, and model governance across labs and instruments.
- Human-in-the-loop design: Systems will prioritize operator trust-traceable reasoning, uncertainty estimates, and fast what-if testing.
- Validation at operational tempo: Grid and nuclear workflows will stress fast, repeatable validation tied to real risk and compliance.
- Skills shift: Demand will grow for researchers fluent in experiment design, automation, and ML ops-not just model training.
- Compute + security: Tight coupling of HPC, secure data enclaves, and automated pipelines will become baseline lab infrastructure.
Related learning
If you're integrating AI into your lab workflows, see the AI Learning Path for Research Scientists. Working with microbes and automated phenotyping? Explore the AI Learning Path for Microbiologists.
Bottom line: PNNL's contributions aim to compress cycle times-from control-room decisions to microscopy feedback loops to automated wet lab iteration-and convert more of your scientific questions into executable workflows.
Your membership also unlocks:
