7/23/2026 Jackson Brunner
Assistant Professor Wenjie Zhou is part of a DOE Genesis Mission project led by Lawrence Livermore National Laboratory to build AI-enabled data pipelines that will make it possible to 3D-print fusion fuel capsules at the scale and precision needed for inertial fusion energy, replacing today's unwieldy mesh-based design files with compact geometry descriptions and AI-driven print planning. Zhou's group at Illinois is building the critical interface that converts planned print paths into machine-readable printer instructions and predicts how each capsule will actually print—catching defects before any material is used.
Written by Jackson Brunner
Wenjie Zhou, an assistant professor in the Department of Materials Science and Engineering at The Grainger College of Engineering, University of Illinois Urbana-Champaign, has been named a co-investigator on a new U.S. Department of Energy project to build the AI-enabled data infrastructure needed to 3D-print fuel capsules for inertial fusion energy (IFE).
The project, "AI-Enabled Data Pipelines for High-Precision Additive Manufacturing of Fusion Targets," is among the first projects selected under DOE's Genesis Mission, a national initiative pairing DOE's scientific facilities and supercomputing resources with artificial intelligence to accelerate discovery in energy, science and national security. It is led by Lawrence Livermore National Laboratory (LLNL), under principal investigator Raspberry Simpson, with co-investigators Xiaoxing Xia of LLNL, Zhou at Illinois and Doug James, a professor of computer science at Stanford University.
LLNL has pioneered a 3D-printing process, using two-photon polymerization, to fabricate the small foam-lined capsules that hold fusion fuel, including "wetted foam" targets, seven of which were successfully shot on LLNL's National Ignition Facility for the first time between fiscal years 2024 and 2026. But scaling this fabrication method up for a future IFE pilot plant is bottlenecked not by the printer hardware, but by the underlying design files: standard 3D-printing pipelines describe target geometries as surface meshes that can balloon past 10 gigabytes at the precision fusion targets require. The project aims to replace those mesh-based files with compact, mathematically defined geometry descriptions, paired with AI-driven print-path planning and a "digital twin" that virtually simulates a print job to catch defects before any material is used.
Within the collaboration, LLNL leads construction of the digital twin and its AI optimization agent, while Zhou's group at Illinois builds a key interface that feeds it. Because effects such as overlapping laser exposure, radical diffusion, and oxygen quenching can cause a printed capsule to deviate from even a correctly planned path — through over-curing, under-curing, or distorted features — Zhou's team is developing the interface that translates planned print trajectories into a sparse volumetric representation of the expected build, enabling virtual inspection for these defects before printing. Zhou’s group will also build the interface that converts Stanford's freeform print paths into machine-readable printer commands. Zhou's contribution draws on his broader expertise in geometry- and topology-aware architected materials, including postdoctoral work with LLNL co-PI Xiaoxing Xia that introduced polycatenated architected materials, a new class of entangled materials featured on the cover of Science in 2025, as well as contributions to a Nature study on scaling up two-photon 3D printing through massively parallel metalens arrays.
"Fusion targets are among the most precise objects people have ever manufactured, and one of the biggest obstacles to printing them at scale now is data," Zhou said. "Our job at Illinois sits in the middle of the pipeline: we turn planned print paths into real printer instructions, and then into a faithful prediction of what the printed structure will actually look like, so the problems are caught before any material is used. If we get that right, target fabrication can become an accelerator on the path to fusion energy."
By the end of Phase I, the team aims to demonstrate at least a 100-fold reduction in file size, a five-fold improvement in print resolution, and a two-fold reduction in print time on a physically printed target — results that would support a Phase II effort extending the work into target design and fusion experiment validation at LLNL.
LLNL is leading ten Phase I Genesis Mission projects in all and contributing to 19 more led by partner institutions.
"These selections reflect LLNL's demonstrated ability to integrate AI, advanced computing, experimental science and multidisciplinary expertise to solve problems of national importance," LLNL Director Kim Budil said in an LLNL press release announcing the awards.
For the Department of Materials Science and Engineering, Zhou's role connects Illinois directly to this national effort to accelerate the path to commercial fusion energy.
Illinois Grainger Engineering Affiliations
Wenjie Zhou is an Illinois Grainger Engineering assistant professor in the Department of Materials Science and Engineering. Zhou is affiliated with the Department of Mechanical Science and Engineering and the Materials Research Laboratory.