U.S. Department of Energy national laboratories are deploying a groundbreaking AI system that processes imaging data in real time, turning weeks of analysis into seconds and enabling scientists to make discoveries faster than ever before. The platform, called SYNAPS-I (Synergistic Neutron and Photon Autonomous Science Intelligence), represents a fundamental shift in how researchers work with the massive data streams generated by advanced scientific facilities.

Why Are National Labs Drowning in Data They Can't Analyze?

Modern X-ray, microscopy, and neutron facilities at U.S. national laboratories produce staggering volumes of high-quality imagery every day. The problem is straightforward: humans cannot keep pace with the sheer speed and scale of information being generated . Scientists face a critical bottleneck where valuable data sits unanalyzed because traditional computational methods are too slow to handle the workload. This creates a frustrating paradox: the more advanced the equipment becomes, the further behind scientists fall in understanding what their experiments reveal.

SYNAPS-I addresses this challenge head-on by integrating data from neutron, X-ray, and microscopy experiments across multiple national laboratories into a single unified AI model. This model analyzes information across different scales and accelerates understanding of complex systems in real time . The platform is being developed as part of the DOE's Genesis Mission, a historic national effort to transform American science and innovation through artificial intelligence.

How Does SYNAPS-I Actually Work at the Beamline?

The SYNAPS-I team, led by Alexander Hexemer at Lawrence Berkeley National Laboratory and including researchers from Brookhaven, SLAC, and Oak Ridge National Laboratories, chose to start with a specialized X-ray technique called ptychography . This technique gathers overlapping diffraction patterns (the distinctive ways X-rays scatter after hitting a material) and uses computation to reconstruct sharp, high-resolution images. The beauty of ptychography is that it sidesteps the physical limits of traditional X-ray optics by using physics and computational reconstruction to achieve detail finer than the beam itself can reveal.

What makes SYNAPS-I revolutionary is speed. The platform delivers high-resolution ptychographic images fast enough to keep pace with experiments in real time. During a recent test at Argonne's Advanced Photon Source (APS), the world's brightest synchrotron X-ray source, SYNAPS-I captured diffraction patterns and reconstructed them into high-resolution images instantly . The system compresses hours or days of analysis into seconds, enabling scientists to see results as experiments unfold rather than weeks later.

The AI engine works by building the physics of coherent imaging directly into the model itself. As one researcher explained, this approach gives the AI the same knowledge a scientist would use, making it far more accurate and efficient when handling the massive data volumes produced at DOE facilities . The platform uses advanced computing resources from the Argonne Leadership Computing Facility (ALCF) and the National Energy Research Scientific Computing Center (NERSC) to handle the computational load.

Steps to Deploying AI-Driven Scientific Discovery at Scale

• Integrate Multimodal Data: SYNAPS-I trains a billion-parameter foundation model on data from more than 100 beamlines across seven DOE facilities, processing different types of data such as text and images simultaneously to create a unified understanding of complex systems.
• Embed Domain Physics: Rather than treating AI as a black box, researchers built the physics of coherent imaging directly into the model, ensuring the system understands the underlying science and can make more accurate predictions and analyses.
• Deploy at Beamline Scale: The platform operates in real time at experiment stations, capturing data and displaying imaging results instantly for researchers to view and guide their next experimental steps without delays.
• Refine Through High-Performance Computing: Saved data is moved to leadership computing facilities where high-performance computing resources refine the models continuously, improving accuracy and expanding capabilities over time.

The recent test at Argonne's 26-ID beamline demonstrated ptychography capabilities that were 10 times higher in resolution than previous methods . More importantly, the test opened the door to real-time identification of defects in materials, enabling researchers to guide manufacturing processes and launch autonomous discovery campaigns. These are largely self-driving research efforts in which AI systems help design experiments, analyze results, and determine next steps, accelerating the search for promising new materials.

"SYNAPS-I is envisioned not just as a tool for analysis and automation, but as a cognitive partner for scientists, capable of generating hypotheses, detecting subtle correlations and helping turn DOE facilities into truly intelligent, self-driving laboratories," said Mathew Cherukara, an Argonne computational scientist and leader of the Argonne SYNAPS-I team.

Mathew Cherukara, Computational Scientist and Group Leader at Argonne National Laboratory

What Industries Will Benefit From This Technology?

SYNAPS-I is designed to accelerate breakthroughs across multiple critical sectors. The platform works across domains including microelectronics, biomedical research, advanced manufacturing, and energy security . By cutting imaging analysis from years to days and enabling real-time, AI-driven materials design, SYNAPS-I could deliver substantial economic gains by eliminating costly bottlenecks and speeding innovation. This capability is particularly important as the United States competes globally in advanced manufacturing and semiconductor development.

The technology also connects to broader national priorities around critical materials supply chains. In a complementary initiative, Argonne is partnering with Aclara Resources to develop an AI-enabled digital twin for heavy rare earth separation . By pairing Argonne's AI-powered process simulation platform with data from Aclara's pilot plant, the team aims to reduce both time and cost as operations scale from pilot facilities to commercial production. This collaboration supports Aclara's plan to build a $277 million heavy rare earth separation facility in Louisiana, the first of its kind in the U.S. with a secured ionic clay raw material feed .

"By combining real-world industry needs with deep expertise in the science of scale-up at national laboratories, we can speed the deployment of advanced technologies that strengthen domestic supply chains," said Claus Daniel, associate laboratory director for Argonne's Advanced Energy Technologies directorate.

Claus Daniel, Associate Laboratory Director for Advanced Energy Technologies at Argonne National Laboratory

When Aclara's Louisiana facility begins production in 2028, it is expected to supply more than 75 percent of U.S. demand for dysprosium and terbium, rare earth materials used in high-strength magnets that enable efficient operation of electric motors . By that time, Aclara will have more than a year of operational data from its pilot plant, enabling faster ramp-up, improved efficiency, and stable long-term operations.

What Does This Mean for the Future of Scientific Research?

SYNAPS-I represents a fundamental transformation in how scientific discovery happens. Rather than scientists waiting weeks or months to analyze data, they can now see results in real time and adjust their experiments on the fly. This acceleration compounds over time, allowing researchers to explore more hypotheses, test more materials, and make breakthroughs faster than traditional methods permit. The platform essentially turns national laboratories into what researchers call "self-driving" facilities, where AI systems actively participate in designing experiments, analyzing results, and recommending next steps.

The implications extend beyond speed. By embedding scientific knowledge directly into AI models, researchers ensure that the technology understands the underlying physics and chemistry, making it more reliable and interpretable than generic machine learning approaches. This is critical for scientific work, where understanding not just what happened but why it happened matters enormously. As SYNAPS-I scales across more than 100 beamlines at seven DOE facilities, it will create an unprecedented archive of scientific knowledge, enabling discoveries that would be impossible with isolated datasets.

The convergence of real-time AI analysis, high-performance computing, and domain-specific knowledge represents the next generation of scientific infrastructure. For the United States, this technology strengthens global competitiveness in advanced manufacturing, materials science, and energy innovation. The partnerships between national laboratories and industry, exemplified by the Argonne-Aclara collaboration, show how federal research investments can accelerate the path from laboratory discovery to commercial deployment, ultimately supporting economic growth and national security.