Berkeley Lab deploys Meta models inside DOE's Genesis Mission

SYNAPS-I fine-tunes SAM 3 and DINOv3 on beamline data, then runs them across 300 Nvidia A100 GPUs inside government infrastructure.

By · Published

Primary source: Meta AI

Why it matters

Genesis is becoming a federal distribution channel for AI that can run behind government firewalls. Meta's deployment shows where open model weights can beat cloud-only APIs: sensitive, specialized workloads that institutions must operate themselves.

Illustration of Berkeley Lab's SYNAPS-I fine-tuning Meta SAM 3 and DINOv3 on beamline data and running them across 300 Nvidia A100 GPUs in DOE Genesis Mission facilities.

Alex Hexemer, the Berkeley Lab scientist leading the SYNAPS-I project, is using Meta's SAM 3 and DINOv3 vision models to turn raw X-ray scans into labeled, three-dimensional images while experiments are still running. Meta detailed the deployment in a July 21st blog post, positioning the work as an early implementation of the Department of Energy's Genesis Mission.…

When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.

Reader comments

Conversation for this story loads after sign-in.