Lium (formerly Astromind) builds an "agentic harness" platform that creates custom AI agents to restructure raw scientific and industrial datasets—such as satellite imagery, seismic surveys, electromagnetic spectrum data, and sparse X‑ray measurements—so large language models can reason over them. The company reports early traction with paying customers, including work with astrophysicists on exoplanet X‑ray analysis, processing terabytes of satellite and weather data for the North Carolina Institute for Climate Studies, and automating electromagnetic spectrum analysis for industrial partner nexGEN Inc. Lium says its agents improve over time as more queries hit a dataset, aiming to reduce LLM hallucinations when handling complex numerical inputs. The startup raised a $5.5M seed led by SJF Ventures to launch and scale this platform. No revenue figures were disclosed in the article.

Total raised

$6M

Funding rounds

1

Latest round

Seed

Latest activity

Jun 2026