Lium (https://lium.ai), a Dallas-based conversational AI platform, has raised $5.5M in seed funding from SJF Ventures, Wavemaker 360, Reach Capital, and GC&H Investments. The platform acts as an 'agentic harness' that ingests large, multimodal scientific and engineering datasets and makes them queryable through natural language. The capital will support commercial expansion after emerging from stealth.
Agentic Platforms Challenge Geospatial Incumbents
The timing comes amid rising investment in AI for physical-world data. Descartes Labs raised $61.3M over five rounds through 2019. Blackshark.ai secured $35M through November 2023. Hex raised $172M through May 2025. Lium differentiates by spanning energy, space, climate, and infrastructure datasets rather than specializing in satellite imagery alone, while using conversational interfaces instead of pre-built dashboards.
LLMs Struggle With Measured-World Data
General-purpose models fail on proprietary formats such as seismic SEG-Y files, X-ray observations, and sensor streams. Early validation includes processing over 100 terabytes of NOAA climate data for NCICS in weeks and building 50 reusable tools. Customers such as nexGEN and Imaged Reality demonstrate demand from energy and geoscience sectors where data engineers are scarce.
Agentic Harness Structures Raw Data
Lium automatically indexes databases, files, and APIs, provisions compute for terabyte workloads, and converts results into reusable artifacts. This approach differs from notebook platforms like Hex that assume clean structured data and from imagery specialists that target single verticals. Analyses become shared organizational knowledge that improves over successive queries.
"The constraint isn't access anymore — it's usability. Lium is fundamentally reinventing data architecture, moving beyond data lakes and data warehouses to create a living, explorable data universe." — Ward Vuillemot, CTO
Mission Investors Back Scientific AI
SJF Ventures and Reach Capital bring sector expertise in climate and enterprise software. Their participation signals conviction that domain-specific agentic systems can unlock value in industries where generic LLMs currently fall short.
Geospatial AI Market Expands Rapidly
The geospatial analytics AI market stands at $45.22B in 2026 and is projected to reach $126.58B by 2035 at 12.69% CAGR. Broader data science platforms are valued at $204B. Structural drivers include exploding volumes of satellite, seismic, and instrument data plus the recognized limits of general LLMs on non-text formats.
Ex-Microsoft Team Brings Domain Credibility
Founders Josh Knutson and Ryan Thill, along with CTO Ward Vuillemot, previously worked at Microsoft, Amazon, and Google. The team developed the core technology on NASA's Chandra X-ray Observatory data before expanding to energy and climate applications.
What's Next
Lium launched publicly in June 2026 with early deployments at NCICS and nexGEN. The platform is positioned for broader adoption across energy transition and scientific research verticals.