Atmo is an AI meteorology company building deep learning models that produce ultra-precise, high-resolution weather forecasts. The company claims its forecasts are up to 50% more accurate, 10x more detailed and roughly 40,000x faster to compute than traditional numerical weather prediction models, making it possible to resolve micro-climate phenomena at fine spatial and temporal granularity.
Atmo ingests real-time observational data from satellites, ground stations, weather radars and ocean buoys, then continuously refines its models. Outputs are delivered as forecasts, nowcasts and risk products tailored to high-stakes operational users. The company also explores weather modification research alongside forecasting.
Atmo was co-founded by Alex Levy (CEO), who previously co-founded AI drug discovery company Atomwise, and Johan Mathe (CTO), a former Google X engineer who worked on Project Loon. The team blends AI research, atmospheric science and applied engineering, and the company was formerly known as Froglabs.ai before rebranding to Atmo.
The company has raised approximately $19.7 million across four rounds, with its latest a Series A in December 2024. Felicis is among the noted investors, and the company has continued to expand its commercial footprint into 2025.
Atmo's traction includes named engagements with the U.S. Department of Defense, top aerospace companies, the government of the Philippines and the United Nations. Differentiators include speed and resolution from pure neural forecasting versus legacy numerical models, plus partnerships with national governments rather than purely commercial users.