We’re bringing AI to the physical economy, at planetary scale.
Atlas AI builds models that detect how places change and forecast what those changes mean. Companies, governments, and development institutions use those forecasts to decide where to build, serve, and invest. The company was founded in 2018 by Stanford scientists and is a public benefit corporation.
We help the people who shape the physical world see what’s coming.
Weather forecasting combines sensors, history, and simulation, so airlines reroute flights, farmers time their harvests, and cities prepare for storms. Atlas AI is building the same kind of forecast for the physical economy: the infrastructure, communities, and markets the world runs on. The Atlas AI platform has informed operational decisions covering more than 190 countries.
The impact we work toward
- Energy
Power grids grow with the neighborhoods and industries that depend on them.
- Connectivity
Telecom networks reach new communities as they form.
- Supply chains
Supply chains keep goods moving as demand shifts across a region.
- Health
Health systems reach every community, including those missing from official maps.
We’re scientists and engineers who build AI that people rely on.
We are guided by four values.
- i
Honesty and reflection
We err toward transparency, we have the hard conversations, and we learn from failures as readily as from wins.
- ii
Embracing the unsolved
We go after problems no one has solved, and we study the whole system around them.
- iii
Improving lives
We respect the nuance of conditions on the ground, and we value technology for what it does for people.
- iv
In service of each other
We work as one team, assume good intent, and invest in where each person is going.
Come build AI for the physical world.
The models you build here inform real decisions in aviation, telecom, public health, and development. Our team works across North America, Europe, and Africa.
What working here is like
- The work
Your models go into production, and you often build them alongside the customer’s own team.
- The data
You work with messy, real-world data: imagery, sensors, surveys, and customers’ own records.
- The team
You work alongside experts in AI, earth science, and economics, and you learn from all of them.
- Your role
You have room to take initiative and shape what the company builds next.
- Released with careWe release a model only when we can manage the risk of its misuse.
- The right AI for each decisionGenerative models where they help, including DiffusionSat and TEOChat from our founders’ research groups at Stanford. Deterministic methods where decisions affect lives or critical operations.
- Clear about performanceWe tell you how each model performs and where it applies.
Our founders helped pioneer the field.
In 2016, David Lobell, Marshall Burke, and Stefano Ermon showed that satellite imagery and machine learning could estimate economic well-being where survey data was missing.
Selected honors
- Google Cloud · 2024Partner of the YearIndustry Solution, Technology: Sustainability
- World Sustainability Awards · 2023Startup of the Year
- Gavi · 2024INFUSE PacesetterOne of four selected from more than 100 applicants
- Geoawesome · 2026Top 100 Geospatial Companies
- NSF · 2020–28SBIR Phase II research grantInstance Segmentation in Support of Sustainable Development
See what we’re working on.
Perspectives, our blog, has essays, research, and product news from the Atlas AI team.