In a world where weather and climate data are pouring in at unprecedented rates, traditional tools are struggling to keep up. But could AI be the key to revolutionizing the future of weather forecasting?
Research suggests the answer is a resounding yes. Now, a newly funded startup called Brightband is stepping up to redefine the landscape by merging cutting-edge machine learning with weather forecasting—and making it open source in the process.
Current weather prediction techniques are primarily built on decades-old statistical and numerical models. While these physics-based models have been effective, they’re far from efficient. Running them typically requires a supercomputer and days of processing time.
That’s where AI comes in. By analyzing vast amounts of weather data, AI has proven its ability to recognize patterns and predict future events with surprising accuracy. So why isn’t AI-powered forecasting everywhere yet?
Bridging the Talent and Industry Gap
“The gap is largely due to talent and focus,” says Julian Green, CEO and co-founder of Brightband, formerly known as OpenEarthAI. “Government agencies and weather companies struggle to attract top-tier AI talent. Meanwhile, big tech companies—who certainly have the talent—don’t prioritize weather because it’s not their core industry. They don’t invest deeply enough to deliver the tools the weather industry truly needs.”
Brightband believes the solution lies in bringing together the best minds in AI, data science, and meteorology within a nimble startup. Their mission? Operationalize AI-based weather forecasting and make it accessible to everyone.
Building on the Shoulders of Giants
Brightband is currently working on its own AI-driven forecasting model, trained on years of global weather observations. However, as Daniel Rothenberg, co-founder and Head of Data and Weather, explains, they’re not starting from scratch.
“We’re standing on the shoulders of giants,” Rothenberg says, acknowledging the long-standing contributions of physics-based models to the field. “These models are incredibly powerful, and AI is a direct beneficiary of that groundwork. The first breakthrough was realizing that AI could learn from these existing models and their massive datasets. We’re building on that foundation and aiming to push the envelope even further. Our goal is to deliver state-of-the-art forecasts that are as good, if not better, than the best global models available today.”
But the real game-changer is speed. AI models have the potential to provide forecasts orders of magnitude faster than traditional methods, allowing for rapid decision-making across industries.
“That’s the core disruption,” Green notes. “AI-based models are not only faster but also cheaper to run, making them ideal for industries that require fast-moving, highly customized forecasts.”
Real-World Applications Across Industries
Brightband sees vast potential for its AI-driven models in a variety of sectors.
“Different industries have specific needs when it comes to weather forecasting,” Green explains. “Energy companies, for example, need to forecast renewable energy supply from wind and solar, while predicting demand for heating and cooling. Transportation companies need to plan routes around extreme weather events, and in agriculture, farmers need to schedule labor for planting, watering, and harvesting weeks in advance based on forecasted conditions.”
By making AI-powered weather forecasting faster and more accessible, Brightband aims to provide businesses with the tools they need to make real-time decisions based on precise, localized data.
Open Source: The Power of Collaboration
One of the most exciting aspects of Brightband’s approach is its commitment to open-source development. The company plans to release its core forecasting models for anyone to use, along with the datasets and metrics needed to train and evaluate them.
“Our goal is to democratize access to high-quality forecasting,” Green says. “We’re not just talking about releasing the models, but also the data used to train them and the evaluation metrics. By open-sourcing the foundational tools, we can accelerate innovation across industries.”
On top of the open-source model, Brightband plans to offer paid services for more specific and advanced forecasting needs, creating a business model that supports both community-driven development and commercial applications.
This open approach extends to data collection as well. Rothenberg points out that a significant portion of weather data—especially historical data from sources like weather balloons and satellites—has been largely ignored because it’s difficult to process.
“There are petabytes of historical weather data that have been overlooked simply because they’re cumbersome to work with,” Rothenberg says. “But for AI, the more data, the better. By including this often-ignored data, we can significantly improve the accuracy of our forecasts. We believe that building a collaborative community around this data will enhance our understanding of the atmosphere at an unprecedented scale.”
Working with Existing Institutions
While Brightband’s mission might sound like the work of a government agency, Green is quick to clarify that they view their efforts as complementary, not competitive.
“We work closely with agencies like the National Weather Service, who are the custodians of vital observational data,” Green explains. “Our goal isn’t to replace them but to extend their efforts by making data more usable and portable for fast-moving industries. We see this as a continuation of the international collaboration on weather data that has always existed.”
A Product on the Horizon
Though still in its early stages, Brightband is making swift progress.
“We’ve only been at this for a few months,” Green admits, “but we’re hoping to have our first model ready by the end of 2025. The goal is to ingest real-time observations, like satellite and radar imagery, and generate accurate forecasts in real time.”
The company is structured as a public benefit corporation (PBC), which Green describes as more of a symbolic gesture than a constraint.
“Being a PBC is about signaling our commitment to transparency and mission-driven work,” Green says. “It means we’re legally bound to consider not only the interests of shareholders but also the broader impact of our work. The $10 million we’ve raised in our Series A round shows that we can attract capital while staying true to our values.”
The round was led by Prelude Venture and included participation from Starshot Capital, Garage Capital, Future Back Ventures, Preston-Werner Ventures, CLAI Ventures, Adrien Treuille, and Cal Henderson.
What’s Next for Brightband?
While the company doesn’t have a strict timeline for its climate-related projects, Green says a weather-focused product will likely come first.
For now, the goal is to have a working model ready for a demonstration by the end of 2025, and it’s clear that Brightband is on a mission to shake up the world of weather forecasting. By combining AI with open-source principles, they aim to not only modernize an industry but also empower a global community to advance the science of weather prediction together.
