Submitted
2025 Global Climate Challenge

SolarTFT

Team Leader
Louiza Ait Mouloud
Our solution is a foundation model designed for solar energy forecasting by replacing outdated, rigid systems with a smarter, data-driven approach. Unlike conventional tools that rely on weather updates every three hours or simplified simulations, our model is trained on multi years of real-world solar farm production data. It delivers probabilistic forecasts, such as “85% chance of 15–22 MW,” giving...
What is the name of your organization?
SOLAR-TFT LLC
What is the name of your solution?
SolarTFT
Provide a one-line summary or tagline for your solution.
Generative AI for Solar Forecasting
In what city, town, or region is your solution team headquartered?
Delaware, USA
In what country is your solution team headquartered?
USA
What type of organization is your solution team?
For-profit, including B-Corp or similar models
Film your elevator pitch.
What specific problem are you solving?
The Solar Energy Paradox In 2023, the U.S. added 32 gigawatts of new solar capacity—enough to power over 7 million homes. Yet, 12% of this energy is wasted every year due to inaccurate forecasting. That translates to $1.2 billion in lost revenue and millions of tons of unnecessary CO₂ emissions as fossil fuel plants fill the gap. Why? Most solar forecasts rely on outdated weather models that update every three hours essentially guessing tomorrow’s solar output based on yesterday’s conditions. In California, this outdated method led to a 14% mismatch between forecasted and actual energy in 2023, triggering $6.8 million in penalties. The ripple effects are massive. In Texas, cloud mispredictions during a heatwave led to $220 million in backup energy costs from gas peaker plants. Residential users lose 22% of stored battery energy from false “sunny” predictions. Farms in regions like Arizona suffer losses from unaccounted local risks, like dust storms, while new projects are rejected due to unreliable forecasting. To make matters worse, many current forecasting tools are black boxes—hard to understand, hard to trust, and extremely expensive. The result? Billions in preventable losses. The solar industry needs a forecasting solution that's accurate, affordable, and transparent.
What is your solution?
Our solution is a foundation model designed for solar energy forecasting by replacing outdated, rigid systems with a smarter, data-driven approach. Unlike conventional tools that rely on weather updates every three hours or simplified simulations, our model is trained on multi years of real-world solar farm production data. It delivers probabilistic forecasts, such as “85% chance of 15–22 MW,” giving operators a more realistic, risk-aware understanding of expected output. What sets this solution apart is transparency. With one click, users can interpret why a specific forecast was made, seeing exactly which variables influenced the outcome. This level of insight is critical as regulatory standards tighten and grid operators demand greater reliability. In an environment where forecasting errors now lead to daily fines, clarity isn’t a luxury it’s a necessity. Purpose-built for small to mid-sized operators, our solution is lightweight, affordable, and stripped of enterprise-level complexity. It empowers users to avoid penalties, reduce reliance on fossil fuel backup systems, and optimize battery storage with smarter, more adaptable planning. By correcting the forecasting blind spots that cost the industry over $1 billion annually, our foundation model offers more than predictions it delivers precision, confidence, and a path toward cleaner, more profitable solar energy.
Who does your solution serve, and in what ways will the solution impact their lives?
Our solution directly serves solar farm operators, grid managers, and renewable energy developers, particularly small to mid-sized entities that are currently underserved by high-cost, opaque forecasting tools. These stakeholders face significant financial and operational challenges due to unreliable solar predictions, leading to penalties, wasted energy, inefficient battery use, and lost trust in solar infrastructure. Today, most forecasting tools are either too expensive, too complex, or fail to capture real-world production patterns. This disproportionately impacts smaller operators who lack the resources to invest in enterprise-grade software or build in-house data science teams. As a result, they suffer from high error rates, regulatory non-compliance, and missed revenue opportunities. Our foundation model levels the playing field by offering transparent, probabilistic, and data-driven forecasts that are easy to interpret and tailored to real-world solar conditions. It empowers users with accurate insights, helps them avoid fines, reduce fossil fuel fallback, and make smarter decisions around storage and dispatch. In doing so, we not only improve the profitability and resilience of clean energy providers, but also contribute to broader climate goals by enhancing trust and reliability in solar power. We are enabling a cleaner, more equitable energy future—starting with those most often left behind.
Solution Team:
Louiza Ait Mouloud
Louiza Ait Mouloud
CEO