This post examines a recent collaboration between climate-tech company Planette and NASA’s Goddard Institute for Space Sciences (GISS) aimed at improving long-range extreme weather forecasting.
Drawing on decades of Earth system modeling and emerging quantum computing approaches, the partnership promises to extend early-warning horizons and support businesses and agencies that manage physical assets exposed to climate risks.
Why this partnership matters
As an atmospheric scientist with three decades in climate risk and forecasting, I recognize that improving lead time for extreme-event warnings is one of the most effective ways to reduce economic and human losses.
Enhanced forecasts out to six months, and the prospect of even longer horizons, change how organizations plan for seasonal and anomalous extremes.
Who benefits and how
Planette targets businesses with tangible exposure: utilities, logistics firms, insurers, and municipalities that operate infrastructure vulnerable to floods, storms, heatwaves and other climate-driven events.
By providing probabilistic forecasts and scenario information well in advance, these stakeholders can stage resources, adjust supply chains, and prioritize protective investments.
Key advantages Planette aims to deliver include:
The science: AI trained on Earth system simulations
Planette trains its AI models on established Earth system simulations such as those from NOAA, which explicitly represent coupled ocean-atmosphere processes that drive large-scale variability.
This approach leverages decades of physically consistent modeling rather than relying solely on observational datasets.
How ocean-atmosphere coupling extends forecast skill
Long-range predictability often stems from slow-evolving components like sea surface temperatures, ocean heat content and large-scale circulation patterns.
By training on models that capture these interactions, AI can identify precursors to extremes months in advance.
CEO Hansi Singh highlights that the heavy computational burden is in the training phase; once trained, generating tailored forecasts for clients is inexpensive and operationally scalable.
Planette is also collaborating with the National Science Foundation (NSF) on another advanced AI forecasting model, expanding the technical portfolio and scientific rigor behind their products.
Quantum computing and QubitCast: predicting rare one-offs
Beyond classical AI, Planette and NASA GISS are developing QubitCast, a quantum computing tool designed specifically to address rare, one-off extreme events that are difficult to sample in conventional datasets and models.
Quantum algorithms offer potential efficiency gains for exploring vast, high-dimensional probabilistic spaces.
What quantum adds to the forecasting toolbox
Quantum approaches could enable more efficient exploration of low-probability tails of the climate distribution — the very events that cause catastrophic impacts but elude typical sampling.
While still experimental, QubitCast aims to complement classical AI by probing scenarios that are computationally prohibitive today.
This holds the promise of longer forecasting horizons and better characterization of extreme event likelihoods.
Practical implications: if quantum-accelerated forecasts reliably identify rare event precursors, emergency planners and infrastructure managers gain critical, actionable time to reduce losses.
Hansi Singh has been clear about the stakes: adapting to climate change will be costly and disruptive.
The smartest way to reduce those costs is through early-warning technologies that give organizations time to act.
Planette’s ambition is straightforward: ensure every agency and business with environmental exposure can use these forecasts to prepare for climate-driven challenges.
Here is the source article for this story: How Planette helps NASA use quantum principles for weather forecasting