In Sanskrit, Vāyuh.ai means the element of wind or air. It is one of the most dynamic of the elements. At Vāyuh.ai Inc., we are scientists and engineers from the UC Berkeley community with a deep understanding of AI and climate. We are here to build the world’s best weather forecasts and climate change analytics using Physics and AI.

Climate change affects us all. One of the challenges we face right now is that of good data that helps us understand the impact of climate change. For example, what is the likelihood of damage due to a specific peril such as a wildfire or extreme weather.

Almost all businesses are affected by climate change. The scale and intensity may vary. Sadly, due to inaccurate accounting and poor practices, climate change's impacts are often poorly understood. Businesses with a lot more physical exposure to climate are more directly affected by climate change. And so, Vāyuh.ai helps energy traders, insurers, mortgage companies and supply chains understand the impact of climate change through better forecasts and climate risk models.

Evidence suggests that climate change impacts many perils such as wildfires, flooding, sea level rise, severe convective weather, hurricanes, and more. The impact of each of these perils differs based on the geographical location. For example, Florida is a lot more affected by hurricanes while California is affected by wildfires. Simultaneously, the floodplains of Bangladesh are impacted by rising sea levels.

At Vāyuh.ai, we aspire to provide forecasts across all timelines ranging from short-term weather forecasts (1-3 days), medium-term weather (14-28 days) and longer-term climate forecasts (annual forecasts). Presently, we are focused on temperature forecasts as they are critical to the energy markets. In addition to forecasts, we calculate derived analytics such as wildfire risk, extreme wind damage, severe convective weather, extreme precipitation, and temperature extremes over longer periods of time (up to 2030+).

Predicting weather is usually done by numerical weather prediction models. These models work very well when 1) all the physics is known and 2) the physics can be simulated efficiently on a supercomputer. Both these assumptions are challenged by climate change. Vāyuh.ai's forecasts have the following ingredients. We incorporate physics-based constraints and physics-based forecasts. We then combine this with observations from satellite imagery and gridded data points. Lastly, we leverage the power of machine learning and AI to generate better forecasts.

Yes. At Vāyuh.ai we work hard to constantly search for the best models, and combine them with our world-class AI to produce meaningful risk models. We understand that estimating peril or climate risk is very crucial to a lot of society and businesses such as insurance, mortgages, and supply chains. Therefore, we classify risk paradigms and compute climate risk using future climate scenarios that are more accurate and do not merely rely on historical observations as was done earlier.

Vāyuh.ai's medium-term weather forecasts and peril risk models are significantly better than those offered by incumbents that largely rely on physics-based simulations. Our severe convective weather risk models, for instance, are state of the art and are unique to Vāyuh.ai Inc. You can read more about it in our case studies here and here.

At Vāyuh.ai, we pride ourselves in holding ourselves to rigorous standards before a forecast or risk model is pushed into production. Almost all forecasts are validated on a held-out test sample of a few years. This means that our customers can confidently use our analytics for business decisions, rest assured that the quality of the analytics is some of the best in the world.

The use of analytics is different for each business. In the energy sector, there is a direct link between weather forecasting and the demand/supply of energy. To elaborate, a temperature forecast is directly predictive of demand in a region. A wind forecast, directly predicts power generated by a wind turbine. For other sectors such as insurance, supply chains, and more - a longer-term (year-on-year) risk model is very useful for planning purposes. For underwriters of insurance, these analytics can provide insights into how to price risk for the next year. Just knowing the frequency of extreme weather events can be highly useful for (re)insurers to deploy appropriate capital in markets based on their risk appetite and the needs of society.

For our customers, our forecasts are available on our web platform in an easy-to-use visual display. In addition, we offer an API to download data. We also offer back casts as part of our trials for customers. Please reach out to us at sales@vayuh.ai for more information about anything you’d like to see on there or a free trial.

Climate change is the most pressing problem of our time. At Vāyuh.ai, we believe that tackling climate change requires us to engage with the free markets and that is why we choose to be a for-profit company. Our initial customers and our market research suggest that there is a clear need for what we are building. We are also pledged to invest a percentage of our profits back into mitigating climate change through projects like Savesoil, Rally for Rivers, and the One Trillion Tree Project.

Improving climate risk models and weather forecasting is a complex problem requiring many scientists and engineers from multiple disciplines. We are always open to hearing from the community and embarking on collaborations. Presently, we collaborate with Oakridge National Laboratory, a Department of Energy lab, and work with their scientists.