AI for a solar powered future

We find all solar PV systems and forecast their output

Solar energy management using Artificial Intelligence and Data Science.

Our Ecosystem of partners, programs, and clients

Our Ecosystem of partners, programs, and clients

<span data-metadata=""><span data-buffer="">Using AI to distribute resources evenly

Increasing visibility of solar generation through cloud movement forecasting

Solutions

State-of-the-art insights for advanced solar energy management

Identification and forecasting for individual sites or entire regions

Awareness

Find the precise locations of all solar and understand your customer base

Accuracy

Forecast solar energy fluctuations in your region and fine-tune your bidding

Adaptability

Be ready for the transition to the solar-powered energy system of the future

<span data-metadata=""><span data-buffer="">Solar Identification

Determine the exact sizes, locations, and orientations of all distributed solar PV

Understand and predict uptake

Differentiate between different asset types

Illustration of identifying solar panels within a region
Solstice AI uses artificial intelligence and data science to forecast solar PV to manage energy distribution within a region

<span data-metadata=""><span data-buffer="">Solar Generation Forecasting

Predict cloud formation and movement using AI

Obtain forecasts specific to a site, network region, or postcode

Prepare for major shortfalls of solar generation

Consulting<span data-metadata="">

Custom solutions for your solar energy management needs

Extract unique and actionable insights from large datasets

Leverage the latest machine learning and AI techniques

Benefit from a highly experienced, PhD-qualified team

Solstice AI provides consulting solutions for clients using Solstice AI's expertise in data science for solar and renewable energy

Get in touch

We’d love to hear from you

Our Team

Meet the founding team

A long track record of extensive industry and research experience

Julian de Hoog
Co-Founder & CEO

Dr. Julian de Hoog completed his PhD in Computer Science at the University of Oxford, and a postdoc at the University of Melbourne studying the impact of electric vehicles on distribution networks. Prior to founding Solstice AI, he spent six years as a senior research scientist at IBM Research, where he worked on renewable energy forecasting and optimal control of energy storage.

Peter Ilfrich
Co-Founder & CTO

Peter Ilfrich is an experienced full-stack software engineer and architect with a German Diploma in computer science. He has worked in multiple domains (e-commerce, banking, healthcare, energy) and is familiar with a broad spectrum of technologies, computing infrastructure and methodologies. He previously worked as senior software engineer for IBM Research.

Maneesha Perera
Co-Founder & Head of Research

Dr. Maneesha Perera completed her PhD at the University of Melbourne in 2023. Her research focused on improving solar power forecasting using artificial intelligence. She has also held prior roles at IBM Research and as a full stack software engineer at Sysco Systems.

Valentin Muenzel
Mentor

Valentin Muenzel is Co-Founder and ex-CEO of Relectrify, a world leader in battery control, backed by leading global investors incl global utility EDP and Toyota Ventures, and Asia Pacific Company of the Year in the Global Cleantech100.

FAQ

Questions and Answers

Everything you need to knoe about Solstice AI™ and what we do

What are Solstice AI's products?

We are currently offering two main products.  

Solstice SkyScan™ scans extended regions (such as entire states) and finds the exact locations, sizes, and orientations of all solar PV systems.  This includes residential (rooftop), commercial/industrial, and large-scale (solar farms).

Solstice SkyFlowforecasts the formation and movements of clouds, the main driver of solar variability, enabling highly accurate solar generation forecasts.

We can provide solar generation forecasts for individual sites (including solar farms), or for entire regions.

We train artificial intelligence models on millions of images of solar PV systems.  They can differentiate between solar PV and solar hot water, and they can determine the orientations of solar PV systems.

If the underlying imagery is good enough, our models have an accuracy of 97%.  

As a result, we can scan large regions and quickly identify exactly where all the solar PV systems are, and what their generating capacities are.

Using a combination of artificial intelligence and cloud motion vectors, we examine real-time satellite images and forecast how clouds will form and move in the next 2-3 hours.  Each artificial intelligence model is trained specifically on the region of interest.  This is important, since local factors such as coastlines, hills, or mountains can affect cloud formation.  The resulting solar generation forecasts are more accurate than if global irradiation models are used.

For solar identification, we partner with several third parties to obtain overhead imagery.  For most populated regions there is high-quality imagery available that is updated regularly.  Some countries and regions have their own local data providers, but this generally has to be checked on a case-by-case basis. If data with sufficient resolution is available, we can procure it and analyse it.

For solar forecasting, satellite imagery is available across the globe from several major satellite data providers.

Solar energy is the world’s fastest growing form of electricity generation, and all major markets need to prepare for a future in which much of our energy will come from solar.

Energy market operators need to forecast solar generation to ensure that electricity supply matches demand at all times.

Energy network operators need to understand where all the solar is (and where it will be) – both for planning purposes and to manage short-term voltage fluctuations in the grid.

Energy generators (including solar farms) and retailers who trade in energy markets need to know in advance if there will be solar energy shortfall or oversupply, so that they can adjust their bidding accordingly.

Entities who manage fleets of solar + storage assets, such as virtual power plants, need solar forecasts to optimally charge and discharge batteries.

We have an extensive background in providing state-of-the-art solutions to the energy sector.

Every region is different and may have different needs and incentives when it comes to analysing and understanding solar generation or other large sources of data (such as smart meter data).

Our PhD-qualified team can assist with applying state-of-the-art machine learning and artificial intelligence methods to industry problems to help our clients reduce their costs and increase revenue.