Illustration: Tiffany Herring/Axios
A growing list of startups is using advances in generative AI, the tech behind ChatGPT, to develop tools to mitigate or adapt to climate change and grow the use of climate technologies.
Why it matters: The changing climate is one of the most complex global problems, and high-powered computing and smart algorithms could do things like help discover new energy materials or unlock young climate tech markets.
Details: Startup Verse, which uses generative AI to help companies buy clean power, recently unveiled its product, along with a seed round of $5.75 million in funding, led by Coatue and with participation from Twine Ventures, MCJ Collective and others.
- San Francisco-based Verse trains models with data about how power purchase agreements work — like PDFs or articles on the subject — and has developed a tool that it says makes the process of buying clean power cheaper, faster and easier.
State of play: Verse is just one example, focused on the specific issue of the complexity of clean power PPAs.
- Vancouver-based startup Good Chemistry uses advanced computing and generative AI to enable researchers to discover and develop new materials.
- Open ESG, a startup, based in Palo Alto, Calif., uses generative AI to help consumers understand ESG details of companies and products.
- Amazon's cloud computing arm, AWS, and an AI research group recently launched a new program that will support startups using AI and advanced computing to combat climate change.
Big picture: The applications from generative AI for climate are just emerging following advances in OpenAI's ChatGPT and others working on generative AI models.
- AWS's head of startup business development, Kathryn Van Nuys, says, "We're seeing a whole new host of use cases and technology around generative AI and climate tech and we're going to see some exciting solutions and big ideas."
Yes, but: While generative AI could aid in the complex issue of climate change, models like ChatGPT also use large quantities of energy to both train and support user queries.
- Companies will need to be able to make training and operating generative AI models as energy efficient as possible to reduce energy use.
