A new frontier: How artificial intelligence is reshaping cleantech innovation

The advances of large language models such as those pioneered by OpenAI’s ChatGPT have made it hard to read much of anything these days without hearing about the vast potential—and extraordinary risks—posed by artificial intelligence (AI). So, why have we chosen to add to the fray? AI presents a truly transformative opportunity to shape our fight against climate change.

Advances in these technologies change how we create, distribute and consume energy. From grid optimization and scientific discovery to climate modelling and predicting extreme weather events, cleantech innovators are integrating various aspects of AI into their products, services and processes. This has led to a surge of investment in AI-enabled cleantech, which reached US$28.5 billion between 2018 and 2023. Going out to 2029, the amount of capital deployed in the space is expected to surge five-fold, with close to US$140 billion in investment expected.

This new emerging field—at the nexus of digital and physical technologies—has applications that cut across industries and sectors. Many of these solutions integrate AI software into cleantech components or leverage AI to enhance the functionality of cleantech hardware. As a result, AI-enabled cleantech solutions range from upstream uses, such as those with large algorithm sophistication, to novel solutions that capture new forms of data. By some estimates, AI-enabled cleantech could help mitigate up to 10% of global greenhouse gas (GHG) emissions by 2030.

But as AI is incorporated into cleantech, it’s important to keep in mind the significant environmental implications of running these models. The data centres needed to run large language models and other AI applications with advanced computational power require substantial amounts of energy and water. To put things in perspective, a standard search on an AI chatbot with natural language processing is said to consume 10 times the electricity of a simple Google search. This is further complicated by shortages of renewable energy on current grids, which make switching to clean energy sources more difficult for big data companies. For their part, technology companies have begun investing in creating renewable energy capacity, especially nuclear power, leveraging small modular reactors (SMRs).

For now, though, AI applications in cleantech tend to focus more on small-language models, as opposed to the large-language models powering the more energy-intensive generative AI applications. As a result, cleantech’s low compute power requirements and relatively lower energy draw mean that it has a smaller carbon footprint. As the scope of AI in cleantech evolves, uses could even lend themselves to solving their own energy challenges; monitoring, predicting and reducing emissions while simultaneously helping decarbonize more efficiently. With use-cases that are wide-reaching and vast, AI-enabled cleantech could, in time, generate net positive carbon emission impacts.

While venture capital activity slowed across a number of sectors, broader cleantech investment remained steady in Canada last year, reaching C$1.2 billion, across 73 deals. At the same time, overall cleantech exports continued to trend upwards, with Canada exporting nearly $21 billion in environmental and cleantech products and services in 2022 (the last year for which data are available), up 17% from 2021. As AI-enabled cleantech applications proliferate, Canadian activity in this area should be boosted by the tailwinds.

The bottom line?

Canada has long been touted as a leader in cleantech innovation and a trailblazer in AI, thanks to the presence of leading research institutions and universities. The development of novel technologies such as carbon capture, utilization and storage (CCUS), long-duration energy storage (LDES) and hydrogen will complement the growth of clean energy solutions. Canada has several strengths in these technologies, which will help secure our progress toward ongoing AI-enabled cleantech advances.

Building these ecosystems together will require improvements in the provision and access to AI infrastructure, alongside affordable access to the computing power needed to run these AI models. It will also require access to the risk capital necessary to scale solutions, necessitating a collaborative approach between government, business, and universities. Ultimately, though, the more fundamental the problem that these technologies help solve, the more exportable the product and service will be.

This week, a very special thanks to Prerna Sharma, senior economist in our Economic & Political Intelligence Centre. As always, at EDC Economics, we value your feedback. If you have ideas for topics that you’d like us to explore, please email us at economics@edc.ca and we’ll do our best to cover them.