Robots & Renewables: What AI Means for the Future of Sustainable Energy (Toronto Climate Week)

Written By Marc Goldgrub

Want to keep up with environmental news, law, and policy? Sign up for the Green Economy Law Monthly Newsletter here.

For Toronto Climate Week 2026, our firm hosted and co-presented (with Avir) this in-depth panel discussion about how AI energy demand is impacting sustainability startups and North America’s future energy mix.

Panelists included:

​The panel was moderated by Green Economy Law’s founding lawyer, Marc Z. Goldgrub. It was filmed by Sofie Mikhaylova.

Key Takeaways

Below, we’ve summarized some of the key takeaways of the panelists’ discussion. Points are organized by subject.

1. AI, energy demand, and climate risk

  • Current and projected AI electricity use is modest relative to total human-caused emissions (on the order of a low single-digit percentage by 2030), with much bigger climate impacts coming from sectors like cement, steel, and industrial agriculture.

  • The main near-term problems from AI/data centers are local (grid stress, water use, siting challenges, Virginia being a standout example) rather than global emissions dominance.

  • There is a big technical and ethical question about what kinds of workloads justify large compute (25,000 GPUs serving billions versus a couple GPUs optimizing an individual building’s efficiency), and panelists argue for making these tradeoffs more transparent.

2. Policy levers and grid modernization

  • A key proposed lever in Canada and the U.S. is a declining carbon-intensity standard for compute: per unit of deployed capacity, emissions must fall over time, while leaving it to industry whether to use renewables, storage, CCUS, nuclear, geothermal, or hybrids.

  • The interconnection queue is a major bottleneck; long timelines to connect new generation push data centers toward off‑grid gas (fuel cells and pipelines) rather than grid-tied renewables.

  • Making it faster and easier to develop off‑grid solar, wind, batteries, and geothermal is seen as critical, alongside broader smart‑grid upgrades and microgrid capabilities.

3. Ontario, startups, and industrial strategy

  • Ontario is described as having a unique combination of AI talent, relatively clean electricity, resources, and an automotive/financial hub, but is losing out because of slow and risk-averse permitting and procurement.

  • Streamlined, province-wide permitting for renewables, storage, and grid‑scale projects (and things like rooftop solar) is a recurring theme, with criticism of current municipal bottlenecks and archaic processes.

  • Government “first customer” procurement, pre‑permitted high‑load industrial sites, and a cultural shift toward saying yes by default are presented as ways to revive climate/AI startup activity in Ontario.

4. Data, microgrids, and privacy

  • Effective AI-based grid optimization (especially reinforcement learning for microgrids) requires granular, local energy data; current regimes in Europe and Canada make access difficult due to strict interpretations of data protection.

  • A proposed compromise is to decouple “energy data” from personal data in regulation, so that microgrid and optimization projects can proceed with lower political friction while still respecting privacy for identifiable residential information.

5. Geopolitics, China, and new energy systems

  • The Iran War and Middle East volatility are expected to raise the risk premium on Gulf hydrocarbons, pushing countries to diversify both into renewables and into “friendlier” fossil exporters like Canada.

  • China is deliberately positioning itself as the industrial core of the clean-energy system (solar, batteries, etc.), much as Britain and the U.S. leveraged coal and oil in earlier eras; this is seen as a conscious power strategy, not an accident.

  • For North America, this implies both more renewables (to reduce dependence on Chinese supply chains) and continued fossil exports (with Canada potentially marketed as a more politically palatable supplier), plus a strong argument that liberal democracies have a moral obligation to lead in clean-tech production.

6. Social impacts, youth, and human agency

  • Younger generations are experiencing significant climate anxiety, including around AI’s energy footprint; one response is to channel that anxiety into climate- and energy‑focused careers and startups.

  • Panelists stress that AI is a tool and citizens still choose the “why”: society can decide to use AI and robotics to free time for care work, creativity, and community rather than purely for profit and surveillance.

  • There is deep concern about concentrated power among a small group of largely unaccountable tech and political elites; one answer offered is building strong “subcultures” and ecosystems (like climate‑AI alliances) that model alternative norms even if they never fully dominate.

7. Robotics and longer‑term opportunities

  • Robotics is seen as being at a “cars in 1895” moment, with general‑purpose and embodied AI expected to become normal within a decade, especially in care work, home assistance, and potentially environmental remediation.

  • Specific long‑term climate applications include forest management, wildfire prevention, and large‑scale reforestation in northern regions where human labor alone is impractical, as well as deploying robots for environmental cleanup.

  • Geothermal—especially advanced and enhanced systems developed by re‑skilled oil and gas workers—is highlighted as a promising “clean firm” baseload option that could scale rapidly if hitched to AI server demand, while also easing political tensions in fossil‑fuel regions like Alberta.

Marc Goldgrub
Next

The State of Canadian Climate Law: 2026 report