Ground-mounted solar panels can both increase our food production and supply power to new data centers.

From: Joshua M. Pearce, holder of the John M. Thompson Chair in Information Technology and Innovation and professor at Western University, Canada The use of artificial intelligence (AI) is exploding. More than 50% of all new internet content was created by AI in 2025, according to an industry report. We are even now training AI on AI-generated content, and although this can impair performance, it continues at a rapid pace. All this AI consumes a lot of energy. It strains the electrical system, increases consumers' electricity bills, and disrupts the planning of large power grids. And the "AI energy crisis" is expanding. The International Energy Agency predicts that global electricity demand from data centers will double by 2030 — to more than Japan's current electricity consumption. At the same time, solar technology — which uses the sun's energy to produce electricity — provides the cheapest energy in the planet's history. The sector is growing rapidly. But both solar and AI projects threaten to take up valuable agricultural land, prompting public protests. A new study I co-authored points to 'agrivoltaics' — the use of land for both electricity production and food cultivation — as a very promising solution. In the first study of its kind, we found that agrivoltaics is a feasible way to meet the growing demand for AI energy in the US while increasing food production. In Canada, agrivoltaics could produce enough electricity to completely eliminate the need for fossil fuels in the power grid — on less than one percent of the country's farmland. Agrivoltaics allows farming communities to generate solar-based electricity while continuing to produce food, sometimes with even higher yields than before. In our study, we examined two types of agrivoltaics: vertical panels and single-axis solar trackers, because both can be integrated into most farms without disturbing farmers. Vertical agrivoltaics are essentially fences made of solar panels. These solar fences are spaced far enough apart so that farmers can operate tractors, harvesters, and other equipment between the rows without hitting them. Single-axis trackers use the same principle — just spaced farther apart when used for agrivoltaics. They follow the sun and thus produce more energy per panel. When crops are grown, they stand upright like fences. Both types of solar agrivoltaics minimally impact the sunlight reaching crops and work well with most types of produce. Multiple studies on a wide range of crops — including basil, broccoli, celery, chili peppers, corn, lettuce, grass, potatoes, spinach, tomatoes, and wheat — have shown that agrivoltaics can increase yields. For example, we demonstrated that strawberry yields in Ontario increased by 18% in a normal year. This is because agrivoltaic solar panels can create a "shade effect," providing a beneficial microclimate where plants are somewhat protected from sun, heat, and wind. This shading effect depends on the weather. For instance, while agrivoltaics generally benefit lettuce, last year's hot summer amplified the shading so much that the fresh weight of lettuce increased by over 400% compared to unshaded control plants and over 200% compared to the national average yield. In our study, we used data on data center energy consumption at the state level and modeled the production potential of agrivoltaics. We examined what portion of the digital sector's demand could realistically be met with agrivoltaics. We also assessed how much farmland would require solar investments to cover AI energy needs in the US states with the largest data centers. Our results showed that vertical agrivoltaics would require only between 0.003% and 2% of farmland across the selected states. That's almost nothing. Single-axis trackers require even less — between 0.001% and 0.548%. The AI energy crisis in the US could be mitigated by installing single-axis trackers on at most 0.5% of land in the less agricultural-rich states. Canada is even more advantaged — with less than 1% of farmland, the country could produce enough electricity to eliminate the need for fossil fuels. This would supply energy for everything — not just AI. Agrivoltaics preserves jobs in agriculture, increases food supply, and significantly boosts farm income due to the high value of solar-generated electricity. It provides a dual income stream: one from selling farm products and another from selling electricity or covering the farm's own energy needs. Unsurprisingly, agrivoltaics is growing rapidly, with the global market already exceeding $14 billion. Even the Vatican now operates on agrivoltaics. In some jurisdictions, however, outdated regulations effectively block new agrivoltaic projects. Ontario, Canada, is an example. Agrivoltaics has been a financial success in Ontario when integrated with grazing sheep and goats to keep vegetation down on conventional solar farms. Unfortunately, this is the only widespread form of agrivoltaics in the province due to restrictions on large-scale solar energy on farmland. To remove this barrier to job creation, food security, and economic development, Ontario could update its regulations to exempt agrivoltaics from current restrictions. This would attract significant investments and enable crop-based agrivoltaics. Specifically, Ontario's government could include agrivoltaics as a "farming-related use" in the provincial policy statement to bypass restrictions on "diversified land use." This way, we could all produce more food and more solar energy to meet the rising demand. The original article was published in English on The Conversation on June 17.