LEAP's climate tech track meets a show built on megawatts

LEAP 2026 is running a dedicated climate tech track covering renewable energy, smart grids, energy storage and sustainability, aimed at the innovators, industry figures, policymakers and researchers working on the energy transition. Among its stated themes is the integration of artificial intelligence and the internet of things into energy systems, and strategies for cutting environmental impact while meeting rising demand. That last clause is doing a great deal of work, and it is worth dwelling on, because the show it appears at is the reason the demand is rising. LEAP's headline announcements this week have been denominated in megawatts: data centre projects, AI compute capacity and the electricity agreements that make them buildable. Those two things — a track about reducing environmental impact, and an exhibition floor selling the most power-hungry computing workload ever commercialised — are in the same building for a reason, and it is not a contradiction so much as the actual problem statement of the decade. The technical content of the track reflects where the practical work now sits. Smart grids and storage are the two subjects that determine whether renewable generation can carry a load rather than merely offset one. Solar is abundant in Saudi Arabia and cheap to build; it is also unavailable at night, when a data centre's draw does not fall. Bridging that gap is a storage and dispatch problem, and it is the difference between AI compute that is genuinely low-carbon and AI compute that is nominally powered by renewables while gas plants cover the hours nobody photographs. The AI-in-energy strand is the other half of the loop and the more immediately deployable one. Machine learning applied to grid operations — forecasting generation and demand, managing congestion, scheduling maintenance, spotting faults before they cascade — is one of the few applications where the efficiency gain is measurable in the same units as the cost. It also produces the mildly uncomfortable situation in which the technology driving load growth is being sold as part of the answer to it. Both claims can be true at once, but the arithmetic has to be shown rather than asserted. Saudi Arabia's position in that arithmetic is genuinely unusual. It has excellent solar resource, a state utility accustomed to planning at national scale, and an explicit policy of siting compute where power can be built for it — a combination most countries chasing AI infrastructure do not have. It also has extreme summer cooling loads and a grid that has to serve them, which is why Saudi Energy's agreements this week on supplying power to AI data centres matter more to the climate question than most of what was said on the track's stage. The commercial end of the category has been present at the show in its own right. Masdar showed data and AI solutions , LivNSense GreenOps reached the Rocket Fuel top 100 with industrial decarbonisation software, and Velaura AI raised a Series A on the premise that AI computing has to become more energy efficient . The last of those is the most direct acknowledgement that the industry's power consumption is now an investable problem rather than a footnote. A conference track is not a policy and does not commit anyone to anything. What it does mark is that the energy cost of the technology being sold on the floor has become a subject the event schedules time for, rather than one it leaves to the utilities to sort out afterwards. TechScoop is reporting from LEAP 2026 in Riyadh.