Maya Pindeus, founder of Another Earth, argues that without AI oversight, the UK electricity grid is at high risk of blackouts because of plug-in solar.
The UK power grid risks future blackouts due to unregistered plug-in solar systems unless advanced monitoring technology is deployed, according to Maya Pindeus, founder of synthetic Earth observation platform Another Earth. Upcoming regulatory changes taking effect on Aug. 27 will permit UK residents to self-install plug-in solar arrays up to 800 watts without using certified technicians. Under the G98 framework, residents are expected to report their systems to network distributors within 28 days after connection, replacing the previous standard where licensed contractors submitted the documentation.
Recent solar output fluctuations during a total eclipse were easily absorbed by the grid because operators had over a year to prepare and centralized utilities continue to supply most of the nation's solar power. However, self-installed systems shifting the reporting burden onto consumers could create significant blind spots, mirroring challenges currently experienced in international markets like Brazil.
In Brazil, over 3.6 million rooftop solar units are connected, but up to 14 gigawatts of generation capacity—comparable to the output of the country's largest hydroelectric plant—remains unmapped because owners expanded systems or added inverters without informing local distributors. As a consequence, grid managers are forced to curtail utility-scale solar farms to manage local oversupply, causing Brazilian utility solar curtailment to double year-over-year to 25% in June 2026 and leading developers to freeze billions in planned investments.
Grid operators worldwide are turning to artificial intelligence and satellite imagery to identify unregistered installations by measuring roof surface areas. However, models trained on mismatched regional data frequently misidentify hardware. In Brazil, AI systems trained on European rooftop data routinely mistake traditional solar thermal water heaters for photovoltaic modules, generating inaccurate capacity counts due to variations in architectural styles and materials.
Pindeus notes that applying borrowed satellite data to British housing stock would produce similar errors due to unique local weather, building densities, and roofing materials. Generating location-specific synthetic satellite data using AI allows algorithms to be trained accurately on regional roof conditions without requiring years of manual data collection, enabling operators to maintain real-time visibility as self-installed consumer solar adoption expands.
Originally reported by Solar Power Portal on Aug 13, 2026.
