Maharashtra State Electricity Distribution Company Ltd (MSEDCL) said it had cracked down on power thieves across the state by leveraging Artificial Intelligence (AI) ‘Super Intelligence’. 

Combining state-wide physical inspections by 190 “flying squads” with AI-driven predictive analysis, MSEDCL said it had exposed 27,964 instances of power theft worth ₹66.28 crore in three months. 

Criminal cases have been registered in 289 instances where offenders failed to pay the assessed electricity charges and compounding fees.

MSEDCL said it had adopted dedicated AI and Machine Learning (ML) module to predict electricity theft accurately. Consumers exhibiting suspicious consumption patterns are automatically flagged under a ‘high-risk’ category, providing field teams with precise target locations without requiring prior manual surveys.

Lokesh Chandra,(IAS) Chairman & Managing Director, MSEDCL in a statement said, “Our target is to reduce MSEDCL’s Aggregate Technical and Commercial (AT&C) losses to below 10 percent over the next three years.”

“Prioritising the recovery of outstanding dues, an aggressive campaign against power theft is underway. The newly developed AI module has greatly intensified our precision and enforcement capabilities against electricity theft across the state,” he stated. 

To execute ground-level actions based on AI intelligence, MSEDCL has deployed over 190 specialized “flying squads” across the state. 

During the three-months from June to August, 2026 joint operations by the O&M and Security & Enforcement divisions exposed power thefts across all major regions: Konkan Region (11,083 cases), Pune Region (5,849 cases), Nagpur Region (5,949 cases), and Chhatrapati Sambhajinagar Region (5,083 cases), the company said.

Out of the total 27,964 cases detected, formal assessment bills amounting to ₹66.28 crore were issued for 24,508 verified theft cases. 

So far, MSEDCL has recovered ₹29.85 crore from 10,810 settled cases. Strict legal action is being initiated against those failing to clear dues within the stipulated period, with 289 criminal cases already registered, it said.