Delhi’s 2.6 lakh CCTV cameras can record nearly everything — but experts say only a fraction can detect threats in real time.
Brajesh Mishra
After the November 10 explosion near Delhi’s Red Fort, police traced a white Hyundai i20 through more than a thousand CCTV clips across Delhi and Faridabad. Investigators said the process relied on manual, frame-by-frame review of recorded footage. The episode has renewed debate about whether India’s surveillance infrastructure — extensive in reach but limited in automation — is designed more for reconstruction than prevention.
Delhi’s CCTV ecosystem is among the largest in the world, but government audits show its efficiency depends on maintenance and integration. Official and industry reports estimate between 2.1 and 2.6 lakh installed cameras, though a July 2025 audit found 32,000 non-functional and 15,000 never installed units.
Feeds come from a patchwork of agencies: Delhi Police, municipal bodies, market associations, and private installations. Data retention varies — as little as three to four days in hospital systems, and longer in police-run servers when capacity allows. Integration remains partial; many private or institutional feeds are not connected to central monitoring systems.
The Delhi Police’s Command and Control Centre (C4i) handles real-time monitoring across select zones, focusing on high-value areas like the Red Fort, Connaught Place, and metro hubs. Public reports describe a system designed to handle hundreds of live feeds at a time rather than all cameras simultaneously.
The Safe City Project, under the Ministry of Home Affairs, introduced 3,500 AI-enabled cameras in October 2025 as part of Phase 1 of its ₹798 crore rollout. These systems include video analytics, facial recognition, and abandoned-object detection but represent only a small fraction of Delhi’s overall network.
The Red Fort case underscored a paradox: Delhi’s surveillance footprint is massive, yet its intelligence layer remains thin.
Experts describe this as a design choice rather than a failure — a network meant to record activity, not interpret it. Technologists and civic groups note that true real-time surveillance requires high-quality cameras, stable connectivity, and integrated software capable of pattern recognition — conditions not yet uniform across Delhi’s network.
Independent researchers have argued that deployment should balance technological capability with privacy and accuracy safeguards. Current AI-driven systems remain concentrated in pilot zones, not yet scaled citywide.
The Red Fort investigation revealed how surveillance can excel at backtracking events, but not always at anticipating them. While the Safe City Mission’s AI rollout marks progress, the proportion of “intelligent” cameras — roughly 3,500 out of more than two lakh — remains minimal.
For Delhi, the challenge ahead isn’t just expanding the number of cameras but integrating them under shared standards for data, storage, and live analytics. As policymakers refine the Safe City framework, experts suggest balancing efficiency with transparency and citizen trust.
If thousands of cameras can record a city’s every move but still miss the warning signs, what does “smart” surveillance really mean — more eyes, or sharper ones?
Official and audit reports
Safe City rollout
Operational and investigative coverage
Context and analysis
Exclusive Content
Sign up for the Daily newsletter to get your biggest stories, handpicked for you each day.
Trending Now! in last 24hrs
