Indian companies are among the most advanced in watching their AI systems, yet they pay the steepest price when things break. That is the central finding of New Relic’s 2026 Observability Forecast: India Edition, which puts the average cost of a single high-impact outage at $3 million.

Bengaluru, November 4, 2026: The report draws on a global survey of 2,575 engineering and IT leaders, of whom 500 are in India, the largest country group in the study. New Relic, an observability platform vendor, describes the result as a paradox: India leads on advanced OpenTelemetry standardisation and AI agent monitoring, but suffers the heaviest operational and financial strain.

The cost figures are stark. New Relic puts the hourly cost of downtime for Indian organisations at $2.43 million, which works out to $40,500 for every minute systems stay offline. Severe outages are no longer rare: 61% of Indian organisations report high-impact outages weekly or more often, and 31% report them multiple times a day, which the company says is a tenfold increase from 2025.

Teams are also slower to catch and fix problems. The average time to detect a high-impact outage has stretched to 55 minutes from 41 a year ago, and the mean time to resolve such incidents has risen to more than an hour. Third-party cloud failures, software updates and network disruptions are the main triggers. Security failures stand out: 32% of Indian organisations cite them as a frequent cause of downtime, nine percentage points above the global average and, in New Relic’s words, the market’s widest operational gap. Developers bear the brunt. Half of their time in India is now lost to dealing with system disruptions, up from 35% last year and nearly double the global median of 30%.

On tooling, the picture is more positive. According to the survey, 56% of Indian enterprises actively monitor AI agents in production, against a global average of 35%. India also leads in AIOps adoption at 54% versus 36% globally, and in machine learning model monitoring at 57% versus 38%. On OpenTelemetry, an open standard for collecting telemetry data, 42% of Indian respondents are fully standardised, nearly double the 22% global average.

That tooling shows up in returns. The share of Indian enterprises realising a 3-5x return on observability spending nearly doubled from 19% to 35% in a year, and the share seeing no financial return or not tracking it fell from 12% to 6%. Globally, New Relic says, teams that monitor production AI agents are more than twice as likely to achieve a 3x or better return than those that do not. Indian chief information officers appear to be less price-sensitive as a result: 39% prioritise AI and machine-learning features in vendors, ahead of cost at 25%.

The weak spot is fragmentation. Indian organisations use an average of six monitoring tools, which the report says leaves standardised data trapped in silos. Only 20% of Indian leaders say AI-assisted observability has significantly improved their event correlation, behind the 27% global average.

“India’s digital economy is scaling at an extraordinary pace, but this high growth has outpaced operational discipline, leaving engineering teams to pay a heavy reactive troubleshooting tax,” said Rob Newell, New Relic’s senior vice president and general manager for Asia-Pacific and Japan. He added that “tool sprawl is still capping the upside” and urged businesses to move from “disconnected point tools to a unified platform.”