What challenges did Delhi Metro Rail Corporation face before AI implementation?

Critical maintenance data was trapped in separate systems. Teams lacked a single real-time view of asset health, making it difficult to make fast, informed decisions and increasing the risk of service disruptions and delays. Maintenance was largely reactive and relied on manual monitoring.

What was the impact of the AI solution?

DMRC is targeting a 10-20% reduction in downtime. The organization reports projected 15% time savings in maintenance planning. It achieved 100% faster access to operational information and a single source of truth for decision-making.

Source

GCPProductionEvidence: Medium65/100

DMRC increases predictive maintenance accuracy using Gemini in Vertex AI (plus BigQuery/Dataflow)

Use case typePredictive maintenanceUpdated Jul 19, 2026

Delhi Metro Rail Corporation (DMRC) built a Data Center of Excellence on Google Cloud to move from reactive maintenance to predictive maintenance across its metro network. The solution uses Vertex AI and Gemini in Vertex AI to identify early fault patterns and interpret depot maintenance logs, while BigQuery and Dataflow unify data and support real-time operational analytics. DMRC monitors 300+ standardized KPIs and aims to reduce downtime for passengers while improving maintenance planning and decision-making.

Organization: Delhi Metro Rail Corporation
Industry: Public Sector
Location: India
Published: July 2026

Reported outcomes

10-20%

downtime reduction targetRisk, reliability & safety

−15%maintenance planning time savings+100%operational information access speed

Strategic outcomes

Cost efficiencyShifted from reactive to predictive maintenanceBetter decisions & insightCreated a single source of truth for operationsScale & capacityBuilt a reusable modular cloud architecture

Why do we believe this?Outcome claims, sources, and evidence checks

Normalized claim

Downtime reduction target: 10-20% decrease

Google Cloud Customer StoriesJul 19, 2026Customer storyExplicit claimMedium evidence strength

target a 10-20% reduction in downtime

Normalized claim

Maintenance planning time savings: 15% decrease

Google Cloud Customer StoriesJul 19, 2026Customer storyExplicit claimMedium evidence strength

projected ~15% time savings in maintenance planning

Normalized claim

Operational information access speed: 100% increase

Google Cloud Customer StoriesJul 19, 2026Customer storyExplicit claimMedium evidence strength

DMRC has achieved 100% faster access to operational information

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Delhi Metro Rail Corporation
Provider: GCP
Maturity: Production
Linked source: Google Cloud Customer Stories

The solution uses Vertex AI and Gemini in Vertex AI to identify early fault patterns and interpret depot maintenance logs, while BigQuery and Dataflow unify data and support real-time operational analytics

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 3

1Predictive maintenance
2Operational analytics
3Decision support

Critical maintenance data was trapped in separate systems.
Teams lacked a single real-time view of asset health, making it difficult to make fast, informed decisions and increasing the risk of service disruptions and delays.
Maintenance was largely reactive and relied on manual monitoring.

DMRC partnered with Deloitte to build a Data Center of Excellence on Google Cloud.
The platform uses Vertex AI to identify early fault patterns across rolling stock, signalling gears, traction systems, lifts, and escalators.
Gemini in Vertex AI helps decipher complex depot maintenance logs and assign the correct sub-systems for repair.
BigQuery, Dataflow, GKE, Cloud SQL, Cloud Run, and Cloud Storage support unified analytics, dashboards, and processing.

Technologies

Vertex AI Gemini BigQuery Dataflow Cloud SQL Cloud Run Google Kubernetes Engine Cloud Storage

DMRC is targeting a 10-20% reduction in downtime.
The organization reports projected 15% time savings in maintenance planning.
It achieved 100% faster access to operational information and a single source of truth for decision-making.

Architecture

DMRC and Deloitte built a Google Cloud Data Center of Excellence that ingests high-volume asset data, uses Vertex AI and Gemini in Vertex AI for predictive maintenance and log interpretation, and uses Dataflow, BigQuery, GKE, Cloud SQL, Cloud Run, and Cloud Storage to support real-time analytics, dashboards, and transactional processing.

Sources & evidence1

Evidence: Medium65/100Evidence strength