Autonomous Vehicle and ADAS Liability Insurance for Indian OEMs and Fleet Operators

Indian passenger vehicle OEMs have accelerated ADAS deployment from 2024 onwards, with Mahindra XUV 700 and XEV 9e, Tata Harrier and Safari, Hyundai Creta and Tucson, Kia Carens Clavis, MG Astor and Gloster, and Honda Elevate and City all offering Level 2 or Level 2+ ADAS suites. Imported Tesla Model Y and Model 3 units operating on Indian roads under the personal import route add a smaller but visible population of vehicles with more capable autonomy stacks. Typical Level 2 features sold in India include adaptive cruise control, lane-keeping assist, automatic emergency braking, blind-spot detection, traffic sign recognition, and some degree of driver monitoring.

Fleet-facing pilots are further along the autonomy curve than most retail buyers realise. Blue Dart has been running geofenced yard automation trials at Bhiwandi and Luhari. Ola and Uber are evaluating Level 3 robotaxi pilots in Bengaluru tech corridors in partnership with Minus Zero and Swaayatt Robots. Delhivery, Ecom Express, and Shadowfax are testing teleoperated and semi-automated delivery vehicles in Gurugram and Hyderabad. On the truck side, Ashok Leyland and Tata Motors are evaluating platooning for long-haul freight on the Delhi-Mumbai Expressway.

The insurance market has not kept pace. IRDAI has issued no specific circular or product filing guidance for autonomous vehicle liability. The Motor Vehicles Act, 1988, continues to govern third-party liability, but its drafters did not contemplate causation disputes between a human driver, a software stack, and a sensor manufacturer. The result is a fragmented coverage stack where OEMs, fleet operators, and tier-1 suppliers each carry partial coverage that does not always respond when a Level 2 feature misfires in monsoon fog or a Level 3 pilot hands back control to a distracted safety operator.

For OEMs the primary exposure is product liability. If a lane-keeping assist steers a Mahindra XUV 700 into the opposing lane because the lane marking was washed out by pre-monsoon rain near Lonavala, and a head-on collision results, the claimant will sue the OEM for a defect in the ADAS feature, the component supplier (Mobileye, Bosch, Continental, Aptiv, ZF) for a defect in the perception or control module, and the policyholder's motor insurer for third-party compensation. Each party will attempt to push liability to another.

Standard product liability policies written in India for auto OEMs carry limits of INR 50 to 200 crore. This is adequate for mechanical defects such as a brake recall or a fuel pump failure, where a single defect population is finite and per-claim severity is bounded. ADAS exposure is different. A software calibration defect can affect tens of thousands of vehicles simultaneously and generate class-style litigation at the National Consumer Disputes Redressal Commission or via writ petitions. OEMs deploying ADAS at scale should underwrite product liability limits of INR 200 to 1,000 crore for Level 2 platforms, and INR 1,000 crore plus for any Level 3 deployment. Reinsurance support via Munich Re, Swiss Re, Lloyd's syndicates, and Scor is available but requires detailed actuarial submissions that include false-positive and false-negative rates for each ADAS feature.

The policy wording must explicitly cover software OTA defects. Most Indian product liability wordings cover manufacturing and design defects but are silent on software. OEMs should negotiate a specific over-the-air update endorsement that covers liability for defects introduced or not corrected through an OTA campaign. Tesla, Tata, and Mahindra all push OTAs that change ADAS behaviour; an OTA that tightens or loosens automatic emergency braking thresholds can create a new defect population overnight.

Sensor failure in Indian conditions is a separate underwriting topic. Monsoon water ingress into radar housings, fog on the Golden Quadrilateral, dust ingress in Rajasthan, and extreme heat deformation of LIDAR mounting brackets in Nagpur summers all degrade ADAS performance. Underwriters typically require proof of IP67 or IP6K9K certification on exposed sensors, validated environmental testing at ARAI or ICAT, and documented operational design domain boundaries that switch the feature off when conditions exceed its design envelope. An ADAS that silently degrades without warning the driver is a far larger exposure than one that disengages with a clear chime.

The Motor Vehicles Act, 1988, section 146, continues to require compulsory third-party motor insurance for every vehicle plying on Indian roads, including those with ADAS or autonomous features. The Act and Tribunals attribute liability to the vehicle owner and driver, not to the OEM or software supplier. Where the driver is not driving in the ordinary sense because ADAS was engaged, causation becomes contested.

The practical pattern in 2025 and 2026 claims has been that Motor Accident Claims Tribunals (MACT) first award third-party compensation against the motor insurer under no-fault and fault-based heads, typically between INR 15 lakh and INR 2 crore depending on earning capacity and age of the deceased or injured. The motor insurer then subrogates against the OEM where evidence suggests ADAS contributed. Early subrogation matters in 2025 involving Tesla imports and MG Astor have been settled out of court, so case law is still thin. A Telangana MACT order in October 2025 involving a Level 2 cruise control incident signalled that Tribunals will entertain OEM-directed subrogation where the vehicle event data recorder shows the feature was active at impact.

Fleet operators should require their motor policies to include an express waiver of subrogation in favour of the OEM for ADAS-related incidents only where the OEM has indemnified the fleet under a manufacturer's warranty extension. Otherwise, fleets preserve the ability to recover. Fleet motor policies for ADAS-enabled vehicles should also carry a software-defect endorsement that keeps motor cover alive even if the OEM initially disputes liability and the claim is slow to settle.

For individual owners of ADAS-equipped vehicles, existing private car motor policies from ICICI Lombard, HDFC ERGO, Tata AIG, Bajaj Allianz, and Go Digit continue to respond to third-party and own-damage claims without feature-specific exclusions, as of Q1 2026. This is a market convention that could change if loss ratios deteriorate. Owners should check renewal terms carefully for any new exclusion that ties coverage to human control of the vehicle at the moment of the accident.

Fleet operators piloting Level 3 and Level 4 automation face a more complex apportionment problem than OEMs. Consider a Blue Dart yard tractor operating autonomously inside a Bhiwandi hub that collides with a parked trailer because the perception stack failed to identify a low-profile chock. The possible liable parties include the fleet (operational negligence, insufficient geofencing, failure to supervise), the OEM or integrator (defective autonomy stack), the sensor supplier (hardware fault), the HD map provider (outdated map layer), and the teleoperator or remote safety driver (if one was on duty).

Contractual apportionment is the first line of defence. Fleets commissioning autonomous pilots should negotiate liability allocation in the master services agreement before any vehicle is put into service. A typical allocation places product defects on the OEM or integrator with a contractual cap of two times annual service fees, operational negligence on the fleet, and data or cyber failures on the technology provider. Residual risk sits with the fleet's liability insurer.

The insurance programme for a fleet operator piloting autonomy should include: motor fleet cover with a software-defect endorsement for the full vehicle population, commercial general liability at INR 25 to 100 crore to respond to bodily injury and property damage outside the strict motor ambit, cyber insurance at INR 10 to 50 crore to cover ransomware and OT intrusion into vehicle control systems, and employer's liability and workmen's compensation for safety drivers and teleoperators. Directors and officers coverage is relevant because autonomous pilots are high-visibility decisions that can invite regulatory inquiry if an incident occurs.

Apportionment between operator and manufacturer is best handled through a joint claims protocol agreed with the OEM at pilot inception. The protocol should specify that the OEM's product liability insurer is the first responder for any claim where the event data recorder shows the autonomy stack was engaged and functioning nominally was not the case, with the fleet's motor and liability insurers responding to the balance. Without such a protocol, claimants face years of inter-insurer litigation and the fleet faces reputational damage even when technically not at fault.

ADAS and autonomous vehicles generate substantial volumes of personal and sensitive data. A single Level 2 vehicle operating on an Indian highway streams driver monitoring camera feeds, GPS traces, cabin audio in some configurations, heart-rate and alertness telemetry where driver monitoring is fitted, and external camera footage that captures third parties including pedestrians, other drivers, and number plates. Event data recorders retain pre-crash telemetry for claims investigation. Fleet vehicles add business-sensitive data such as route optimisation, customer drop-off locations, and load manifests.

The Digital Personal Data Protection Act, 2023, and its 2025 draft rules apply to most of this data. Driver identity, biometric alertness markers, and cabin audio are personal data. Footage of third parties raises collection and consent questions that no Indian OEM has cleanly solved. Fleets acting as data fiduciaries must provide notice, obtain valid consent from drivers, maintain data localisation where applicable, and respond to data principal rights requests within the statutory timelines.

Event data recorder admissibility in Indian courts is still being tested. In motor accident proceedings, Tribunals have begun to accept EDR printouts produced by the OEM's technical team as evidence, subject to section 65B of the Indian Evidence Act certification. OEMs and fleets should have documented chain-of-custody procedures for EDR extraction and a legal hold protocol for vehicles involved in serious incidents. An EDR that is overwritten or tampered with can convert a defensible claim into a conceded one.

Cyber insurance for OEMs and fleets must cover DPDP penalties up to INR 250 crore per contravention, data subject compensation claims, forensic investigation costs, and regulatory defence. Standard cyber wordings from Tata AIG, ICICI Lombard, and Bajaj Allianz now include DPDP-specific endorsements filed after the 2023 Act, but limits below INR 25 crore are usually inadequate for an OEM or a fleet with more than 1,000 connected vehicles. A cyber policy should also respond to loss arising from OT intrusion, meaning a cyber event that reaches the vehicle control network, because such an event can simultaneously trigger motor third-party claims, product liability claims, and regulatory action.

Four Indian regulators hold pieces of the autonomous vehicle insurance puzzle, and none has yet produced a unified framework. The Ministry of Road Transport and Highways (MoRTH) released a draft concept note on autonomous vehicle testing in late 2024 and a follow-up consultation in Q3 2025 that proposed a sandboxed testing regime with state-level approvals, minimum insurance limits for test fleets, and mandatory EDR fitment. The final rules are expected by late 2026.

The Bureau of Indian Standards has issued IS 17852 for lane departure warning and IS 18091 for automatic emergency braking, aligning broadly with UN-ECE Regulations 79 and 152. Conformity with these standards is voluntary but insurers increasingly require it as a condition for premium discount on ADAS-equipped fleets. The Automotive Industry Standards Committee is working on a draft AIS standard specifically for Level 3 handover protocols, expected in 2026 or 2027.

The Directorate General of Civil Aviation is relevant only for airport apron vehicles and airside autonomous tractors at airports such as Delhi IGIA and Mumbai CSMIA. Fleets operating airside must comply with DGCA airside vehicle permit conditions in addition to state transport approvals.

IRDAI has been notably silent on autonomous vehicles. No product filing exists specifically for AV liability, no circular has addressed causation apportionment, and no minimum limit has been prescribed beyond the standard motor third-party regime. The insurance industry is proceeding via bespoke manuscript wordings filed under the file and use regime for commercial lines. Senior underwriters at General Insurance Council meetings through 2025 have called for IRDAI to issue framework guidance before a high-severity Level 3 incident forces reactive rulemaking. OEMs and fleets should track the IRDAI Insurance Advisory Committee agenda and the expected 2026 exposure draft on emerging mobility risks.

Other jurisdictions offer partial templates. In the United States, the National Highway Traffic Safety Administration has used its recall authority to force OTA fixes and has treated ADAS misuse and abuse as a driver issue. State-level liability follows product liability and comparative negligence doctrines. California, Arizona, and Texas have developed specific regulatory regimes for AV testing with minimum liability coverage of USD 5 million per vehicle.

The European Union's General Safety Regulation mandated ADAS features on all new vehicle type approvals from July 2024. The Product Liability Directive revision of 2024 explicitly extended strict product liability to software and to defects introduced through updates, closing the OTA gap that Indian law still contains. The UK's Automated and Electric Vehicles Act 2018 created a first-instance liability on the motor insurer for automated driving incidents, with rights of recovery against the manufacturer.

China's regulatory regime under the 2022 Intelligent Connected Vehicle Road Testing and Demonstration Application Management Specification places primary liability on the AV operator during testing, with recovery against the OEM for proven product defects. China has also mandated data localisation for AV-generated data, a model that DPDP rules may eventually echo.

For Indian policy makers and underwriters, the UK first-instance model is administratively simple but requires a clear statutory amendment to the Motor Vehicles Act. The EU OTA liability extension is a natural fit for Indian product liability reform. The US comparative negligence approach maps well to the existing MACT jurisprudence but leaves causation disputes unresolved. The practical market approach in India until legislation catches up is to contractually allocate risk via master services agreements between OEMs and fleets, and to underwrite manuscript policies that name all likely defendants as additional insureds. This is not a durable long-term solution but it allows commercial deployment to proceed while the regulatory framework matures.

For a passenger OEM selling Level 2 ADAS across 200,000 to 500,000 vehicles annually, the baseline programme should include product liability at INR 500 to 1,000 crore limit structured as INR 50 crore primary with excess layers placed with international reinsurers, an OTA software defect endorsement, a recall expense cover at INR 100 to 300 crore, cyber insurance at INR 50 to 100 crore, and directors and officers at INR 100 to 250 crore. Annual premium outlay for such a programme is in the range of INR 40 to 100 crore, with premium sensitive to claims history, ADAS feature mix, and the proportion of ADAS-equipped units in the annual production volume.

For a fleet operator running 5,000 to 20,000 ADAS-enabled vehicles with a small Level 3 pilot, the programme should include motor fleet cover with software-defect endorsement, commercial general liability at INR 25 to 100 crore, cyber insurance at INR 25 to 50 crore, employer's liability, and a contingent product liability difference-in-conditions layer that responds if the OEM's policy is exhausted or denies the claim. Annual premium outlay is typically INR 15 to 40 crore for a fleet of this size.

Key benchmarks for underwriters and risk managers. Motor fleet loss ratios on ADAS-equipped vehicles have trended 8 to 12 percent lower than non-ADAS fleets based on 2024 and 2025 data from Bajaj Allianz and Tata AIG internal books, driven by reduction in rear-end collisions and single-vehicle run-off-road incidents. This discount is not yet fully passed to customers but is emerging in fleet renewals. Product liability severity on ADAS claims remains uncertain because claim development tails are long, so reserves should be set conservatively.

Practical steps to reduce total cost of risk. First, mandate EDR fitment and maintain documented chain-of-custody procedures. Second, require OTA update logs to be retained for the lifetime of the vehicle plus three years, matching the likely limitation period for defect claims. Third, negotiate OTA liability endorsements at every renewal rather than accepting the standard silent wording. Fourth, run joint tabletop exercises with OEMs, tier-1 suppliers, and insurers annually to test incident response and subrogation protocols. Fifth, maintain a live register of ODD boundaries for each ADAS feature deployed, with evidence that drivers are trained and warned when conditions exceed the envelope.