Yulu's $93 Million Round Exposes the Insurance Gap in Shared EV Fleets

Yulu closed $93 million on 11 August 2026, made up of $63 million in equity led by GEF Capital Partners and $30 million in debt, at roughly $170 million post-money. The company runs about 50,000 electric two-wheelers today and has said it intends to reach 200,000 within two years, while logging roughly 1.6 million zero-emission miles a week and powering more than 750,000 deliveries a day. It has also launched Yulu Express, a higher-speed electric scooter aimed at longer-haul delivery work, with about 500 units already running in Bengaluru and trials underway in three more cities.

Two details in that round matter more to an insurance buyer than the headline number. The first is the debt component. Yulu's chief executive has said this is expected to be the final equity raise before an eventual listing, with further fleet expansion financed mainly through debt and lease financing. A debt-and-lease-funded fleet means lenders and lessors sitting behind the asset, and lenders ask for named-interest endorsements, agreed valuation bases, and evidence that the cover actually responds. The second is the growth rate. A four-fold fleet increase inside twenty-four months is not a renewal adjustment. Every frequency-driven exposure in the programme, third-party bodily injury above all, grows at the same pace, and the loss record that supports the next renewal is being written during the expansion.

For context, the same week put Yulu's raise at the top of thirteen Indian startup rounds totalling $242.55 million, which tells you how much of the sector's capital is now flowing into physical, moving, insurable assets rather than software.

An Indian commercial motor policy is built around a stable triangle: a registered owner, a vehicle with a fixed usage pattern, and a driver who is either the owner, an employee, or someone driving with the owner's permission and a valid licence. Underwriting, pricing, and claims handling all assume that triangle holds for the policy year.

A vehicles-as-a-service fleet breaks all three corners at once:

• The owner is not the user. Yulu is the registered owner and the party on the certificate of registration, but the person on the vehicle is a customer who unlocked it minutes earlier through an app.
• The user is not employed. The rider is not on payroll, is not a named driver, and in the delivery use case is usually a gig worker engaged by a third-party platform, not by the fleet operator at all.
• Usage is not fixed. A single vehicle can be ridden by six or eight different people in a day, across mixed urban traffic, at different hours, with no continuity of driving behaviour between hires.

The result is that the standard commercial motor policy wording can be technically in force and still leave gaps that only surface at claim stage: who was driving, whether they held a valid licence at the moment of the accident, whether the use at that moment fell within the declared purpose, and whether the operator's own verification process can evidence any of it months later.

Under the Motor Vehicles Act, 1988, third-party liability cover is compulsory and the liability itself is unlimited in amount for death and bodily injury. Compensation is determined by the Motor Accident Claims Tribunals, and awards turn on the deceased or injured person's income, age, and dependants rather than on any policy limit chosen by the buyer. Claims run for years and settle long after the vehicle has been retired from the fleet.

This is the line item that a four-fold fleet expansion multiplies most directly. Two-wheelers in dense Indian city traffic sit at the sharp end of the accident statistics, and a shared fleet puts riders of widely varying experience on identical machines. As the fleet moves from 50,000 to 200,000 vehicles, the operator is not simply buying four times the number of policies. It is accumulating four times the annual crop of long-tail third-party liability claims, each of which stays open across several renewal cycles and shapes the loss ratio the market will see when the company eventually approaches a listing.

Three consequences follow for the programme design:

1. Reserve visibility matters more than premium. The operator needs claim-level data from the insurer, including outstanding reserves on tribunal matters, so that the true cost of a policy year is visible before it hardens into a renewal quote.
2. Recovery rights need protecting. Where an accident involves a defective component, a third-party vehicle at fault, or a rider acting outside the terms of hire, the insurer's subrogation position depends on the operator preserving evidence and vehicle telemetry rather than putting the machine straight back into service.
3. The programme should be structured, not renewed vehicle by vehicle. A fleet at this scale belongs on a declaration basis with agreed additions and deletions, because vehicles will be entering the fleet weekly during the expansion.

In a private-vehicle policy, the compulsory personal accident cover follows the owner-driver. On a shared fleet, the owner never rides. The rider is a customer, and in the delivery segment a gig worker whose engagement sits with a delivery platform.

That creates a three-way question the operator has to answer before an incident, not after: if a rider is injured on a Yulu vehicle while completing a delivery for a platform, whose cover responds? The fleet operator's motor policy addresses third parties and the vehicle. The platform may carry group personal accident cover for its delivery partners, which is the usual arrangement described in our note on group personal accident cover for gig platforms. The rider's own health cover, if any, sits behind both.

Shared-mobility operators generally address this with a rider accident benefit attached to the ride itself, funded per trip or per active vehicle, with a fixed benefit for death and permanent disablement and a smaller medical reimbursement limit. Two design points decide whether it is worth having:

• The trigger. Cover has to attach on ride start and detach on ride end, defined by the same event the billing system uses, so that there is one authoritative record of whether a ride was live at the moment of the accident.
• The interaction with platform cover. Where the rider is also covered by a delivery platform, the operator's benefit should be written to sit alongside rather than duplicate, and the contract with the platform should say which policy is intended to respond first.

There is a commercial reason to carry it beyond risk transfer. Rider trust affects utilisation, and utilisation is what makes a vehicles-as-a-service unit economic model work at all.

On a shared fleet, own-damage cover stops being a per-vehicle purchase and becomes a portfolio decision. Losses are high-frequency and low-severity: kerb impacts, drops, panel and light damage, vandalism, and theft of vehicles or components from parking zones. A conventional own-damage policy with a small excess on every one of 200,000 vehicles turns into an administrative burden that costs more to run than the losses it transfers.

The realistic structures are a large per-event deductible with the frequency layer retained by the operator, or an aggregate deductible where the operator absorbs the first tranche of losses across the year and the insurer responds above it. Either way the operator needs a maintenance and repair network capable of absorbing routine damage internally, which most fleet operators build anyway.

The battery decides the severity tail. On an electric two-wheeler the pack is the single most valuable component, and in a swap model the pack in the vehicle at the time of loss may not even be the pack that vehicle started the day with. Three points have to be settled in the wording:

• Whether the battery is insured with the vehicle or separately. In a swap network the circulating packs are better treated as a distinct insured item with its own basis of valuation, since they move independently of the vehicles.
• How the sum insured is set. A depreciating insured declared value on a fleet the operator intends to keep in service, and on which lenders hold security, sits awkwardly with the replacement cost of the pack. Our note on insured declared value and battery treatment on electric commercial vehicles sets out the mechanics.
• What the fire and thermal exclusions actually say. A pack that fails in service is the loss most likely to be argued over, so the wording should be read for any exclusion covering internal battery defect, charging outside prescribed conditions, or use of non-original components.

Because the next phase of expansion is being financed with debt and lease finance, lenders will want their interest noted, and that in turn forces the valuation basis into the open. It is better to have that conversation at placement than after a claim.

A fleet of this size is supported by a network of swap and charging sites, each of them a small urban footprint holding a number of charged lithium packs. Insurers do not read those sites as retail premises. They read them as an accumulation of stored energy, closer in character to a small battery store, where the question is not whether one cell can fail but how far a failure spreads and what it takes with it. Our detailed treatment of that exposure is in the post on insuring battery-swapping network operators.

For a shared-mobility operator the practical consequences are:

• Per-location limits have to reflect real accumulation. The limit at each site should be set against the number of packs held there at peak, plus the charging equipment, plus the fit-out, not against an average across the network.
• Third-party property is exposed. Most swap points sit inside leased space in mixed-use buildings, retail parking, or basements. A fire that spreads to the landlord's building or neighbouring premises is a public liability claim, and the lease will usually make the operator responsible for it.
• Interruption is a network effect. Losing one node inconveniences riders. Losing a hub that serves a cluster of nodes strands vehicles across a whole service area, so business interruption sizing should reflect the revenue at risk across the affected cluster and an indemnity period long enough to cover pack resupply lead times.

Pull the pieces together and a shared electric two-wheeler operator at Yulu's intended scale needs a programme with six components, placed as one structure rather than assembled piecemeal by department.

1. Fleet motor on a declaration basis, covering compulsory third-party liability plus own damage, with additions and deletions notified monthly and a deductible structure that leaves routine damage with the operator. This is the anchor placement.
2. Rider accident benefit attached to the ride, triggered by the same ride-start and ride-end events the billing system records, and drafted to sit consistently with any delivery platform's group personal accident cover.
3. Property and fire cover across the swap and charging network, with per-location limits set on peak pack accumulation and a schedule that keeps pace with new sites.
4. Business interruption sized on cluster loss, not single-site loss, with an indemnity period matched to battery resupply lead times.
5. Public and product liability, covering both premises exposure at swap points and the operator's exposure as owner of packs and vehicles that fail away from any site. See our overview of liability insurance for the structure.
6. Directors and officers cover appropriate to a pre-listing company, since the stated plan is a listing after this round, and the diligence that precedes it looks closely at whether the insurance programme matches the asset base.

The sequencing point is the one operators most often miss. Insurance capacity for a fleet quadrupling in size is priced on the loss record of the smaller fleet that preceded it. An operator that goes into the expansion with clean rider verification data, telemetry, claims discipline, and engineered swap sites gets a very different market response two renewals later than one that adds 150,000 vehicles first and tries to explain the loss ratio afterwards. Broadly comparable questions arise across electric fleet operators generally, as set out in our risk profile of electric bus fleet operators.