The Parked EV Fleet: A Grid Storage Asset the Grid Hasn't Turned On

Roughly 250,000 electric vehicles already capable of selling power back to the grid sit parked across the United States tonight. The grid draws on almost none of them.

Vehicle-to-grid is an old idea with a recently hardened business case. A parked electric car is a battery on wheels, wired to the grid through a bidirectional charger; a fleet of school buses or factory shuttles holds more usable storage than most dedicated utility batteries. The University of Delaware has run its fleet this way for years, with PJM Interconnection paying close to $1,200 per vehicle each year for the hours those cars sit plugged in and available.

What kept the idea in laboratories for two decades was nothing physical. The chargers worked, the cars worked, and the market hardly did. Grid operators had no reason to buy flexibility that needed standards nobody had agreed on, tariffs no utility had written, and software no vendor had shipped. All three shifted in the last eighteen months, which is why the conversation moved from "will V2G ever happen" to "which companies monetize the parked fleet first."

The raw material is already on the road. The International Energy Agency's 2026 vehicle-to-grid assessment counts 22 commercially available models with two-way charging capability. That is a narrow sliver, and the agency is explicit about it: those 22 models account for less than 1.5 percent of all EV models on sale.

What matters is who the 22 are. The capability landed first in volume formats (Renault's compact city car, GM's mass-market SUVs, Nissan's fleet stalwart), not in luxury trims. The hardware pipeline is widening faster than the market's ability to absorb the idea, which is the normal shape of an infrastructure transition: the first two years feel slow, then the curve bends.

The capacity math is elementary. A 60 kWh battery, the compact-class EV, holds more than a day of average household consumption. A ten-bus school fleet sits on more than 300 kWh of storage between morning and evening runs. The White Plains school district's battery buses have fed New York's Con Edison grid for years, the first such arrangement in the state. Brooklyn tested its first wave of Nissan Leafs at 45 kW under a Revel, NineDot and Fermata Energy partnership. None of it required new physics. It required willingness to plug in, and a counterparty willing to pay.

No utility wants to negotiate with a million parked cars. The IEA defines the workaround precisely: an aggregator is a platform that combines many small electricity resources into a single participant in the power market. The electric analogue of an asset manager assembling retail savers into a fund. The analogy holds in a useful way, because the fees and the risk live in the same place.

The intelligence layer is new. Battery management used to be physics-based estimation from voltage, current and temperature. The IEA describes AI models that learn state and degradation from charge and discharge history instead, and that matters twice over: predictive degradation modelling decides whether an owner will allow deep cycling, and the same models forecast when a fleet can profitably discharge. The two questions are the same question.

The market is already populated. ChargeScape, incorporated in 2024 out of Austin by BMW, Ford and Honda, operates as a shared vehicle-grid integration platform across three automakers' enrolled vehicles. Evert in Norway runs AI-driven V2G energy management on a €1.6 million pre-seed. Carbvolt in Spain prices bidirectional energy with machine-learning forecasts. South Korea's VPP Lab links V2G fleets into virtual power plant (VPP) dispatch, and Australia's V2Grid Australia is building the same stack around Mitsubishi's commercial vehicles. None of these companies is a household name. That is usually the moment before the market consolidates.

France got the first consumer-grade V2G service. Renault's 52 kWh R5 E-Tech ships with two-way capability, and Mobilize with The Mobility House sells it as a bundle: the car, a bidirectional charger, an energy contract and software that manages the charge and discharge automatically. Expansion into the Netherlands, Germany and the UK was planned through 2026. In Utrecht, more than a hundred Renault 5s are live in Europe's largest V2G car-sharing project, heading toward 500 vehicles.

Germany's first commercial offer came from BMW with E.ON, using the Wallbox Professional, one of the first bidirectional CCS wallboxes on the market; BMW's Neue Klasse platform will carry the capability from the outset. Mercedes-Benz announced the same for the electric GLC through MB.CHARGE Home with The Mobility House. In Denmark, Nissan, Enel and Nuvve operate the world's first fully commercial V2G hub, where the grid operator can charge or discharge plugged-in vehicles on demand.

The US market moved in June 2026. GM activated V2G for existing customers using its bidirectional hardware, with no new hardware required where utility programs support export. Its headline numbers were unusually concrete for an automaker: a quarter-million bidirectional vehicles in circulation, theoretical storage capacity to power 120,000 homes for a week, and a Pacific Gas and Electric (PG&E) collaboration aimed at 52,000 of its vehicles participating in grid balancing by 2030. DTE Energy is testing the same idea across employee homes in Michigan. It is still small relative to an installed base measured in millions, but it is no longer a roadmap slide.

China is scaling the pilot layer at a cadence that makes Western programs look cautious: 30 V2G pilots across nine cities announced in 2025, with a target of 5,000 V2G charging facilities by the end of 2027. Its constraint is standardisation, the same one everywhere else: bidirectional protocols for the GB/T standard are not fully settled. Asia Pacific is moving in parallel: Hyundai and Kia run AC V2G pilots aligned with Korea Electric Power Corporation's (KEPCO) grid-code update proposal in South Korea, Australia's biggest utility is trialling V2G subscription bundles built around the BYD Atto 3, and Brazil's ANEEL authorised a 2026 V2G sandbox that lets distributors blend EVs, solar generation and storage into one aggregator billing structure.

The strongest current number comes from energy modelling consultancy E3, whose July 2026 study found that vehicle-to-grid integration delivers five to fifteen times more value per EV than one-way managed charging in many electricity markets. The reason is diversification: the same battery earns from frequency regulation in the morning, voltage support at noon, a demand-response call in the afternoon and time-of-use arbitrage overnight, then sits as an emergency reserve the utility pays to keep available. A one-way charger captures exactly one of those streams.

As we wrote in August, Noon Energy booked 1.1 GW of storage in four months, with AI data centers flagged as the next buyers. Stationary storage sets the benchmark for what an equivalent parked fleet should earn. The V2G pitch is that comparable megawatts already exist in driveways, already purchased, with the purchase price sunk.

Nuvve supplies the early commercial proof. The aggregator behind the Denmark hub and the Delaware fleet reported its Q2 2026 update in August and is expanding under its OMNIA partnership: a third European battery project in weeks, combined capacity crossing 150 megawatts, a 40 MW Austrian deal, and a 20-year merchant aggregation agreement in Japan. The structure is not exciting. These are long-lived contracted assets with merchant upside, which is precisely what institutional capital tolerates. The pattern belongs in the same family as a data center REIT or a pipeline: slow, contracted, and boring in the best sense.

The people hostile to this thesis should also own it, because the downside is easy to price: nobody is buying a car to sell kilowatt-hours. V2G revenue is a secondary asset on a primary purchase, and secondary assets get ignored in the owner manual. The market's job over the next cycle is to prove owners actually capture the revenue, the same adoption curve demand-response programs have climbed, contract by contract, for two decades.

The first wave of revenue splits across three claimants. Charger hardware captures the installation sale, the aggregator captures a recurring slice of every discharged kilowatt-hour, and the utility captures the deferral of grid investment. The interesting margin sits in the middle, because hardware heads toward commodity pricing while software compounds.

California's residential market is already testing that split. Bidirectional Energy, run by veterans of V2G programmes, is deploying residential bidirectional charging with Wallbox, pairing two-way wallboxes with a tariff structure that pays the homeowner to discharge. The message from the company is explicit: the value is not in the box, it is in the orchestration behind it.

Charging manufacturers are racing the same way. Star Charge, one of the largest charging-infrastructure operators by global footprint, is expanding its V2G position through certifications and partnerships across Asia, Europe and Australia. Nissan earned the UK's first G99 grid certification for AC V2G, a bureaucratic milestone that quietly opens one of the most conservative jurisdictions. Kia and Hyundai are offering a customer V2G service in the Netherlands on the EV9 and IONIQ 9, bundled like a phone plan rather than sold as a feature.

None of these moves is dramatic on its own. Taken together they describe a market sorting into a familiar shape: the automaker supplies the battery and the warranty, the hardware maker sells boxes, and the software layer, the aggregator that bids the fleet, forecasts prices and times the discharge, carries the widening margin. That last layer is where private capital should be looking, because it is the only piece with a compounding cost structure.

Interoperability is the largest wall. The IEA's assessment is blunt: the communication layer between charger, aggregator and grid is today proprietary or absent, and commercial V2G offers do not extend beyond a single country, typically tied to one utility. The United Nations Economic Commission for Europe (UNECE) has proposed harmonising grid codes and communication links, but the standard layer today resembles telecommunications before number portability.

OEMs are the second brake. Eleport's 2026 market survey counts a short list of cars that actually export power: Renault's R5, BMW's iX3, Kia's EV9, Hyundai's IONIQ 9, GM's Equinox and Sierra and LYRIQ. Most new EVs still ship with one-way duty only. The commercial V2G model today is locked to approved chargers and approved tariffs. Rivian's R2 is expected to become its first V2G-capable model; Polestar and Clever are piloting the Polestar 4 in Denmark. The momentum is there; the reach is not.

The third brake is regulatory fragmentation. Grid codes differ country to country, and a bidirectional vehicle approved in Germany assumes nothing about Poland. Revenue pools exist in the US through the wholesale layer, yet the E3 study is blunt that owners struggle to earn compensation for capacity under current utility programs, and the monetization pathways it charts still need regulatory evolution. Every pilot is also a test of patience with interconnection paperwork.

The sceptical case deserves equal weight. The capable fleet is concentrated in a handful of models; the IEA counts 22 of them against a global catalogue in the thousands. GM itself notes EV sales dropped this year and utility adoption of full-scale V2G programs remains limited. The value math is contingent on market design. Payments for flexibility default to zero anywhere no such market exists. And the security surface is real: published research has demonstrated wireless attacks capable of disrupting EV fast-charging sessions, a vulnerability class that scales to fleets of grid-connected batteries.

That counter is weaker than it was two years ago, but it is not wrong. The honest read: V2G is a demand-response product with a vehicle on top, and demand-response economics have always scaled slowly, market by market. The question for the next two years is more administrative than technological: tariffs that price discharging, grid codes that talk to each other, warranties that allow it.

The parked fleet is the only storage resource that is already paid for. The question nobody outside this market is asking yet is where the value accrues between the socket and the wholesale market. The answer, increasingly, sits in software that does not yet have a household name.