Which manufacturers actually deployed V2V — and what architecture each one used.

Every implementation in this audit resolves to one of these four paths. The difference between them is the message path, not the vocabulary the manufacturer uses to describe it.

Latency is measured in tens of milliseconds. Coverage does not exist as a concept: either the two vehicles are within radio range or they are not. There is no subscription, no roaming, no backend, and nothing to switch off remotely. How the V2V message set works

Latency is seconds, not milliseconds, which rules out crash-imminent warnings and rules in road-condition and hazard information. It works at any range, including tens of kilometres ahead, which direct radio cannot do. It requires mobile coverage, a connected-services account, and a manufacturer willing to keep the backend running. Why the network path is a different tool

This is the architecture that actually scales first, because the value does not depend on how many other cars are equipped. One instrumented ambulance benefits every equipped vehicle it approaches. Why V2I scales linearly and V2V does not

A second vehicle being detected is not a second vehicle communicating. Advanced driver assistance and automated driving belong in this row no matter how capable they are, and a fleet that learns from aggregated driving data still belongs in this row, because the learning reaches the other vehicle as a software update rather than as a message about a hazard that exists now. The strongest argument for building it this way

Class letters refer to section 3. Deeper technical detail for each row is on its own case study rather than in additional columns, because a twenty-column table is unreadable on every device anyone actually uses.

Mercedes-Benz Car-to-X Communication. Offered since the 2016 E-Class and installed as standard in combination with navigation. When the vehicle's stability-control sensors identify low friction, that observation and its position are sent over the mobile network to the manufacturer's vehicle backend, anonymised there, and distributed to vehicles that are approaching. The customer needs a connected-services account and has to activate the service; it is included for an initial period and chargeable afterwards. Euro NCAP, assessing it, described the field generally as one where different communication strategies may be used — a mobile network, or a more direct but shorter-range radio transmission — which is exactly the boundary this audit is drawn along. The Mercedes-Benz Car-to-X case study

Volvo Connected Safety. Hazard Light Alert and Slippery Road Alert, first offered on the 90 Series in Sweden and Norway in 2016 and extended across Europe from April 2019, available on Volvo models built on the SPA and CMA platforms. When a car switches on its hazard lights the event goes to the cloud and out again to nearby connected Volvos. Volvo has been unusually explicit that the data is pooled and has invited others to join, and has shared the resulting road-condition data with municipalities and a navigation platform. The Volvo Connected Safety case study

Neither system should be dismissed. Both cover distances no radio can reach, both work on the first day a single car is equipped rather than waiting for a second one, and Volvo's slippery-road function in particular addresses a hazard that direct V2V handles badly, because friction is a property of a road segment rather than an event at a moment. What they cannot do is warn a driver about the car braking hard forty metres ahead, and no amount of backend engineering will change that.

General Motors equipped the Cadillac CTS with DSRC-based V2V from the 2017 model year, as standard equipment, and continued through model year 2019. In 2018 it announced plans to expand across the Cadillac fleet from 2023.

The National Transportation Safety Board's February 2024 review of the Department of Transportation's V2X deployment plan records what followed, in three sentences that between them explain the entire American position: General Motors never initiated the announced expansion; Toyota suspended its April 2018 plan to equip part of its United States fleet, citing regulatory uncertainty and the absence of deployment plans from other automakers; and Ford's 2019 plan to equip part of its fleet with C-V2X by 2022 did not occur.

Around this, the regulatory floor moved. The Federal Communications Commission reduced the transportation safety band from 75 MHz to 30 MHz, allocated the remaining 30 MHz to C-V2X alone and ordered the discontinuation of DSRC. Because the existing deployments were DSRC, infrastructure owners had to end them — and in 2021 that meant 34 states running 57 operational projects with more than 15,500 instrumented vehicles and 6,200 intersections and roadside units. In November 2023 the National Highway Traffic Safety Administration withdrew the 2017 proposal to mandate V2V in new light vehicles.

The NTSB's own assessment of what remains is blunt: the plan targets about 6 per cent of United States passenger vehicle models with 5.9 GHz capability by 2030–2034, and at the time of writing no automaker had committed to the near-term deployment the plan assumes. The Cadillac V2V case study · United States regulation in full

Tesla. No 5.9 GHz V2X deployment, no V2X radio hardware identified in any shipping vehicle, and not a member of the industry alliance whose roster contains almost every other manufacturer named on this page. Tesla is active in the energy modes — vehicle-to-load and vehicle-to-home through Cybertruck Powershare, and vehicle-to-grid programmes with utilities in Texas and California — but those share an abbreviation with V2X rather than a technology.

The confusion is understandable and worth naming precisely, because four separate things get read as V2V and none of them is:

• Autonomy is not communication. A vehicle that drives itself extremely well is still deciding alone. Full self-driving capability makes the sensor-only row of the diagram in section 2 better; it does not move a vehicle out of that row.
• Fleet learning is not V2V. Aggregated driving data improving a model that is later shipped as an update is a development process. V2V is a message about a hazard that exists now, arriving in time to act on it.
• Robotaxi coordination is not V2V. Vehicles being dispatched and routed by a backend are clients of a central service. Two of them benefiting from data that originated in one another does not make a direct link between them, and a cooperative-safety receiver cannot use it.
• Camera-based emergency-vehicle detection is not V2X. Recognising a police car is perception. Receiving an authenticated message from one is communication.

None of this is a criticism of the architecture. There is a serious engineering argument for building a vehicle that depends on nobody else, and it is set out at length elsewhere on this site rather than caricatured. It is simply not a V2V deployment, and it should not be counted as one. Tesla and the sensor-first argument

Hyundai, Kia and Genesis. Hyundai Mobis launched production C-V2X modules in August 2023, but the customer manufacturers were not disclosed, and no Hyundai, Kia or Genesis passenger vehicle with active factory V2X could be verified. The group's public vehicle-to-everything activity is concentrated in the energy modes.

Nissan. An alliance member with trial participation in Japan. No production deployment verified in any market.

Indian manufacturers. No production V2X fitment by any Indian manufacturer could be verified. The activity is regulatory and pre-competitive: a research demonstration, an academic collaboration, and a draft standard. India's V2X framework

Truck and bus manufacturers. No series-production cooperative V2V could be verified for any major commercial-vehicle manufacturer. Platooning, the programme that would have driven it, was wound down — Daimler's truck division stopped its platooning work in 2019 after concluding the fuel case did not hold at realistic following distances. The truck-side V2X that does exist in the field is public-sector and infrastructure-flavoured: instrumented snowplough and queue-protection fleets operated by state transport departments, which belong in category C.

A negative finding is a finding. Each of these was investigated against the same sources as the positive rows, and the absence of evidence is recorded rather than filled in with inference.

These terms are not interchangeable, and this audit does not treat them as such.

• IEEE 802.11p is the physical and link layer. DSRC is the North American system built on it with the IEEE 1609 WAVE stack and SAE J2735 messages. ITS-G5 is the European system built on the same link layer with ETSI messages and GeoNetworking. Wi-Fi p is the name Volkswagen's own customer material uses for its implementation. They share a radio; they are not one specification.
• C-V2X is the 3GPP family. PC5 is its direct device-to-device interface — the part that is comparable to ITS-G5. Uu is the ordinary path to a base station, which is what carries V2N. LTE-V2X is Release 14 onward; NR-V2X is the 5G sidelink from Release 16. A 5G modem provides Uu. It does not provide PC5 unless it was specified to.
• Japan's 760 MHz ITS band is neither of the above. It is a national allocation with its own standards, which is why an ITS Connect vehicle exported from Japan carries a radio that other countries may not permit to be operated at all.

No production deployment identified in this audit uses 5G NR-V2X sidelink. Every direct system found is either 802.11p-family or LTE-V2X PC5. The full DSRC and C-V2X comparison · Spectrum by region

• A cloud-mediated feature has a backend compatibility lifetime. Audi's own technical service bulletin, filed in May 2023, records that traffic-light information stopped working on 2017 model-year A4 and Q7 cars built before a specific production week, because those vehicles carry software no longer compatible with the backend that provides the data, and that there is no software or hardware update that resolves it. The service itself continues; it was that population that lost it. A direct-radio system has no equivalent failure mode; its exposure is certificate expiry, which is a solved problem with a known remedy.
• Spectrum policy is product policy. Toyota's system cannot leave Japan, and the United States industry stopped when its band was reopened. Nothing in either case was an engineering failure.
• Infrastructure does not create a fleet. China allocated spectrum in 2018, built more than eleven thousand roadside units and ran twenty pilot cities, and vehicle penetration was still around four per cent in 2025. Roadside deployment and vehicle fitment are separate problems and need separate instruments.
• The first deployable value is not car-to-car. Every large operating deployment in this audit that is not Volkswagen's derives its value from an asset that is not an ordinary passenger car — an ambulance, a work zone, a signal controller, a road-condition aggregate. That is what linear returns look like next to quadratic ones.
• An option box is a fitment strategy. Volkswagen reached two million vehicles by making the function standard on one mass model and a cheap option across the rest of the range. The manufacturers who offered it only on a flagship never reached a density at which anything happened.
• Cross-manufacturer interoperability is the whole point and is almost never tested in public. Volkswagen states its exchange process is open and standardised across manufacturers. In Europe that claim can in principle be exercised. In the cloud-mediated systems it cannot: a Volvo warns Volvos, a Mercedes warns Mercedes, and the pooling that Volvo has publicly invited has no equivalent in the direct world because the direct world got interoperability for free from the standard.

1. Volkswagen Group — A milestone for greater road safety: more than two million vehicles produced with Car2X, 27 October 2025. Vehicle count, Wi-Fi p, 800 m, no mobile network, anonymisation, user activation, cross-manufacturer exchange, model list.
2. Toyota Motor Corporation — Toyota Bringing Advanced ITS Technology to Mass-market Models, 2015. ITS Connect launch, 760 MHz, V2V and V2I functions, initial models.
3. Waka Kotahi NZ Transport Agency — Vehicles fitted with ITS Connect. Regulator confirmation of the 760 MHz band, the affected Toyota and Lexus models from 2015 onward, and the New Zealand licensing position.
4. National Transportation Safety Board — Response to the USDOT V2X Deployment Plan, 2 February 2024. Cadillac CTS model years, the unexecuted GM expansion, Toyota's suspension, Ford's unmet commitment, the FCC band reduction, the 2021 infrastructure figures, the NHTSA withdrawal, and the 6 per cent target.
5. Audi of America — Technical Service Bulletin 91 23 91, 2070216/1, 5 May 2023, filed with the National Highway Traffic Safety Administration. Withdrawal of Traffic Light Information Online from early 2017 model-year A4 and Q7 vehicles for backend incompatibility.
6. Euro NCAP — Advanced Reward assessment of Mercedes-Benz Car-to-X Communication. Independent assessment of the system and of the distinction between network-mediated and direct short-range strategies.
7. Volvo Cars — Volvo models across Europe to warn each other of slippery roads and hazards, April 2019. Europe-wide launch, cloud architecture, SPA and CMA platform coverage, 2016 origin.
8. HAAS Alert — Industry-leading V2X activation equips 1.8 million Stellantis vehicles with Emergency Vehicle Alert System, 23 May 2023. Vehicle count, brands, model years, markets, alert originators and the platform path.
9. BMW and JOYNEXT — BMW launches the all-new 5 Series with V2X technology in China, April 2025. First BMW model with built-in V2X, China market, supplier.
10. Renault Group — “Renault works with SCOOP to prepare infrastructure for tomorrow’s autonomous, connected cars”. The thousand-vehicle ITS-G5 fleet trial.

External links are provided so that every figure can be checked at its origin. They do not imply any relationship with, or endorsement by, the organisations named, and the trademarks referenced belong to their respective owners.