When Fisker Inc. filed for Chapter 11 in June 2024, roughly 11,0001 Ocean SUV owners found themselves holding vehicles priced between $40,0001 and $70,0001 with no manufacturer support and no clear path forward. Digital rights advocate Cory Doctorow described the Ocean as a “software-based car” where virtually every subsystem required periodic cloud connectivity. The owners’ response: fund the reverse engineering themselves.
The Brick Problem
When Fisker’s cloud infrastructure went dark, owners lost more than remote access and OTA updates. Basic diagnostic and service workflows required cloud-authenticated sessions pointing at dead infrastructure.
American Lease bought Fisker’s US assets and paid an extra $2.5 million1 specifically for source code and cloud access. The company made an informal agreement with the Fisker Owners Association to extend connected services, then demanded the FOA cover 58% of operational costs1 while refusing to provide itemized invoices. The deal collapsed. For private owners, American Lease subsequently revoked cloud API access entirely.
That sequence is worth noting: the acquirer held both the source code and the cloud infrastructure, made a handshake deal with the community, then used cost-sharing terms to kill the arrangement. Whatever survives of the Ocean’s software stack does so despite, not because of, the asset buyer.
The Fisker Owners Association as a De Facto Automaker
The Fisker Owners Association grew to roughly 4,0001 members and structured itself as a nonprofit that functions somewhere between a car club, a tech startup, and an independent automaker. The comparison to the latter is not overstatement: the organization is doing failure analysis, parts sourcing, software maintenance, and field repair coordination, functions a manufacturer handles internally when it still exists.
Vitalik Buterin tweeted in July 2024: “We really need much more open source in the auto industry. Really sad that if the manufacturer disappears, the car is useless now has seemingly so quickly become a default.” The FOA represents the practical answer to that observation: a community large enough to fund and coordinate engineering work that would otherwise require a company to perform.
Reverse-Engineering the Stack: From Cloud API to CAN Bus
GitHub user MichaelOE reverse-engineered the My Fisker mobile app API and published a Home Assistant integration under Apache 2.0. The repository has 1352 commits and 20 releases2. The integration works by polling the cloud service for the car’s digital twin, which, given American Lease’s API revocation for private owners, constrains its current utility to whatever residual cloud access still exists.
The deeper work is on the CAN bus layer. The Ocean runs four buses: CCAN, PTCAN, Inverter CAN, and BCAN, all at 500 kbps1. GitHub user puddletools has published DBC files to support CAN viewer filtering. DBC files map raw frame IDs and byte offsets to human-readable signal names, decoding which bytes represent state of charge versus motor temperature is tedious enough that sharing the maps is a genuine contribution.
Third-party tooling has also appeared. Freesker (from Armenian Solutions for Advanced Logic and Analytics) offers browser-based DTC read/clear, ECU firmware checks, and service actions. It requires a Vgate vLinker FD+ OBD dongle and is VIN-locked to a single vehicle. Pricing runs €99 for the Lite tier and €6783 for Pro. The VIN lock means that comprehensive diagnostic access carries a per-vehicle cost that not all owners will absorb, and it forecloses the community sharing a single license.
The Flying Doctors, Group Buys, and Parts
The FOA’s most immediately practical contributions are not software. In Europe, the organization created the “Flying Doctors” program: a mobile repair network where technically skilled members travel to assist other owners with repairs. This is mutual aid organized around automotive triage.
On the parts side, FOA coordinated bulk purchasing that dropped replacement key-fob prices from roughly $1,0001 each to a fraction of that figure. The organization also hosted free global pairing events that saved individual owners between $1001 and $2501 per visit. These are not dramatic engineering wins, but they reflect where most of the $40,0001 to $70,0001 ownership cost shows up when support infrastructure disappears: in the unplanned individual repair bill.
Why This Matters for Software-Defined Vehicles
The Fisker case forces a specific question for any automaker currently shipping cloud-dependent critical-path features: what is the explicit commitment to maintaining that infrastructure over the vehicle’s operational life, and what happens when the company is acquired, restructured, or dissolved?
The answer most automakers currently give is implicit: nothing is promised, and the EULA disclaims liability for service discontinuation. Fisker did not invent this position. Tesla has shipped OTA updates that alter or remove features on delivered vehicles. GM shut down OnStar connectivity for older hardware. The Ocean is the most acute version of a pattern that has been visible for years.
What the FOA demonstrated is that the failure mode is not terminal when the owner community is large enough and technical enough to reverse-engineer the stack. Roughly 4,0001 members, organized as a nonprofit, can fund API reverse-engineering, publish CAN bus maps, coordinate a mobile repair network, and negotiate, however unsuccessfully, with asset acquirers. That is not a small community. Most vehicles that go out of production do so with owner bases that are smaller, less technical, and less motivated.
The Legal and Technical Limits Ahead
The community has done what could be done without manufacturer cooperation: mapped the CAN buses, reverse-engineered the mobile API, built third-party tooling, and created a self-organizing repair network. What it has not done, and likely cannot do without the source code American Lease holds, is restore OTA update infrastructure or rebuild the full cloud-dependent feature set from local components.
The legal position is murky in ways that are structurally similar to other automotive right-to-repair fights. DMCA Section 12011 exemptions for vehicle software have been extended and contested cyclically. The specific acts that MichaelOE and puddletools have taken, API reverse-engineering, CAN bus sniffing, sit in territory that has not been tested in court against a motivated asset holder.
American Lease’s behavior during the cloud-service negotiation suggests the company is fully aware of the infrastructure it controls. Demanding 58%1 of operational costs without itemized invoices is not a good-faith negotiation posture. It is a posture that expects the counterparty to walk.
The second-order consequence that matters for the industry is direct: when a manufacturer ships critical-path functionality through a cloud service it controls and then files for bankruptcy, the running infrastructure becomes an asset in liquidation proceedings. That asset can be bought, withheld, and used as a condition in negotiations against the people who bought the hardware. The Fisker case is the clearest illustration of that dynamic yet produced. It will not be the last.
Frequently Asked Questions
How is the Fisker case different from GM or Tesla discontinuing connected services?
GM’s OnStar shutdown affected vehicles at the end of their supported hardware lifecycle after multiple product generations had superseded them. Tesla’s OTA changes alter features on vehicles whose manufacturer still exists and maintains the update pipeline. Fisker is the first case where the manufacturer folded entirely and an asset buyer then revoked cloud access while holding the sole source code, on a vehicle that hadn’t completed its first full year in customers’ hands.
Can OTA updates ever be restored without American Lease’s cooperation?
Practically no. Pushing firmware to ECUs requires the manufacturer’s signing keys, the build toolchain that compiles ECU images, and the secure-boot chain that validates them, all of which reside in the $2.5 million source-code package American Lease purchased. The community’s CAN bus maps and Freesker’s service actions operate within existing firmware capabilities and cannot inject new code. The diagnostic layer is forked open; the update layer remains locked.
Does the Ocean’s CAN bus documentation help owners of other vehicles?
The published DBC files map frame IDs to signal names for the Ocean’s specific ECU firmware version and would not transfer to any other vehicle, Fisker or otherwise. The reverse-engineering methodology—observing correlations between physical actions and CAN traffic—is portable across makes and models, but the resulting maps are model-specific and must be regenerated from scratch for each platform.
What should EV buyers verify before purchasing from a startup automaker?
Check whether core driving, charging, and diagnostic functions operate without cellular connectivity, since the Ocean demonstrated that cloud-dependency becomes a brick risk when a manufacturer fails. Buyers considering Rivian, Lucid, or Polestar vehicles should note that no current automaker provides software escrow, source-code commitments, or guaranteed offline fallback for critical-path systems. The practical floor is a vehicle whose basic transportation and charging functions run entirely on local firmware.