Milan knows how to dress a wound. The fashion houses of Brera understood long before the engineers of Cupertino that the most beautiful objects are often the ones that never reach production — that the sketch is more honest than the car, the sample more true than the collection. It is in this spirit that we present our taxonomy: ten electric technologies that were conceived, announced, funded, poached, delayed, revised, abandoned, and eventually absorbed — like so many runway pieces — into the vast archive of almost.

These are not failures in the ordinary sense. Failure implies a lack of ambition. What we document here is something more interesting: the overreach of the perfectly dressed. Projects that wore their promises like Milanese tailoring — impeccable on the surface, with nothing holding the seams together underneath.

Apple launched Project Titan in 2014, assembling more than a thousand automotive specialists at a classified facility near its Cupertino campus. The ambition was total: a fully autonomous, steering-wheel-free electric vehicle that would do to transportation what the iPhone had done to communication. The metaphor was irresistible, and so it was repeated — endlessly, uncritically — until it became its own kind of architecture.

What happened inside those walls over the following decade was less a car programme and more a philosophical negotiation. Should the vehicle be fully autonomous? Semi-autonomous? Should it have a steering wheel, offered as a kind of retro luxury, like a manual transmission on a Porsche? The questions cycled through leadership teams with the regularity of Milanese seasons. No answer was ever finalised. The sketches accumulated.

By 2022, the envisioned price had dropped from $120,000 to $100,000. The target launch had been pushed from 2025 to 2026, then to 2028. The fully autonomous dream had been quietly interred somewhere around 2019. What remained was essentially a very expensive EV with uncertain ambitions — a concept car, sustained by corporate confidence and accounting conventions.

The ending arrived on February 27, 2024. COO Jeff Williams and VP Kevin Lynch summoned the remaining 1,400 engineers to a meeting. The meeting lasted twelve minutes. No questions were taken. Most of the team was redirected to generative AI. Six hundred Santa Clara employees were laid off by April. The folders were archived. The facility was repurposed.

By 2024, Ford had cut Lightning production from a projected 150,000 annual units to approximately 40,000. The next-generation electric truck platform was delayed. Losses at Ford Model e — the EV division — had reached $1.3 billion in a single quarter. The company announced it would extend its hybrid vehicle production timelines indefinitely. The Lightning persisted, but at a fraction of the scale that had made it culturally significant.

The architectural lesson is one of load-bearing promises. A building can be announced with its façade. The façade can be beautiful, iconic, photographed, published. But if the structural system behind it hasn’t been resolved, the moment someone moves in, the problems begin. Ford had designed the Lightning’s exterior — its cultural positioning, its reservation system, its press strategy — before the structural economics were sound.

General Motors announced in January 2021 that it would phase out all petrol and diesel vehicles by 2035. It was one of the most aggressive decarbonisation commitments from a legacy automaker — a Damascene conversion, apparently, from the company that had famously crushed its EV1 electric cars in the 1990s. Mary Barra presented the pledge with conviction. The press received it with something approaching relief. The future, finally, had been scheduled.

The reversal was not announced. There was no counter-press conference, no formal withdrawal. The pledge simply ceased to be operational. In 2024, GM extended its hybrid production timelines, quietly de-prioritised Ultium battery investments, and began the process of making the 2035 deadline meaningless through incremental revision. The language of commitment remained; the commitment itself had been dissolved.

Cruise represented the furthest frontier of electric mobility: not merely a zero-emission vehicle, but a vehicle that required no human driver at all. By mid-2023, Cruise was operating hundreds of autonomous electric robotaxis in San Francisco, accumulating ride data at scale, proving out an operational model that most of the industry had theorised but never deployed commercially.

In October 2023, one of Cruise’s vehicles struck a pedestrian who had already been hit by a different car, and dragged her approximately twenty feet before stopping. The subsequent investigation revealed that Cruise had provided incomplete information to regulators. The California DMV suspended the company’s permit to operate driverless vehicles. GM paused operations nationally. The CEO resigned. Hundreds of employees were laid off. The robotaxi fleet, so recently a proof of concept, became a monument to the gap between capability and accountability.

For architects, the Cruise episode carries a specific lesson. A building can be structurally sound but socially illegible — its relationship to the street, to pedestrians, to the human grain of the city, can be wrong in ways that no engineering specification captures. The Cruise vehicle functioned. Its function was the problem.

In late 2021, freshly emerged from bankruptcy, Hertz announced it would purchase 100,000 Teslas — the largest EV fleet acquisition in automotive history at the time. The stock surged. The press declared Hertz’s reinvention complete. The CEO appeared on the cover of things.

The problems were structural and they arrived immediately. Tesla’s repair network was not calibrated for high-frequency rental use. Damage rates were higher than anticipated. Repair costs were catastrophic — a minor collision that would cost $3,000 to fix on a conventional vehicle could cost $15,000 on a Tesla. Rental revenue did not compensate. By early 2024, Hertz announced it would sell approximately 20,000 EVs from its fleet — primarily Teslas — to cut its losses and repurchase conventional petrol vehicles.

The fire sale proceeded at prices well below acquisition cost. Elon Musk, noting that Hertz had never actually secured a volume discount on its original purchases despite press reports to the contrary, offered no public sympathy. The EV transformation, so confidently announced, had lasted approximately twenty-four months before the economics of the real world reasserted themselves with characteristic indifference.

Production delays accumulated. Software updates bricked vehicles. Owners reported doors that wouldn’t open, screens that froze, range estimates that bore no relationship to real-world conditions. Service centres didn’t exist in most markets. Fisker attempted to negotiate a rescue partnership with Nissan — the talks failed. A potential deal with a large OEM collapsed. By June 2024, Fisker had filed for Chapter 11 bankruptcy, leaving approximately 5,000 Ocean owners with vehicles of declining value and no support infrastructure.

For architects, the lesson is the difference between a beautiful rendering and a building. The Ocean existed in its fullest, most complete form on a turntable at a motor show in 2020. Every subsequent version — the prototype, the pre-production unit, the delivered vehicle — was a degradation of that original image. Architecture that only works as an image is not architecture. It is illustration.