Here is a take that used to get me labeled an Apple apologist, and now finally looks like plain common sense: the Apple Car was never a failure. It was a ten-year, ten-billion-dollar research lab wearing a car-shaped disguise, and its real output is now arriving in the company's most important silicon ever. The M7 and M8 chips that Apple is rushing toward are, in a very real sense, the ghost of Project Titan.
Bloomberg's Mark Gurman reported this week that Apple is designing its next-generation M7 and M8 processors with AI support as the headline priority, and that the designs trace directly back to the company's self-driving car efforts. The M7 generation reportedly skips the usual obsession with raw speed and power efficiency in favor of serious Neural Engine upgrades. Gurman also says Apple is skipping the Pro, Max, and Ultra versions of the M6 entirely to accelerate the M7, which should arrive in the first half of 2027.
I have been saying for years that killing the car was the smartest thing Apple ever did, and this is the week the receipts showed up.
The Car That Was Really an AI Lab
Let us remember what Project Titan actually was. It ran for roughly a decade, consumed around $10 billion, and produced exactly zero cars. By every conventional measure, it was the most expensive failure in Apple's history. But that framing has always missed the point.
To build a self-driving car in 2014, Apple needed to process enormous amounts of camera, lidar, and sensor data locally, in real time, with no cloud connection and no room for latency. That problem forced Apple to invent the Neural Engine, the dedicated on-device AI processor that debuted with the A11 Bionic in the iPhone X and powered Face ID, Animoji, and augmented reality.
Tim Cook essentially told us this back in 2017, when he described autonomy as "the mother of all AI projects" and admitted it was "probably one of the most difficult AI projects actually to work on." Nobody believed him then. Everyone was convinced Apple had been caught flat-footed by the AI boom. The truth, as usual, was stranger and more patient: Apple had been building the hardware foundation for years while the rest of the industry was still treating AI as a cloud service.
When the car was finally cancelled in 2024, its engineers were not scattered to the wind. They were redeployed directly into the AI team led by John Giannandrea. The people who spent a decade solving real-time on-device inference problems became the people building Apple Intelligence.
Why the M7 Story Matters Now
The M7 is reportedly the first Apple chip generation designed around AI first, with the traditional priorities demoted. That is a bigger deal than the specs suggest. It means the on-device AI philosophy that the car project forced into existence is now the company's central strategy, not a side project.
Here is what we know about the M7 family, mostly from Gurman's reporting and the coverage around it:
- The base M7 should land in the first half of 2027 with significantly upgraded Neural Engine performance.
- An M7 Ultra is reportedly in the works as the basis for a new Apple server product, with support for up to 1.5TB of unified memory.
- Early work on M8 processors has reportedly already begun, continuing the same AI-first design philosophy.
That last bullet deserves attention. Apple building its own AI server silicon is the quiet revolution underneath all the consumer chip drama. If the M7 Ultra really powers a new server line, Apple is doing what it always does best: vertically integrating its way out of dependency, this time on Nvidia and the cloud.
The timing is also fortunate. Apple was late to the hyperscale data center spending spree, and that has been used against it for two straight years. But the math has started to flip. In June 2026, reports showed OpenAI losing $1.25 for every dollar earned, and investors have begun drifting back toward Apple precisely because Apple never burned its balance sheet on subsidized AI.
I am not going to pretend the software side is fixed. Apple Intelligence has been uneven, Siri still embarrasses itself, and Gurman's own reporting admits Apple's AI software efforts lag the industry. Hardware alone does not win an AI war.
But here is my thesis, and I am sticking with it: the company that owns the on-device inference layer, the privacy story, and now its own server chips has a structural advantage that no amount of cloud capex can copy. The dead car project handed Apple the Neural Engine, the M7's design language, and a trained generation of AI engineers. For $10 billion, that is the bargain of the century.
The Apple Car never drove anywhere. It did not need to. It was carrying the M7 the whole time.
Comments