Apple's abandoned electric car may be gone, but the engineering behind it isn't going to waste.
According to TechRepublic, Apple's failed car project reportedly shaped the company's Neural Engine, its future M7 chips, and its plans for more powerful AI server infrastructure. In other words, work that was meant to help a vehicle drive itself is now being redirected toward the silicon that powers Apple's artificial intelligence ambitions.
That pivot is showing up in the chips themselves. Tom's Hardware reports that Apple's rumored M7 Ultra is being designed to support up to 1.5 terabytes of unified memory and to push AI performance toward the class of Nvidia's Blackwell accelerators — the high-end hardware that currently powers much of the AI industry's heavy lifting.
There's a big catch. Tom's Hardware describes the M7 Ultra as a 2028 offering whose arrival would depend on the current memory shortage easing. Memory is one of the most supply-constrained parts of the AI hardware boom, so even an ambitious design can stall if the components aren't available at scale.
It's worth stressing that these are reports and rumors, not official Apple announcements. The details come from claims cited by Tom's Hardware and TechRepublic rather than confirmed product plans.
Why it matters: if the reporting holds, it shows Apple trying to build serious in-house AI muscle to rival Nvidia — and reusing years of canceled car research to get there, rather than letting that investment disappear.