Apple's car flop gave birth to M7 Ultra AI chip with 1.5TB RAM

Apple’s ambitious self-driving car project never hit the road, but it left a powerful legacy: the company’s AI chips. Early in development, Apple realized it needed massive on-device AI processing for autonomous driving. Though the car processor was never finished, that effort directly led to the creation of the Neural Engine, debuting in 2017 with the iPhone X’s A11 Bionic. Originally used for computer vision tasks like FaceID, Animoji, and AR, the Neural Engine later became the backbone of Apple’s M-series chips, bringing on-device AI to desktops and establishing Apple as an early hardware leader. This foundation also allowed Apple to emphasize privacy by minimizing cloud data transfer.

Now Apple is doubling down on AI hardware. According to Mark Gurman, the company is skipping the Pro, Max, and Ultra variants of the upcoming M6 chip to accelerate development of the M7 lineup, expected in the first half of 2027. The M7 Ultra will feature major Neural Engine upgrades and support up to 1.5TB of RAM, making it suitable not only for Macs but also as the basis for a new Apple server product. This move signals Apple’s intent to compete more aggressively in AI infrastructure while maintaining its privacy-first approach. Though Apple’s AI software efforts have lagged behind competitors, its hardware trajectory could shift the balance in the on-device vs. cloud AI debate.