How Apple's cancelled car project fueled its AI future

Because they needed this real-time processing, Apple invented the Neural Engine. This is a specialized portion of a computer chip designed specifically to handle complex AI math. While this tech was built for a car that never arrived, it became the heart of the iPhone X and later migrated to Mac computers. Now, Apple is looking toward the future of data centers by accelerating the development of the M7 Ultra chip. Expected to arrive in 2027, this powerful chip is intended for server products with the massive memory capacity needed to run the most demanding modern AI.

A standard computer chip is like a general-purpose clerk that tries to do everything, from running your web browser to playing games. But AI tasks, like recognizing your face or writing text, require thousands of simple calculations to happen at the exact same moment. The Neural Engine is a dedicated section of the chip optimized just for this math. By putting this hardware directly inside your phone or laptop, your device can process AI tasks instantly without relying on a remote data center—called the cloud—to do the heavy lifting. This is why features like Face ID work the moment you look at your screen. It also keeps your data private, as your personal information does not need to leave your device to be analyzed.

While much of the industry is focused on massive software models run on internet-based servers, Apple is building a bridge between its local hardware and more powerful infrastructure. The move to develop the M7 Ultra, which will support huge amounts of memory—up to 1.5 terabytes—shows a clear strategy: Apple wants to control the AI experience from the device in your hand all the way to the server rack. This suggests that the future of computing will not just be about clever software, but about which companies control the specialized physical silicon needed to make those models run quickly and privately.