June 26, 2024 – According to a tweet published by the source @MappleGold on June 25th, Apple’s upcoming A18 chip features a neural engine that is even more powerful than the one found in the M4 chip. This suggests that the A18 will be capable of running AI models locally, enhancing its ability to handle various AI tasks.
Looking into Apple’s official press releases, we learn that the M4 chip is equipped with Apple’s most advanced neural network engine yet, boasting a computational speed of up to 38 trillion operations per second. This marks a significant improvement compared to the initial neural network engine in the A11 Bionic chip, with a speed increase of up to an astonishing 60 times.
If the information provided by @MappleGold is accurate, it implies that the neural network engine in the Apple A18 chip could potentially exceed the computational speed of 38 trillion operations per second.
The source also hinted that Apple might adopt a different marketing strategy for the A18 chip, possibly steering away from the conventional “powerful” and “highly efficient” narratives seen in previous SoC launches.
Furthermore, the performance enhancements in future Apple chips, such as the A19 or A20 series, will largely depend on Taiwan Semiconductor Manufacturing Company’s (TSMC) 2nm process technology. Apple is expected to leverage this advanced manufacturing process to further elevate processor performance in its future chip generations.
