The new semiconductor expands on-device artificial intelligence capabilities with upgraded CPU, GPU, and Neural Engine cores.
On 10 September, Apple advanced its artificial intelligence strategy with the arrival of the A20 Pro chip. Manufactured using TSMC's 2-nanometer process, the processor shifts more artificial intelligence workloads directly onto mobile devices, reducing dependence on cloud data centers.
The A20 Pro features a six-core CPU and a seven-core GPU, delivering up to 40% higher GPU performance than the previous generation. Apple also expanded memory bandwidth by 50% compared to the A19 Pro. For artificial intelligence processing, the chip incorporates a Dual Neural Engine with 32 dedicated cores, which doubles the core count of the previous version.
The company pairs the denser 2-nanometer architecture with a redesigned packaging system and a new vapor chamber. This thermal setup allows the device to sustain intensive computing workloads for longer intervals without throttling performance due to heat buildup.
Apple is using a hybrid architecture where supported tasks run locally to lower latency and improve privacy, while heavier tasks route to Private Cloud Compute. This setup balances battery consumption, processing power, and corporate data control requirements.
Newsletter
Markets in your inbox, weekly
LATAM-focused analysis, investing ideas, and the week in finance.
Keep reading