New M4 iPad Air Officially Launched
Apple has officially launched the new iPad Air with the M4 chip, starting at $599. The device boasts a faster Neural Engine for on-device AI tasks, up to 30% faster CPU performance than the M3, and 50% more unified memory bandwidth. Early benchmarks and reviews highlight desktop-class performance for machine learning workloads, making it a powerful target for AI-driven apps.
The M4 chip is manufactured using TSMC's second-generation 3nm process, packing 28 billion transistors. This iteration moves to the ARMv9.2-A instruction set architecture, featuring improved branch prediction and wider decode/execution engines for its performance cores. A core focus of the M4 is machine learning, with its 16-core Neural Engine capable of 38 trillion operations per second (TOPS). This is more than double the 18 TOPS of the M3's Neural Engine and a significant leap from the M2's 15.8 TOPS, enabling more complex on-device AI models. The iPad Air's specific M4 configuration is a binned variant with an 8-core CPU (3 performance, 5 efficiency) and a 9-core GPU. This differs from the top-tier M4 in the iPad Pro, which can have up to 10 CPU cores (4 performance, 6 efficiency). Unified memory in the new iPad Air sees a 50% increase to 12GB, coupled with 120GB/s of memory bandwidth. This larger memory pool is critical for reducing model swapping in AI applications and handling larger assets in creative and development workflows. Connectivity gets a future-proofing upgrade with the inclusion of Apple's N1 wireless chip, enabling Wi-Fi 7 support. Cellular models incorporate the custom C1X modem, which provides faster 5G speeds with lower energy consumption. The 9-core GPU architecture brings hardware-accelerated ray tracing and mesh shading to the iPad Air for the first time. This allows for more realistic lighting and complex geometries in graphics-intensive developer tools and applications.
