Apple Unveils M5 Pro and M5 Max Chips for On-Device AI
Apple just unveiled its new M5 Pro and M5 Max chips, designed for a major leap in on-device AI performance. The company is touting massive boosts for local large language model (LLM) execution, including 8x faster image generation and 6.7x faster LLM performance, positioning its new hardware for the next wave of AI applications.
This push into on-device AI places Apple in direct competition with a rapidly crowding field. Qualcomm's Snapdragon X Elite boasts a 45 TOPS (Trillions of Operations per Second) Neural Processing Unit (NPU). Meanwhile, AMD's Ryzen AI 9 HX 370 touts up to 50 TOPS from its NPU, and Intel's Core Ultra processors feature their first integrated NPUs, all aimed at accelerating local AI workloads. The focus on local processing marks a significant strategic shift from cloud-based AI. By handling tasks on the device, companies can offer enhanced user privacy and reduce the costs associated with data center processing and transmission. This approach also minimizes latency, enabling the real-time responsiveness required for next-generation AI applications. At the core of Apple's on-device AI strategy is the Neural Engine, a specialized component of its chips. The M1 chip's 16-core Neural Engine was a major step, and the M2's offered a 40% performance increase. The M4 chip's Neural Engine is capable of 38 trillion operations per second, a significant jump from the M3's 18 trillion, laying the groundwork for the M5's focus. For the business and creative sectors, these performance gains are critical. The M4 generation already brought significant speed boosts, with the M4 Pro being twice as fast as the M2 Pro in some tasks. The M5's claimed 8x faster image generation and 6.7x faster LLM performance point to a future of more complex, instantaneous AI-powered creative and productivity tools running natively on MacBooks and iPads.
