M5 Max Benchmarks Leak, Rivaling Nvidia's 5080

The M5 Max's 18-core CPU not only surpasses its 16-core M4 Max predecessor but also outperforms the 32-core M3 Ultra in some multi-thread workloads. Geekbench 6 results show the M5 Max reaching multi-core scores as high as 29,644, slightly edging out the M3 Ultra's typical scores. This performance leap is achieved despite having fewer cores, highlighting significant architectural efficiencies. On the graphics front, the M5 Max's integrated GPU scored up to 232,718 in Geekbench's Metal benchmark. This places its compute performance ahead of Nvidia's discrete GeForce RTX 5070 and competitive with the RTX 5070 Ti. While not a direct match for the top-tier RTX 5090, achieving this level of performance in an integrated, power-efficient chip is a notable engineering feat. This hardware advancement is directly tied to Apple's strategy of enhancing on-device AI capabilities, ensuring user privacy by processing data locally. The powerful Neural Engine, a core component of Apple Silicon, is designed to execute machine learning algorithms efficiently, minimizing latency for AI-driven features. This approach contrasts with cloud-reliant AI strategies, offering a significant advantage in responsiveness and data security. The foundation for these gains lies in TSMC's advanced manufacturing processes. Apple is expected to be the first company to utilize TSMC's 2nm process node, which will feature new Gate-all-around (GAA) transistors for reduced power leakage and increased performance. Securing a significant portion of TSMC's initial 2nm capacity ensures Apple maintains its hardware lead for future devices, including the iPhone 18 and subsequent M-series chips.