Reviews of Apple's latest M6 System on a Chip (SoC) are emerging, showcasing impressive performance that, while not directly benefiting PC gamers due to platform limitations, carries significant implications for the future of PC processor technology. The M6 is built on TSMC's cutting-edge N2 silicon node, a 2nm class process.
Historically, advancements in silicon manufacturing nodes reliably delivered improvements in transistor density, power efficiency, and cost. However, Moore's Law has faltered in recent years, with companies like Intel experiencing prolonged struggles on older process nodes. The successful implementation of TSMC's N2 node in the M6 chip signals a potential return to substantial gains from new manufacturing processes.
Performance benchmarks highlight the M6's capabilities. According to Notebookcheck, the M6 achieved a single-core score of 4,698 in Geekbench 6.7. This significantly surpasses the 2,872 points scored by AMD's Ryzen AI 9 HX 470, which is among AMD's top laptop CPUs. The M6 also shows a roughly 10% improvement over the previous Apple M5 chip in this benchmark. Similar advantages are observed in Cinebench tests.
In multi-threaded workloads, the M6, equipped with 12 CPU cores compared to the M5's 10, offers approximately 50% greater overall CPU performance. Furthermore, the M6 demonstrates nearly 20% better performance per watt than the M5 and is twice as power-efficient as the AMD Ryzen AI 9 HX 470. These gains underscore the health and potential of TSMC's N2 silicon.
The critical question for PC users is when this advanced silicon will power their hardware. The ongoing demand for AI-focused chips may influence allocation of the scarce N2 manufacturing capacity. AMD has announced that its server variants of the next-generation Zen 6 CPUs are already in production on N2, suggesting that PC versions of Zen 6 might also utilize this node, potentially bypassing TSMC's N3 process.
Nvidia's current Blackwell graphics chips are on N4, with upcoming Rubin GPUs slated for N3, and an N2 GPU potentially not arriving before 2030. Intel, already using TSMC's N3 for Lunar Lake, may employ N2 for some variants of its upcoming Nova Lake CPU family, expected early next year.
While PCs are unlikely to be flooded with N2 silicon in the immediate future, its arrival seems inevitable. The performance and efficiency demonstrated by the M6 suggest that next-generation PC chips built on this node will offer substantial improvements. A continued wait might even allow for a stabilization of chip prices.