Apple M5 Pro chipset will feature separate CPU and GPU for server-grade performance
Unlike previous SoCs, Apple is reportedly planning to separate the CPU and GPU in its upcoming M-series chips to take advantage of TSMC’s latest chip-packaging technology, known as SoIC-mH Is.
listen to the story

Apple is known for its system-on-a-chip (SoC) design, which is the main feature of its A-series and M-series chips. This design integrates major components such as the CPU (central processing unit) and GPU (graphics processing unit) into a single chip. Apple has improved performance and power efficiency across all of its devices by combining these components and tuning them together. However, a recent report from renowned Apple analyst Ming-Chi Kuo suggests a major change with the upcoming M5 Pro chip. Apple may move away from its fully integrated SoC design and instead adopt a new approach that separates the CPU and GPU.
This change is reportedly being made for two major reasons – better performance and better production yields. By separating the CPU and GPU, Apple can take advantage of TSMC’s latest chip-packaging technology, known as SoIC-mH (System-on-Integrated-Chips-Molding-Horizontal). This advanced method of packaging the chipset is said to significantly improve thermal performance, meaning the chip can run at full power for longer periods of time without overheating. Additionally, the SoIC-mH process reportedly reduces manufacturing defects, resulting in higher quality chips and less production loss.
According to Kuo, this new design will be used not only in M5 Pro but also in M5 Max and M5 Ultra variants. These high-end chips will be built on TSMC’s N3P node, which is an advanced chip-manufacturing process. Mass production timelines suggest the base M5 chip will launch in the first half of 2025, followed by the M5 Pro and M5 Max later that year, and the M5 Ultra in 2026.
Interestingly, the report also links this development to Apple’s artificial intelligence (AI) plans. The M5 Pro chip is expected to power the Apple Intelligence Server, which is part of the company’s Private Cloud Compute (PCC) infrastructure. These servers are designed for AI inference, which is a critical task for processing AI-powered tasks. Once these high-end chips are ready, Apple plans to enhance its AI infrastructure, furthering its ambitions in the AI ​​field.
This potential change in design is a significant deviation from Apple’s usual approach. Although Apple hasn’t confirmed these plans, rumors suggest the company wants to adapt and innovate over time. It also highlights the increasing complexity of balancing performance, efficiency and production challenges in the tech industry.