A lot of speculation has popped up online already about what this could mean for the PlayStation, but what it probably doesn’t mean is a significant change to the existing PS5. It’s too soon to expect Sony or Microsoft to debut the equivalent “Pro” models for their platforms; the PS4 Pro and Xbox One X appeared three years into their console life spans. In late 2022, the PS5 will just be turning one.
The 6nm node this new PS5 is expected to target offers up to an 18 percent density improvement compared with previous 7nm technology, but no claimed improvement to clock speed or power consumption. Oberon, the CPU inside the PS5, is a 308mm2 SoC, compared with 360mm2 for the Xbox Series X. Oberon is still smaller than either the PS4 Pro (322mm2) or the PS4 (348mm2).
The only reason for Sony to be respinning the SoC this quickly is if it’s trying to cut costs. Analysis has suggested that only the full-fat version of the PS5 actually breaks even at sale; the PS5 Digital Edition may currently be sold below cost. This may explain the entirety of why Sony is pushing for a redesign in the first place. A modest reduction in PS5 die size would allow them to build more dies per wafer.
Given that TSMC doesn’t expect any power consumption or performance improvements, we’re assuming the updated variant wouldn’t offer any. We might see a very small single-digit bump in performance depending on the exact characteristics of the silicon, but console manufacturers who use die shrinks for this kind of shrink typically go to some trouble to ensure the experience is as identical as possible.
Sony has noted that the PlayStation 5 hardware is profitable, but takes care to note that such profitability is “inclusive of peripherals.” The company’s last quarterly earnings call referenced upcoming improvements to PS5 profitability in its second year, and a modest die shrink would do that. A 10-percent die size reduction would cut the size of a PS5 SoC to ~278mm2. A simplistic wafer calculator suggests this could boost Sony’s total recoverable dies per wafer (assuming equal defect density) from ~177 to ~200, with the number of good dies estimated at 152, up from 131. These values are mathematical estimates based on modeling, but they give some idea of what kind of improvement Sony might see if it shrinks the PS5 die.
As for whether the die shrink would improve availability, that depends entirely on which components are slowing production. Bottlenecks in areas other than wafer production have emerged across the industry in everything from display drivers to Wi-Fi chips. If wafer yields are slowing Sony’s PS5 production, boosting it with smaller chips might help assuming equal yield. If other low-level components are holding back shipments, the die shrink won’t help much.