A new theory suggests that Apple’s M3 Ultra chip might be designed as a chip rather than being made up of two M3 Max dies. This idea comes from Vadim Yuryev of Max Tech, who discussed it in a post on X. Yuryev referenced a tweet by @techanalye1 indicating that the M3 Max chip may no longer include the UltraFusion interconnect. He speculated that the upcoming “M3 Ultra” chip may not consist of two Max chips within one package. This could mean that the M3 Ultra will operate as a chip for the time.
More About the M3 Max Chip
If this theory pans out Apple could have the flexibility to customize the M3 Ultra for tasks tailoring it to meet demanding needs. For example, Apple might choose to remove efficiency cores and focus on performance cores while increasing the number of GPU cores. Consequently, a single M3 Ultra chip designed in this way could offer performance scalability compared to the M2 Ultra when compared to the M2 Max since there would be no efficiency losses related to the UltraFusion interconnect.
Furthermore, Yuryev suggested that the M3 Ultra could possibly include its UltraFusion connection making it easier to combine two M3 Ultra units into one package to double the performance capabilities of an M3 Extreme” chip. This strategy would provide better performance scalability than combining four M3 Max units and might also support memory capacities.
Although limited information is currently available about the M3 Ultra chip, a report from January hinted that it would be manufactured using TSMC’s N3E node, akin to the A18 chip anticipated to debut in the iPhone 16 lineup later this year. Consequently, the M3 Ultra would mark Apple’s inaugural foray into the realm of N3E chips. The M3 Ultra is rumored to make its debut in a forthcoming release.
