A project that meets professional expectations: Apple’s M4 Extreme chip promised to revolutionize computing performance, based on an innovative architecture composed of four interconnected chips. Destined for Mac Pro models, they will have to meet the most resource-intensive needs of professionals in graphic creation, video editing, or scientific computing. It promised enticing specs: up to 64 CPU cores and 160 GPU cores. It would have offered, quite simply, power that nobody had ever seen before. But these ambitions came square into the middle of big technical and logistical roadblocks that made this project too complex and expensive to execute in the short term.
A Strategic but Costly Cancellation
According to information obtained by The Information, Apple has decided to remove resources from its work in the M4 Extreme chip in order to work on an AI server chip. The strategic decision could thus allow it to compete with players like Nvidia and Google, which currently top this booming market. But this comes at a cost: giving up the M4 Extreme chip risks disappointing professional users and makes Mac Pros even less popular, since they are having a hard time finding their place in people’s minds compared to the Mac Studio. This about-face nevertheless shows the increased place AI is taking in Apple’s strategy.
Historically, the Mac Pro has always been the presentation of Apple’s know-how, dedicated to a professional public looking for power and modularity. Take away the M4 Extreme chip, and this premium variant loses its appeal. Of course, the new Mac Pro should come with an M4 Ultra chip, just like the Mac Studio, if the line between the two products isn’t to become blurred. And that would weaken the platform the Mac Pro is standing on, considering it was already panned when refreshed for changing too little. While users had been dreaming about a machine capable of holding a candle to the most powerful workstations, this ambition seems now compromised.
The future of Extreme chips at Apple
If the M4 Extreme were to be abandoned, this would not absolutely bury the idea of an ultra-high-performance chip at Apple. The brand has already shown several interests in this type of project, as evidenced by the abortive attempt of the M2 Extreme. If the current challenges are too great, there is nothing to rule out that the M5 series or another future generation could integrate an “Extreme” version. Meanwhile, Apple is making its bet on the development of a chip for artificial intelligence servers, an area which, at some point, might turn as strategic as Macs to their future. This choice is typical of the balance the company has to find between innovation, technical feasibility, and strategic priority.
