Meta and Apple bring large Muse AI model to Mac

By Splitfeed EditorialLast updated Oct 15, 2026, 18:40 UTC

On October 15, 2026, Meta released an open AI model called Muse Spark, paired with a partnership with Apple to run it on Mac Studio hardware. The model demands 1TB of memory, placing it at the extreme end of what consumer-grade machines can handle locally. The collaboration is aimed at developers who want to run large models on their own hardware rather than routing inference through cloud services.

The release carries a tension worth noting. Meta made Muse Spark openly available while simultaneously keeping its most capable AI technology proprietary and closed. That distinction matters because it shapes how the open-source AI community and competitors can engage with what Meta is actually building. Muse Spark may be large by local standards, but it is not necessarily Meta's frontier model, which means the openness on offer here is partial. The Apple partnership is also notable as a business signal, representing an unusual alignment between two companies that have historically operated in separate lanes, one building hardware ecosystems and the other building social and advertising platforms.

No published reactions from the Pro-AI, Anti-AI or Middle Ground camps had appeared by the time this story was filed. The Pro-AI camp would typically treat a release like this as evidence that capable AI is becoming more accessible and decentralized, reducing dependence on a handful of cloud providers. The Anti-AI camp would likely raise concerns about the concentration of power that still exists when only a 1TB Mac Studio can run the model at all, and would scrutinize what Meta means by open when its best technology stays closed. The Middle Ground camp would probably focus on the practical implications for developers and ask whether the hardware requirements make local deployment realistic for most teams.

The argument about what Meta's openness actually means will sharpen as more technical details about Muse Spark emerge, particularly around licensing terms and how its capabilities compare to the closed models Meta is not releasing. Developer adoption numbers and any follow-on announcements from Apple about broader Mac compatibility would also help clarify whether this partnership represents a lasting strategic shift or a limited promotional release.