Apple and Meta are collaborating on a new AI model

By Splitfeed EditorialLast updated Oct 16, 2026, 12:01 UTC

On October 16, 2026, three separate outlets, Semafor, CNBC, and the Wall Street Journal, reported that Apple and Meta are collaborating on a new AI model. The Wall Street Journal added Nvidia to the picture, describing all three companies unveiling new open models together. Beyond the fact of the collaboration and its open-model framing, no specifics about the model's architecture, capabilities, release date, or licensing terms have been published.

The significance here lies largely in the companies involved and what their partnership implies about the direction of the AI industry. Apple and Meta have spent years publicly clashing over privacy, data collection, and platform control, most visibly during the App Tracking Transparency dispute. That two companies with such a documented adversarial history are now co-developing an AI model signals how much competitive pressure the AI race is generating. The open-model framing is also notable: if accurate, it would mean Apple, historically protective of its ecosystem, is moving toward a more permissive release strategy, at least in this context. Whether the model is truly open-source or open in a more limited sense remains unresolved.

Because no camp reactions have been published yet, what each group would typically argue can only be anticipated. The Pro-AI camp would likely read this as a positive sign that major players are pooling resources and releasing models openly, which could accelerate access and capability for developers worldwide. The Anti-AI camp would be expected to raise concerns about the concentration of power among three of the largest technology companies and question what user data or platform leverage underpins the collaboration. The Middle Ground camp would probably welcome openness in principle while pressing for transparency about training data, safety evaluations, and what conditions govern the model's use.

The details most worth tracking are the actual release of the model, any accompanying technical documentation or model card, and the specific licensing terms, since those will determine whether the open framing holds up and how the collaboration affects the broader competitive landscape among AI developers.