A partner willing to customize around OpenAI's specific system requirements rather than protecting its own AI roadmap.

In my opinion, OpenAI is going to whomever is willing to accommodate their system and product specific needs.

Qualcomm and Apple, and we can add here Samsung, maybe even Google, have their own GenAI interests to protect.

Continuing to evolve the machine learning AI capability in their mobile platform is a vital necessity for MediaTek too, and not just for the application processor, but also for the basement processor to continue to drive the air interface optimization.

And what would an orchestrator actually look like in practice? Say, for example, a message request that routes itself to the optimal channel based on constraints.

I can see myself saying, no, instructing the personal AI residing on whatever device platform may be the case.

I need to send a message to my friend Johnny.

It's for his eyes only and should reach him immediately with delivery confirmation, in which case the orchestrator will analyze the time of day, the known available means of contacting my friend, and combined with the other constraints I placed in my request, and may end up dialing my friend's phone number.