Since Apple Intelligence launched last fall, the company’s AI efforts have mostly amounted to a string of setbacks. The biggest one? The full suite of Apple Intelligence features still hasn’t rolled out and likely won’t for another nine months. That puts it about a year behind schedule, a kind of delay I haven’t seen in all my years following Apple.
The reason for the holdup is simple but difficult: integrating GPT and building their own models that capitalize on the personal data that makes Apple’s AI opportunity so unique is hard. They know giving they have one of the worlds most trusted brands, they need to get it right, whether it’s their own model or a third-party one, especially when it comes to handling personal data, intent, and privacy.
Until recently, based on Apple’s capex outlook, I had all but ruled out the company offering its LLM to developers. But according to reporting from Bloomberg’s Mark Gurman, it now sounds like Apple will start offering its smaller models, and eventually its full LLM, to developers. The announcement is expected at WWDC. The big unknown is timing of when will developers actually be able to build these models into their apps. My guess is we’ll start to see the smaller models go live late this year.
Shares of AAPL barely moved on the news, likely because investors still see Apple’s AI push as a “show me” story after so many recent delays.
At WWDC 2024, Apple didn’t give developers much in terms of tools for adding AI to apps. Last year’s event was more about the broader Apple Intelligence vision. If Apple does end up releasing its own LLM, it matters for two reasons:
-
It’s key to winning over developers in the long run.
-
It would mark the first real sign of progress in AI since last September — something investors are eager to see.