Apple and Google – Desperate Times

Apple does a deal with the devil.

• Using Google as the foundation of its AI is far more consequential than search because if AI becomes a point of differentiation, then Apple has just ceded any edge that it had to Google.
• It is also an indication of the size of the hole that Apple finds itself in, as AI is the last thing that it should be relying on a competitor to provide.
• Apple and Google have confirmed a multi-year deal (cancellable presumably) with Apple effectively stating that Google has the best models available today.
• No financial terms were disclosed, but I suspect that Apple will pay Google a couple of billion a year to have access to its latest models and the tools that it needs to tune them to its own requirements.
• Given what Apple has promised to date, I think that Apple will now take the Gemini foundation models and use them to replace the models that are currently powering Apple Intelligence, while the infrastructure private cloud compute will remain largely unchanged.
• In effect, Apple is in the pits for an engine change, but the car itself will remain the same.
• It is also likely that the signing of this deal will kick off a period of fine-tuning where Gemini’s base functionality is tuned to Apple’s requirements.
• The result will be Gemini with an Apple wrapper, and while this will help fix Apple’s short-term problems, it sets up a dependency on a competitor that could be highly problematic.
• If the glitterati of AI are to be believed, then fairly soon, we are going to be using agents to facilitate the humdrum digital tasks of our daily lives.
• This means that between the traditional user experience and the user will sit an agent, and with Apple dependent on Google, there is no way that its agent will be better than Google’s.
• Furthermore, Google will be incentivised to ensure that Android gets the latest AI features before Apple does, meaning that there could easily be a situation where Apple’s AI is worse than what is available on Android.
• With users spending more of their time with agents, they are certain to start to question what is so special about the Apple user experience and why they are paying a premium price for the Apple products.
• This is why it is imperative that Apple creates (or buys) its own AI and gets off Google as soon as it possibly can.
• Failure to do this could result in gross margin and brand erosion and market share loss to high-end Android products, where Samsung and Qualcomm would be the main beneficiaries.
• Although I see this as a real risk, I do not think that it is close to materialising as AI currently has no impact on the user’s decision on where to live his or her digital life and what smartphone to buy.
• Hence, for the next year at least, I do not see this risk being realised and I think agent-based user experiences will arrive first on smart glasses and automobiles.
• Apple is not strong in either of these categories, and so I think it has time to bring its competency up or acquire one of the other players.
• This is not a respite, and if I were at Apple, I would be looking to redouble my efforts to achieve AI independence as soon as possible, as the consequences of failure could be dire.