Let’s be honest—Apple has always played the long game.
Remember when everyone mocked them for skipping 5G? Or for refusing to add USB-C while the rest of the world moved on? Yet, somehow, they always landed on their feet. Not because they were first—but because when they did arrive, they did it with polish, control, and that unmistakable Apple sheen.
But this time… this time feels different.
For the first time in memory, Apple isn’t just late to the party. It’s standing outside, hat in hand, knocking on Google’s door asking for a favor. And not just any favor—they’re reportedly paying $1 billion a year to license a custom version of Google’s Gemini AI to power Siri.
Yes. Apple. Paying Google. For AI.
If that doesn’t make you pause, maybe this will: the deal reportedly includes running Google’s trillion-parameter model on Apple’s private cloud, all while trying to maintain that carefully curated “privacy first” brand image. It’s a delicate dance—one that speaks less to strategy and more to desperation.
So what happened? How did the company that redefined smartphones end up outsourcing the very intelligence it promised would redefine the next decade?
The Promise vs. The Reality
Back in June 2024, Apple unveiled Apple Intelligence—a sweeping vision for on-device AI that would transform Siri from a sleepy voice assistant into a proactive, context-aware companion.
You’d ask, “Summarize unread emails from my boss,” and it would—using only your local data without sending anything to the cloud. You’d ask, “Plan a budget-friendly weekend getaway based on my calendar,” and it would draft options, book your partner’s tickets, and book flights.
Fantastic, right? This was the standard Apple was trying to meet.
But they never did.
By early 2025, Robbie Walker, Apple’s senior director of AI, was quoted saying the project delays were “ugly and embarrassing.” He went on to characterize a company’s decision to announce features that were not ready as “vaporware,” and calling it an “absolute disaster.”
They were not the only ones. Lawsuits for false advertising were growing in number as they were not living up to expectations. Customers expected artificial intelligence for the iPhone 16. All they got was old Siri that still couldn’t play “Smelly Cat” on command, and played “Friends Forever,” a random YouTube clip instead.
The Google assistant on a Pixel 9 Pro XL had no problem. It found the exact song, gave trivia about the show, and did a full spec comparison of phones without asking for anything. It had no problem.
It was humiliating.
The Internal Melt Down No One Saw Coming
Chaos was a certainty.
Apple's AI team was not only struggling, but fragmented. For years, the Siri team and foundational AI team functioned under the thinned metaphorical walls of an empire, with different roadmaps, duplicated efforts, and, most importantly, no communication. One feature competitors had moved past years ago—removing the "Hey Siri" wake phrase—took two years to remove.
Having to retrofit generative AI into a system designed for deterministic, rule-based responses resulted in chaos. Bugs and unreliable outputs legally and socially branded the company as Apple Unreliable.
The jargon focus shifted to 'Key Talent' as Key Talent began to exit the company. Apple’s 100-person LLM team was led by Ruoming Pang, who joined the developer team at Meta. The shift in employees was not for the 'better' offer. It was a ship, a sinking ship of morale and likability.
Blame? The corporate kind. AI engineers lost the marketing team. 'Overpromising' was their crime. The marketing team, as the legal loophole claimed, were 'given timelines by you'.
With internal efforts floundering and shareholder pressure mounting, Apple explored offerings.
Anthropic was likely too expensive. Microsoft’s too intertwined with Windows. OpenAI is too associated with Microsoft and owns a Windows app. Also philosophically too far apart on privacy with Apple.
So the last option was Google.
Apple’s custom Gemini model will run on Apple’s private cloud infrastructure. Apple will uphold user privacy with the model rumored to be 1.2 trillion parameters. It’s Apple leasing the brain without the body. Google will not be able to access the data.
Google pays Apple 18 billion dollars a year. Apple pays Google 1 billion a year for AI. It’s more than irony. It’s like a couple of old rivals realizing they need each other more than they are willing to admit.
Apple will not tell anyone Gemini is in their SYNCED servers or privately on cloud. No, “powered by Google AI.” Just a Siri who will finally work without a connection.
Do Consumers Want AI on Phones?
This is a difficult, uncomfortable truth. According to a CNET survey from early 2025, only 11% of U.S. smartphone users upgraded their devices because of AI features. This is a 7% drop from the year before the survey.
Even more telling? It’s hard to ignore the 30% who don't find mobile AI useful and don't want more of it.
Samsung learned the hard way. They overestimated the value of Galaxy AI and thought most consumers would upgrade their devices. Instead, their earnings calls reported “market weakness” and “economic headwinds.” It turns out, customers value battery life, camera quality, and the price over the tools their phone has to summarize a PDF.
So, why do most customers Apple devices say kr Apple?
Not having AI systems integrated in devices is a liability. To investors and tech press, having a phone without smart capabilities is dated, even if customers do not want to use it. It’s like having a car without Bluetooth in 2015: technically functional, but ridiculous.
Apple is not focused on customer delight or even on improving Siri, which is barely functional. Their objective is to buy time.
The Bigger Picture: The Philosophical Side of AI.
For quite a while, we’ve been told that every major tech company has to develop its own AI, as if we are in some kind of arms race, and that if they are not spending billions on data centers, chips, and personnel, they are falling behind.
However, Apple’s recent strategy seems to point to a different conclusion: perhaps it’s not necessary.
Consider a smartphone structured in three layers:
- Hardware (the physical device) → analogous to AI data centers.
- Operating System (iOS or Android) → comparable to foundational AI models, like Gemini, GPT, or Claude.
- Apps (Instagram, WhatsApp) → provide AI-powered features like Siri, photo editing, and writing tools.
For a long time, Apple built the first two layers. But when it comes to AI, perhaps the company now sees the OS component, the LLM, as something that has become a basic commodity. Not in quality, but in accessibility. If Google, Anthropic, and OpenAI offer enterprise-grade models through APIs, where’s the sense in spending $50 billion on something that has been offered at much lower prices?
This is not a sign of weakness, but perhaps a sign of wisdom.
Apple has a long history of focusing on user experience rather than the technical bragging rights. If they can obtain a top-tier model and adapt it with a focus on privacy so that it runs on-device, they can provide seamless performance, and that’s something to be proud of.
Samsung did the same thing with Galaxy AI, also using Gemini under the hood. Instead of building their own LLM, they built a better application of someone else’s intelligence.
And possibly the future will not be about the biggest models but about the best usage of a model.
No one knows the best answer about “privacy,” “trust,” “tightrope,” or any other of those unethical words that will get you extensions if you say and make it your business ethics, and anyway, it won’t difference if Apple covered one of the underlining elements of the business, it still loses the distinguishing factor.
That’s why the on-premise detail of the cloud is important since Apple doesn't send your sensitive voice data to Google’s storage. It is a compromise, but one that is defensible, non-invulnerable.
There will always be someone to “cry hypocrisy.” “You built your brand on not being Google,” they will say, and they are not wrong.
This, as they say, is a “take deal.” You can always “patch the hole.” And you will keep on “building.” Apple suffers and loses its customers because it has to suffer a plateau without an on-device trillion or expensive parameters.LLM. Since it’s 2023, customers get a broken assistant from Apple as they wait. Competitors overtake Apple and it has to suffer a broken assistant.
Let’s not mistake Apple’s collapse as it’s own.
Sales of the iPhone are up 29% worldwide annually, especially from China. All iPhone 17 units are being sold. Due to Windows 11’s intrusive updates, user logins, and telemetry, frustrates and is forcing customers to leave, and so if apples.(And yes, I too have worked on Windows, and I can feel that pain as well.)
Tim Cook does not lose sleep over these “issues”. Apple’s cash reserves can fund a small country. This is not a crisis for Apple, it is a normal correction in their business strategy.
This is a signal, a quiet acknowledgment that in the era of AI, control does not always mean creation. Sometimes it means knowing whom to partner with, when to wait, or when to swallow your pride for the benefit of your users.
If Apple, with all of their cash, talent, and engineering, and then some, is not building their own frontier AI model, what does that say for the other thousands of startups with VC money building their own LLM?
The real innovation is not in the model, it is in the application. In the UX, the privacy guardrails, and the human-centered design.
Look at it this way, technology is not about who has the biggest engine. It is about who builds the car that people actually want to drive.
If Apple really gains from Google’s AI Labs, even if it’s not entirely theirs, then maybe, just maybe, it’s not a retreat at all.
Maybe it’s a first step towards a more advanced type of intelligence. The kind that understands its boundaries and knows when it needs assistance.
0 Comments