Why Apple’s $1 Billion Deal With Google Proves Siri Was Always Broken
I spent two weeks comparing AI models and discovered the uncomfortable truth about big tech’s innovation problem.
A genuine thunderclap just hit the tech world: Apple and Google, yes, the two giants, just announced they’re joining forces on artificial intelligence.
AI is now part of our daily conversations, at the heart of our connected lives, and whether you’re terrified or fascinated, this will certainly take up more and more space in our societies.
So what are the stakes and affects for Apple, but also Google, as well as other AI heavyweights like OpenAI, the company behind ChatGPT? But above all, what does this mean for us, the users?
Get comfortable. I have a lot to explain and decode.
The Partnership I Saw Coming a Few Months Ago
To understand what just happened a few hours or days ago, we need to go to the source. And the source is this joint statement uniting Google and Apple. Here’s what it says:
This is not a small deal at all.
And you’ll see with all the data I’m going to share that, an extremely important page is being written.
First, I think for over a year now, many have judged that Apple was behind, including myself. Behind on AI, on what they announced with Apple Intelligence, because AI is artificial intelligence, it was rather clever on their part. And indeed, Siri is totally failing at the moment I’m talking to you.
Personally, I use an iPhone. Siri is completely broken. Yes, Apple Intelligence plans announced in 2024 were postponed by over a year. Yes, there were also departures of several artificial intelligence executives within the company. All that’s true.
But never forget that Apple has a genuine war chest. And that war chest is its iOS and Mac device park. We’re talking about approximately 2.35 billion active iOS devices. And that really allows us to see the impact, influence, and weight that Apple carries today.
That many devices constitute the entry points for AI. Then, also remember that Apple hasn’t unnecessarily burned financial resources and, in reality, will be less subject to turbulence on the AI bubble if it were to burst or at least deflate somewhat.
Apple started by associating with OpenAI initially some time ago with ChatGPT, and now turns toward Google, which has taken the technological lead.
Apple Dumps Siri for Google’s Gemini — What This Deal Really Means
Why Apple is outsourcing its voice brain, and how this will reshape Siri, iPhone UX, and privacy promisesmedium.com
The $20 Billion Context Nobody Mentions
For memory, remember that Google already pays nearly $20 billion per year to Apple for Google to be the default search engine on iOS devices. I don’t know if people realize that $20 billion is not a small sum.
And in this new deal, whose exact contours we don’t know, but insiders from the tech world mention that Apple would pay roughly $1 billion for using the Gemini model to power its new Siri.
So, for Apple, in reality, this would be a hell of a good deal.
And the union between Google and Apple in terms of AI could really create something highly potent.
Ultimately, we need to understand what this involves. Actually, Gemini, which is Google’s artificial intelligence model, will be at the heart of the new Siri, but it won’t be visible to the end user. There won’t be like “Gemini speaking to you” and that, no.
It will be a model with approximately 1.2 trillion parameters that will bring Siri 2026 to life. Siri 2026, by the way, should probably arrive with the iOS 26.4 update expected for spring, and there it will be — a Siri that’s totally redesigned and should finally show its muscles.
Because this model will run on Apple’s servers while preserving users’ raw data to respect the privacy of iPhone, iPad, and Mac owners. The company has always been sticking to this.
We could also say that Apple has been improving its chips for months for this internal AI computing leap. Because in a way, one paradigm many hope for is that eventually AI will be as reactive and fast as possible, but above all, as centralized at the very heart of your device without necessarily needing to connect to the internet.
I spent two weeks testing both AI models last summer.
Before the heart of this deal and decoding it, there have been many reactions in the space of a few hours, and the competition couldn’t stay idle. Among them is Elon Musk, who’s also in the AI race with Grok.
Here’s Elon Musk’s reaction to the Google and Apple news:
And it’s true that when there’s a lot of talk about technological monopolies, because these companies, OpenAI, Nvidia, Apple, Microsoft, and the company, today have absolutely stratospheric economic power, but also a power I’d say almost social and therefore somewhat political.
Because when you’re at the heart of the connected lives of billions of people, with the tools you develop and the biases you can integrate into your software, that obviously raises proper questions.
Now, let me offer you a sort of experience feedback. I genuinely benchmarked, meaning tested in depth, both ChatGPT and Gemini, which are today the two main public AI models. Each time in the paid versions, to have the most precise, sharpest perspective on each one’s strengths and weaknesses.
Here are my findings.
I’ll start with Gemini because we’re talking about it specifically. I believe it seems clear now. Each time, I made uncommon requests and compared them between ChatGPT, Gemini, etc., so Gemini Advanced and ChatGPT o1 specifically.
And clearly, I can say that most of the time, not always, but most of the time, Gemini’s responses were superior in quality, relevance, and I’d even say the way of structuring thought.
And indeed, to be very clear, and this also comes with the paid models of these AI models, I was in reasoning mode, meaning not standard mode or just conversational mode, which I’ll come back to in a few moments, which are certainly much faster but don’t go to the heart of things.
When we switch to reasoning mode, it takes a bit more time, but we get a more reasoned analysis of the questions we can ask. And on that, clearly, Gemini had taken the lead over ChatGPT. It’s what emerges overall from all the very technical tests, but also in terms of personal feeling.
Gemini also leads in the quality I find of Spanish or French voice synthesis, which has less of an American accent than ChatGPT’s, which is very nice but has an extremely pronounced accent.
However, in a more public-facing way, I find ChatGPT o1 definitely hasn’t said its last word.
Because I find ChatGPT keeps a huge lead in usability comfort. I’m not even talking about the interface and everything because that’s tastes and colors, but it’s, for example, on voice capture.
I used these AI models enormously with voice, meaning without having to type a long prompt or long question, but extremely naturally by dictating my question. Once again, I specify, not necessarily in discussion mode but by dictating my questions, and sometimes I asked questions that were blocks of text, long reflections, etc.
And on that, there’s no doubt, Gemini was completely off, totally off.
And it really amazed me to see how much Gemini was better in its responses, but not very relevant in the way of capturing voice. So I did it sometimes outside, so you can say there’s a bit of noise, etc., and indoors, well, each time one word out of four or five is completely wrong.
Is the iPhone Falling Behind? Android’s AI Leap Explained.
This article explores Google’s strategic AI advantage over Apple, showing how Google’s AI integration surpasses…ai.gopubby.com
As soon as you start to lower your voice, it launches the prompt, but you haven’t finished, so you’re like “no wait, I haven’t finished” etc., whereas ChatGPT, in contrast, is amazing, amazing, whatever the condition outside or inside. Frankly, 99% all the words, the vocabulary, everything was perfect. It captured what I was saying extremely naturally, so I didn’t have time to waste correcting after what I wanted to say.
On that, frankly, there’s an enormous lead, and I found that on the fluidity of the voice as well, it was better there, too, on ChatGPT. The voice has a minor accent, which is actually quite pleasant.
I believe ChatGPT is better for routine inquiries. I’m talking about fairly mainstream requests and not very high-level scientific stuff where unanimously Gemini has the lead, but for everyday use, I don’t know, you want to do your shopping, you want to calculate your kilocalories, you have some professional or business questions etc., frankly both were very good but Gemini a tiny better but less well on the delivery.
The Power Concentration Problem
So, in early 2026, it’s most interesting to see that there’s still a match and that, in a way, it’s extremely intellectually stimulating because each one trades blows.
ChatGPT and, therefore, OpenAI, really led the dance for many months, and still today it’s OpenAI that dominates in terms of number of users across the planet. Sam Altman, who became a personality in the tech world in the space of just a few months and heads OpenAI, can boast about that.
However, we can notice these last months a sort of exhaustion, and Gemini is gaining more and more importance, helped by Google’s results. That means the technical results are better for Gemini.
And in terms of the number of users, Gemini will be in the background, the engine that will command iOS and Apple Intelligence, which will rise.
Because I think there’s one point we really mustn’t look away from: Apple will catch up almost instantaneously on its technological delay to date. It will offer totally transparent integration for billions of users who don’t care at all about what’s under the hood.
Whether it’s Gemini, ChatGPT, Claude, or others, they don’t care. The important thing is that it works on the device you own.
And on that, it’s going to be extremely interesting to see this with Apple’s user interface overlay, which is very good at this level, with a Siri that until now was catastrophic, I tell you as an Apple user, it really just served me to set my alarm and stuff because everything else is really failing every time.
You’re like, “Why are you so bad? Why are you so bad?”
Well, maybe in a few months, they’ll finally be back at the level. It’s going to be extremely interesting to see OpenAI’s response and others’ and see how all this will grow.
What This Actually Means For You
Here’s what nobody’s saying clearly: this partnership doesn’t change your AI experience in the short term. It changes who profits from it.
Currently, to access good AI on your iPhone, you need to open the ChatGPT or Gemini apps directly. Come spring 2026, you’ll use Siri powered by Gemini without knowing it. Your experience might improve, but Google gets billions of user interactions to train its models, Apple gets to claim competitive AI without building it themselves, and you get… a slightly better voice assistant.
That’s the trade. Better Siri for Google, having even more data about how billions of people use AI.
Is that worth it? Depends on how much you value convenience over corporate power consolidation.
The Uncomfortable Questions Nobody’s Asking
Let me ask some things that are bothering me about this entire situation:
Why couldn’t Apple build competitive AI themselves? They have infinite money, top engineering talent, and years of warning that AI would be critical. They’re licensing Google’s models, which suggests either incompetence or deliberate strategy. Neither is reassuring.
What happens when this partnership ends? Multi-year doesn’t mean forever. If Google decides to renegotiate aggressively or Apple wants to switch to a different provider, what happens to all those Siri interactions and learned behaviors?
Who actually controls the data? Apple claims privacy preservation, but Gemini needs data to function. Where’s the line between “running on Apple servers” and “feeding Google’s training data”? The technical details matter enormously here, and they’re conveniently vague.
What does this mean for OpenAI? Apple already partnered with them for some ChatGPT integration. Now they’re primarily using Google. Does OpenAI become a secondary provider? Do they get cut out entirely? This feels like Apple hedging bets rather than committing to either platform.
Why should users trust this? Both companies have histories of privacy issues, monopolistic behavior, and prioritizing profits over user interests. Now they’re combining forces on the most powerful technology of our generation. What could go wrong?
Could a Small Issue Bring Down Apple? The $3 Trillion Question.
Investigating the Subtle Issue That Could Affect Apple’s Future.medium.com
The Reality Check
What’s certain is that AI is taking an important place in our daily lives.
You’ll tell me in the comments what your AI usage is today. Is it a little, a lot, moderately, or not at all? Are you fascinated? Are you terrified? Certain aspects can be, and others, in contrast, can be seen totally differently.
Because AI is a very general umbrella term, there are distinct applications, such as generative AI for image creation, text generation, sound synthesis, and health analysis, among others.
My honest assessment after two weeks of intensive testing: both AI models are impressive. Both have significant flaws. ChatGPT wins in usability and voice interaction. Gemini wins on reasoning quality and response depth.
For most people, ChatGPT remains the better everyday choice because the interface matters more than marginal improvements in reasoning quality. But in complex analytical tasks, Gemini genuinely performs better.
The Apple-Google partnership puts Gemini’s superior reasoning behind Siri’s interface with Apple’s privacy promises. If they deliver on those promises, it could be the best of both worlds.
But I’m skeptical. Corporate partnerships of this magnitude rarely prioritize user interests over profit extraction. Soon, we’ll learn if this helps Siri or merely empowers two tech giants.
Either way, the AI wars are heating up, and we’re all beta testers whether we consented or not.
Thanks for reading me, I hope you enjoyed it. What do you think? Can Apple catch up, or has it already lost the AI battle? Share your thoughts in the comments. See you in the next one!