Apple lost the AI race. That feels pretty obvious, right?
The company announced a much smarter Siri, failed to deliver it for two years, and ended up rebuilding with help from Google. Meanwhile, ChatGPT, Gemini, Claude, and the rest kept moving. Apple Intelligence has hardly looked like the next big thing.
But there’s another version of this argument where Apple is still in a pretty good position. Maybe even a winning position.
The difference comes down to what you mean by the AI race. Building the best AI assistant is one competition. Making the hardware people use to run those assistants is another. And losing the first doesn’t automatically mean losing the second.
The Case Against Apple Is Pretty Straightforward
When ChatGPT arrived, it was immediately clear that something significant had changed. This wasn’t just another app getting some attention. It looked like a new way to interact with computers.
The response was fast. Google went into overdrive. Meta ramped up its efforts. Bing had a particularly wild couple of weeks. ChatGPT’s growth was extraordinary, with an estimated 100 million monthly active users just two months after launch.
And Apple? For a while, basically nothing. The company seemed reluctant even to say “AI” at its events.
Now, the obvious defense is that this is what Apple does. It usually isn’t first to a new technology. It waits, lets early adopters discover the benefits and the bugs, and then arrives with something more polished and more tightly integrated into its ecosystem.
Fair enough. But that explanation only works if the eventual product is actually good.
Once the technology has reached hundreds of millions of people, “we’re waiting for it to mature” starts to sound less like patience and more like being behind. A company with Apple’s resources should be able to recognize that generative AI is here to stay and build a cohesive version for its own products.
So far, Apple Intelligence hasn’t made a convincing case that Apple’s usual late-arrival strategy is working here.
But Apple Is Still the iPhone Company
Here’s the counterargument: not every new technology has to be treated as an existential race by every company.
For a search company, a chatbot that changes how people find information is a direct threat. For a company selling AI subscriptions, the quality of its models is central to the business.
Apple’s position is different. Its core business is still selling hardware, especially the iPhone.
That leads to a slightly uncomfortable hot take: Apple Intelligence might not have to be amazing for Apple to keep winning as a business.
It might only need to be good enough to reassure investors that Apple is competitive and doing something meaningful with AI. If people continue buying iPhones, Apple can remain successful without owning the best chatbot.
That is a much lower bar than building a model that beats ChatGPT, Gemini, and Claude. It’s also a very different definition of success.
Are AI Features Actually Making People Switch Phones?
Of course, there’s a problem with the “people will keep buying iPhones” argument. What if they don’t?
If AI is your top priority in a phone, there’s a reasonable case for looking at Google, Samsung, or Xiaomi before Apple. Their assistants, writing tools, and AI photo editing can make Apple’s efforts look behind.
But how many people are actually choosing an entirely different phone because of those features?
That’s where the argument gets interesting. A feature can be better without being important enough to change a purchase. AI writing tools and photo editing can be useful, but they can also feel like gimmicks. Being able to demonstrate something impressive is not the same as giving someone a reason to leave the iPhone.
From that perspective, Apple’s feature list starts to look like an effort to stay comparable:
- Writing tools? Covered.
- AI photo editing? There’s a version of that.
- A smarter Siri? That’s the promise.
- Genmoji? Of course, Genmoji.
The question isn’t whether Apple has the strongest version of each feature. It’s whether the gap matters enough to make people buy something else.
The Other AI Race Is About Hardware
There’s a second competition happening underneath all of this: who makes the devices that run AI?
For tasks a local model can handle well, running on the device offers an appealing combination of speed and keeping data local. Bigger, more capable models are what you turn to the cloud for.
Now extend that idea forward. As on-device models improve, more tasks could happen locally. You might need the cloud less often. In the most optimistic version of that future, most of what you want an assistant to do happens right on your own machine.
That future isn’t guaranteed. But if it arrives, somebody still has to sell you the hardware.
Machines with enough memory to run capable local models. Phones and computers that can handle useful AI tasks without sending every request to a remote server.
Apple can be one of the companies selling those machines, even if it doesn’t build the best model running on them.
Think about search. Apple doesn’t need to make the best search engine to sell you the device you use to Google something. The same basic logic could apply to AI.
Nvidia’s push into local AI hardware makes sense through this lens, too. There’s a hardware business here, separate from the competition to make the most capable assistant. That opportunity isn’t limited to the United States.
You can lose the model competition and still sell the machine running the winning model.
Two Races, Two Very Different Finish Lines
So it helps to separate the argument into two parts.
Race One: Build the Best AI Software
In this race, Apple is behind. Its competitors have established products and capabilities that Apple is still trying to match.
Apple can assemble a more competitive suite over time, but relying on another company’s technology doesn’t suddenly make it the leader. If company A is paying company B a billion dollars a year for the AI it needs, it’s pretty reasonable to say company B has won that particular part of the competition.
Apple can still deliver a useful product through a partnership. That’s just different from leading the technology itself.
Race Two: Own the Device People Use for AI
This finish line is much farther away. And Apple has a meaningful starting advantage because it already makes the iPhone.
It has a device people use constantly, an ecosystem built around it, and every reason to protect that position. New products and features can follow the familiar Apple playbook: make them work best with the iPhone.
But that advantage isn’t a permanent guarantee. An AI company could try to build its own hardware and give people a reason to move away from Apple’s devices. Reports and rumors about OpenAI exploring hardware make that possibility more than a purely abstract thought experiment.
The important distinction is between making a device and making one compelling enough to replace an iPhone. Those are very different achievements.
Siri’s Integration Advantage Has Limits
Apple’s strongest software argument isn’t necessarily that Siri will be smarter at everything. It’s that Siri can work more deeply with the things already on your phone.
The newer Siri beta points toward that advantage by drawing on iMessage, Apple Calendar, and Photos in ways a standalone ChatGPT app on the iPhone can’t match. That kind of integration could make an assistant useful without requiring it to lead every category of AI performance.
But access to apps and information isn’t the same as being the best personal assistant.
The Siri experience being tested still has meaningful limitations compared with ChatGPT and Gemini:
- It isn’t a comparable tool for advanced work such as coding apps.
- Large requests don’t offer the same ability to leave them running in the background and return later.
- Remembering information about you depends on creating notes in the Notes app, rather than offering the same kind of persistent personal context.
That last point is especially important. Siri may have deeper access to parts of the iPhone, while a competing assistant may still be better at remembering who you are and carrying that context into future conversations.
And those competing assistants already run on the iPhone. You don’t necessarily have to change your hardware to get better AI.
The Real Question Isn’t Who Has the Best Chatbot
Apple’s AI shortcomings are real. Calling the company’s position strategic doesn’t erase missed promises or make an underwhelming feature suite impressive.
But declaring the entire business a loser because it doesn’t have the best AI model is too narrow, too.
The more useful question is which of these happens first: an AI company builds hardware good enough to pull people away from the iPhone, or Apple makes its AI good enough to keep them there.
One side has to make a compelling new device. The other has to improve the software on a device people already own and like.
So, did Apple lose the AI race? The software race so far, sure. But the race to remain the device in your pocket is still very much on.
