Why Apple is the Smartest Company in Tech Right Now

Big tech is betting everything on AI. In 2023, when ChatGPT began attracting hundreds of millions of users, every major tech company was forced to explain how their business would adapt in the age of AI. And in just three years alone, in total they have spent nearly 1 trillion dollars. And then there was Apple. While every other company rushed to launch a chatbot, Apple waited more than a year before finally making its move. And when they eventually did, it was a total disaster. In the summer of 2024, they announced Apple Intelligence, but many of the biggest features they promised never arrived. Siri was still as useless as ever, and all people were left with were generative emojis, some photo editing tools, and a painfully slow writing assistant that relied almost completely on ChatGPT. In fact, Apple overpromised so much that they were even sued for false advertising. On the surface, it may seem like Apple is falling behind, as every other Mac 7 company stock has outperformed it. And while Amazon, Google, Meta, and Microsoft are collectively spending around 700 billion dollars on AI this year, Apple has barely spent 3% of its revenue on CapEx. So, did they simply get too overconfident and become greedy? Or are they actually being far smarter than every other company in tech? Because the closer you look, the more you realise, instead of missing out on the AI boom, they're profiting on it far better than anyone else. But it's also a strategy that comes with risks far greater than anything they've done until now. Here's why Apple is the smartest company in tech, at least for now. To understand why Apple isn't spending hundreds of billions of dollars on AI infrastructure, you first kind of have to understand why everyone else is. Ultimately, it's because they need to build data centres. And there's two main reasons for that. Firstly, products like ChatGPT or Gemini are general purpose AI models that can handle an enormous range of tasks. And training an AI like that requires thousands of specialised chips. And secondly, think of a company like OpenAI. They don't make phones or computers, so the only way they can serve ChatGPT's almost 1 billion users is via the internet. Which means OpenAI needs to build massive data centres. Even a company like Microsoft isn't that different. Yes, Microsoft makes laptops, but they aren't powerful enough to run big AI models on their own. So they also have no choice but to build data centres. And these data centres can get very expensive very fast. Between construction materials, staff, networking equipment, energy infrastructure, and thousands of GPUs, the cost for a single AI data centre can reach billions of dollars. And these companies are building hundreds of them. That's why the AI race has turned into a bit of a Now, because Apple doesn't have a general purpose chatbot like ChatGPT or Claude, you might think that the reason they're not spending billions of dollars on AI is simply that they're not building any. But they absolutely are. But before we look at that, I wanted to mention the sponsor of today's video, Abacus AI, and their AI Agent, which is basically a general purpose AI that can build entire apps for you, full stack, with hosting, a database, authentication, and even strike payments, all from a single prompt. But this is really just scratching the surface. You can automate routine tasks, generate deep research reports, build rich presentations with real data and charts, and even create mobile apps. What I've actually been using this for is the automation side. I set up an agent that handles a bunch of repetitive stuff I was personally wasting hours on every week. Things like scheduling posts, pulling data, you know the drill. A massive bonus is that it connects to basically every tool I already use, so it just slots in. It also gives you access to all the top AI models, like ChatGPT, Claude, Gemini, and Grok, plus image and video generation tools, all in one place. And the basic subscription is just $10 a month, which includes three free agent tasks. Go click the link in the description or the pinned comment and try it out for yourself. Thanks to Abacus AI for sponsoring this video. Now let's get back to looking at how Apple is outsmarting their competitors. It's just that their approach to it is completely different. Companies like Google and OpenAI are in a constant race to build the most advanced AI models so that they can monetize them mostly through ads, subscription, or usage. But Apple's business isn't building frontier AI models, it's building great consumer products. And those products don't necessarily need the They simply need useful features. Things like removing objects and photos, understanding what you're looking at through your camera, and helping write messages or summarize notifications. On an operating system level, those are the kind of features that make sense for Macs and iPhones. So instead of trying to build one giant general purpose AI system that can literally do everything, Apple is focusing on smaller, more specific models designed for smaller tasks that improve their user experience. Secondly, the chips inside modern iPhones and Macs are already very powerful, which means Apple doesn't need to deliver the AI over the internet, it can run it directly on your device. So while companies like Google and OpenAI are betting on cloud-based general intelligence, Apple is focused on something very different altogether, on-device task-specific AI. And instead of monetizing that as a standalone product, they're simply integrating it directly into their devices as built-in features, for free, just to make their existing products more compelling. Now, to be clear, part of that strategy is adding more general AI features, and that's where Siri comes in. Apple hasn't completely abandoned the idea of having its own general purpose AI assistant, but when it comes to that, they're making a totally different bet than the rest of the industry. Apple instead seems to be making a very specific bet about the future of AI. Instead, they believe the advanced large language models companies like OpenAI and Google are spending tens of billions of dollars to build what will eventually become a commodity. Think of it like the iPhone's camera sensor. It's an important component of the device, but Apple doesn't make it themselves, they buy those sensors from Sony. However, if one day Sony decided to hike up the prices to an insane degree, Apple could just as well start buying its sensors from Samsung. In other words, Apple believes the technology behind general purpose AI assistants will likely be available from multiple providers, all competing at roughly the same level of capability. And if that's true, then it may not make sense for them to spend billions of dollars developing their own front-end model for Siri. Instead, they can simply buy the technology from any number of companies at a competitive price. So essentially, Apple is making a bet that cuts against almost the entire tech industry, while everyone else is acting as if the future belongs to whoever owns the most advanced AI models. Apple is acting as if the model itself may eventually matter less than people think. And instead, the real winners may not be the companies that build the models, but the companies that distribute them. And that's exactly what they're doing. In November 2025, Apple reportedly cut a billion-dollar deal with Google to make Gemini the foundation of the future version of Siri. From Apple's perspective, it's a much cheaper way to deliver a general AI assistant. They've let Google burn the cash to train AI, paid them once for it, and now they're free to monetize the finished experience. Now taking all of this into account, you can start to see how Apple is approaching AI in a completely different way than their competitors. But really, the smartest part about their strategy is how it reinforces Apple's existing business. And that all comes down to one thing. Vertical integration. You see, Apple has always approached technology a little differently. Instead of building standalone products, they build tightly integrated systems, where the hardware, the software, and the experience are all designed together. And today they have what you might call a trifecta. They build the hardware. iPhones, Macs, iPads. They control the operating systems those devices run on. iOS, macOS, iPadOS. And crucially, they design the very chips that power that hardware. The A-series and the M-series processors. That level of vertical integration is extremely rare in the tech industry, and it leads to a very different kind of AI strategy. We usually think of AI infrastructure as massive data centers filled with Nvidia chips. But from Apple's perspective, that infrastructure is already sitting in your pocket. And that means Apple can have its cake and eat it too. Instead of having to build expensive data centers, they can simply put their focus on improving their chips. And on top of improving AI performance, it improves the entire device. Better chips means longer battery life, better video editing, better gaming, and better performance across their entire operating system. And if they keep improving their chips year after year, they might even be able to have their own powerful chatbot app that runs completely on device, without even an internet connection. And this also happens to reinforce one of Apple's biggest selling points, privacy. Because when AI runs directly on your device, your data never has to leave it. When you think of it that way, Apple's real competition isn't in AI, it's Nvidia. Nvidia benefits from a world where AI requires more cloud servers, more infrastructure, and more capex. Whereas Apple benefits from a world where AI becomes light and efficient, enough to carry around in your pocket. Because ultimately, their AI strategy is simple. The goal is to sell more iPhones and more Macs. And to do that, they just have to do exactly what they've been doing since before the AI boom. Design better chips. It goes to show just how powerful Apple's vertical integration is. Now, from a product standpoint, this is by all means a viable strategy. But once you realize Apple doesn't need to build data centers, you also start to realize something else. Financially, it puts them in an incredibly powerful position. For companies like Google or OpenAI, building AI means adding an entirely new cost to their business, data centers, tens of thousands of Nvidia GPUs, and huge teams of AI researchers, etc. But since Apple can focus on improving its chips, building for AI doesn't change the fundamentals of the business. Instead, they can continue operating with gross margins over 40% while generating enormous amounts of cash flow. And that raises an obvious question, what does Apple actually do with all that extra cash? The simple answer is that they can reinvest that cash to keep improving their chips and their products, or even venture into entirely new product categories like the Vision Pro or rumored AI glasses. But there's also a second place that cash goes, share buybacks. Over the past decade, Apple has reduced its total share count by roughly one third, making it arguably the best buyback machine in corporate history. Take 2025 as an example. While companies like Google and Amazon and Meta were spending hundreds of billions of dollars building AI data centers, Apple spent just $12.7 billion on CapEx. And because of this, they were able to buy back $110 billion worth of their own stock, the most a company has ever done in history. And for Apple's shareholders, it means the supply of shares go down, and their existing shares become more valuable. So this is the core of Apple's strategy, they improve their chips, they run smaller, task-specific AI directly on their devices. Those features make the products more compelling, which sells more devices, which keeps the company growing. And because they avoid the massive infrastructure costs of their competitors, they remain incredibly profitable, letting them increase shareholder value through buybacks. Now, even though the strategy makes sense on a conceptual level, it does come with some very serious risks. Firstly, Apple's entire AI vision depends heavily on being able to run powerful AI on device. But if AI models keep getting bigger and more computationally demanding, Apple won't be able to offer the most advanced features to their customers directly. And the power will remain in the hands of companies like OpenAI and Google. But the most important risk is that Apple is simply wrong about how the AI market will evolve. Right now, they're acting as if the most advanced models will eventually become cheap, abundant, and widely available. Something that they can license the same way they license their sensors in the camera, like they're doing now with Siri. But if the best AI stays concentrated in the hands of a few companies, then Apple could end up deeply dependent on technology it doesn't control. And for Apple, dependence is dangerous. This is a company that's built its entire business model around controlling the full experience end-to-end. From the hardware, to the software, to the chips, the whole ecosystem and store. And if the single most important technology of the next decade ends up living outside Apple's walls, Apple could suddenly find itself years behind the rest of the industry, structurally dependent on the very companies they're trying to compete with. So in many ways, Apple's entire strategy hinges on a single assumption. If it turns out to be wrong, the company spending hundreds of billions of dollars today might end up with an insurmountable lead. And of course, it's also worth mentioning that AI isn't the only challenge in Apple's path. The company itself is on the cusp of an entirely new era. There have been many drastic changes to its leadership recently, and with rumors of Tim Cook stepping down from his role of CEO soon, whoever leads Apple next will have to guide the company through what very may well be one of the most significant technological shifts of the past few decades. So Apple's future is still far from guaranteed, but if their strategy works, they might once again prove that just like in the stories, slow and steady really does win the race.