This morning I read about Microsoft's latest super app plans and their $3.2 billion slash at Anthropic. It made me want to throw up. Not because they're investing in AI—but because they're doing it the wrong way. Again. Let me be clear about what's happening. AI agents are not about adding a chatbot to every Office document. They're about creating an entirely new paradigm for how humans interact with technology—a layer that understands context, intent, and your life. Here's why that matters: whoever nails this gets to own the next decade of computing. Microsoft is trying to buy that future. Apple needs to build it. Microsoft is throwing money at partners—OpenAI, Anthropic, whoever will take the check. They logged $3.2 billion from their Anthropic investment (Source: industry earnings reports, 2025). That's not a strategy, that's a portfolio. You don't create great products by spreading bets across dozens of teams. You pick a direction and march your best people into it. Microsoft's approach is the anti-thesis of focus. They have three different Copilot brands, two different AI model families, and a super app that tries to be everything. It's a committee's dream and a user's nightmare. What are the implications of Microsoft's AI investment? They're repeating their Windows playbook—licensing technology and pretending that makes them the leader. But the world has changed. OpenAI could cut them off tomorrow. Anthropic might prefer a deal with Amazon. Microsoft doesn't control the whole widget—they don't own the silicon, the models, or the hardware that runs them. They're a middleman in a business where the margins go to those who own the stack. Can tech giants collaborate on AI? Short answer: no. Real breakthroughs require a single obsessive vision, not cross-company committees. Apple learned that the hard way with the iPhone—we owned everything from the glass to the operating system. That's why it worked. Microsoft is doing the opposite, and it's going to bite them. Now, how is Apple using AI? Quietly. Siri is still a disappointment, but Apple has been embedding ML into the chip for years—the Neural Engine, on-device processing, privacy-first. That's the right architecture. The problem is they haven't shipped a breakthrough AI experience. They're afraid of shipping something half-baked, so they ship nothing. That's also wrong. The iPhone was half-baked at launch—no copy-paste, no third-party apps. But it had a vision. Apple needs a vision for AI that's not a chatbot. It needs to be the operating system's nervous system. Here's the pros and cons of Microsoft's push: Pro—they have speed and distribution. They can shove Copilot into every enterprise contract. Con—it's disjointed, feels like a bolt-on, and completely lacks soul. Apple's silence is the opposite extreme. My rating: Microsoft's AI push gets a C-. They'll probably make money. But they won't change the world. Apple gets an incomplete. They have the foundation but no product yet. Final advice: Stop worrying about partners. Don't try to build a super app. Build an agent that lives on the device, learns from you, and disappears into the experience. Invisible intelligence. That's the only AI that matters. FAQ Q1: How much did Microsoft invest in Anthropic? Microsoft logged a $3.2 billion gain from its Anthropic investment in 2025, according to industry earnings reports. That’s a financial win, but it doesn’t translate into a coherent product strategy. Real investment should be in your own people and chips, not in buying influence. Q2: What can Apple learn from Microsoft's AI mistakes? Apple should learn that you can't partner your way to greatness. You have to own the hardware, the model, and the experience. Microsoft's approach is fragmented and fragile. Apple's vertical integration is its biggest advantage—use it.