# A Bull & Bear case for Apple

Source: https://rickyrichards.com/blog/a-bull-and-bear-case-for-apple
Author: Ricky Richards
Published: 2026-08-26
Category: Investing
Keywords: Apple AI strategy, AI inference costs hardware, personal AI brain layer, enterprise AI spending, Apple privacy advantage

> Apple's hardware economics and consumer trust give it a path to AI dominance—but only if it moves faster than Nothing, OpenAI, and emerging rivals.

For a while now, I’ve been telling friends that I believe Apple could be disrupted. To start with, Apple has largely bowed out of the AI infrastructure play when compared to the rest of the Magnificent Seven. It hasn't dedicated serious resources to the build-out, yet despite this, it has been rewarded by the markets.

Market's optimism aside, it's clear that Apple is behind the curve on integrating AI into its products, a trajectory that is speculated to change with the appointment of [John Ternus as chief executive](https://www.youtube.com/watch?v=DWdPxX7FvCk). Ternus came up through hardware engineering, but at heart, he's a product person, whereas Cook was an operator.

Where Apple goes from here, only time will tell, but I’ll do my best to speculate on some possibilities on both the bull and bear cases.

The bull case is that Apple is very good at producing hardware.

Just the other day, they announced the new Mac Studio M5 Ultra, which starts at $5,500 and, specced to its full capacity, comes in at $18,300. There's a 512-gigabyte configuration arriving in October that will comfortably clear $20,000.

[https://www.youtube.com/watch?v=3uAIqqg8ZHo](https://www.youtube.com/watch?v=3uAIqqg8ZHo)

On the surface, that is a ludicrous figure. However, it stops being ludicrous the second you look at what enterprise AI is actually costing companies today. I know from firsthand experience that it is not uncommon for a designer or engineer to run up a bill of over $6,000 a month in AI spend. Run that across thousands of team members, and you’re looking at a serious line item.

All of a sudden, dropping $18,000 means that you can load a leading frontier open-source model onto this hardware and go from millions a month in variable burn to a fixed asset that pays for itself in a matter of months.

If that plays out at scale, there is enormous upsdie for Apple.

Now, the obvious argument against this bull case is that inference costs are falling fast, and there's a plethora of startups working on efficiencies in chip design that will pass savings back to the model providers in time. If tokens get cheap enough, the entire capex argument evaporates. 

Despite this, there’s still a window of opportunity. It's just uncertain how long it will stay open.

The bear case is the software.

Apple's adoption of AI across its own apps has been abysmally slow. And I don't think it's unrealistic for an ambitious company to look at the iPhone as it exists today and the default applications that ship on it, and believe they have the ability to rebuild the whole experience from first principles through the lens of AI within a short period of time.

On the day I'm writing this, Nothing just shipped their OS 5, built on Android 17. Included in their release are north of eighty changes, with AI threaded through every core experience rather than bolted onto the side. They're obviously not trying to compete with Apple at the highest end. But it is already a step ahead of where Apple is today, and that should be uncomfortable for Apple executives.

[https://www.youtube.com/watch?v=aYoWAJz8LFw](https://www.youtube.com/watch?v=aYoWAJz8LFw)

Then there's OpenAI, working with Jony Ive on a pocket-sized device. We don’t know if this is intended to replace or supplement the iPhone just yet. Either way, that’s Apple's greatest designer at the helm of arguably Apple's largest emerging competitor.

[https://www.youtube.com/watch?v=J4r3Yth10SA](https://www.youtube.com/watch?v=J4r3Yth10SA)

Then the one nobody is watching, Hark. Brett Adcock's new company. Best known for Figure AI, he recently raised somewhere in the region of $1 billion to build out a family of AI devices. If the robotics work is anything to go off, there's a good chance they’ll build something genuinely beautiful, with AI powering the heart of the product.

[https://www.youtube.com/watch?v=0H1LSLipOVI](https://www.youtube.com/watch?v=0H1LSLipOVI)

None of these is likely to kill Apple overnight. After all, Apple has lock-in and a huge customer base. But while lock-in is a fantastic defence in the short term, it’s not a long-term strategy. And over time, if stagnation creeps in, competitors will eat away at market share.

Despite this, there is one aspect of the AI race that Apple is better positioned to win than any other company on earth, and I don't think they've taken advantage of it yet.

The AI brain.

One of the greatest advantages of AI is that it remembers things when humans are fallible enough to forget. Right now, every company you interact with keeps its own private silo of your data and uses it to build enterprise value for itself. You get a marginally better product. They get the asset.

What you actually want is the inverse. One secure, central place that holds everything about you, that doesn't look to monetise you, but that you can point at any product to make it instantly personal. The equivalent of signing in with Google, but for your entire life’s worth of data, rather than just your password.

Andrej Karpathy has already sketched a version of what this looks like with his [LLM wiki](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f), a knowledge base that an AI model maintains for you. The downside of this is it lives on your machine, at the scale of one person, which is wonderful in some respects but terrible when you want to plug that knowledge into third-party tools. It’s a great proof of concept, but lacks adequate integrations.

The reason I believe Apple is well positioned to build this is nothing technical. It's that Apple is the only company most people trust with their data.

I cannot picture a world where OpenAI or Anthropic offers to hold everything about your life and the public says yes at scale. Apple has spent the best part of several decades making privacy the product. And it already has a device in your pocket that can say more about you as a person than you could articulate about yourself.

I suspect that whoever ends up owning this personal brain layer is going to be integral to the AI ecosystem. Every asset required to make it happen, Apple already owns but isn’t yet leveraging. The silicon. The device in your pocket and, most importantly, years of trust that no competitor can buy their way into possessing.

If Apple can get its act together, they have all the ingredients to remain at the frontier for decades to come. If they don’t, then there’s a chance they die via a thousand paper cuts.