⚠ The Software HalfModerate threat

The fit between hardware and software is a moat only if both halves are excellent, and the uncomfortable truth is that Apple's software has lately been the weaker half. The most advanced neural engine in the world does little for the customer if the intelligence meant to run on it disappoints — and Apple's long struggle to make Siri genuinely¹ capable, and its visible stumbles in delivering the AI features it has promised, expose the risk plainly: the company owns superb silicon whose potential its own software has not fully unlocked.

The danger is asymmetric in an unfamiliar way. For years the story was hardware excellence carrying the whole; now the binding constraint has shifted to software, and software is not where Apple's cultural strengths of restraint and industrial design most naturally apply. Building large AI models and the services around them is a different discipline, and one in which nimbler, more focused rivals have plainly moved faster. The catch-up now shows in the accounts: research spending rose 32% in the June 2026 quarter, which Apple attributes to infrastructure costs including investments in artificial intelligence².

Against this stand vast resources, the world's most valuable software distribution channel, and a long habit of arriving late and polished rather than early and rough. The hardware foundation is already laid, waiting for the software to catch up, and Apple can lean on partners for models while it builds its own. The co-design still produces real, felt smoothness across everything the products already do.

Moderate, then — with the next Siri as the test. The fit remains a genuine advantage in everything Apple already does well, but it is only ever as strong as its weaker half — and for the first time in a long while that weaker half is the software, in exactly the area, artificial intelligence, that may come to define the next decade of the products.