What should Apple do with $100B? Hosts debate M&A, vertical integration, and AI strategy

Summary

Apple has $100 billion in cash. The hosts debate whether a transformative acquisition or capital blitz makes strategic sense for the company.

Take Lui, author of The Nvidia Way, proposed a three-part strategy: quadruple iPhone RAM to 32GB, buy Mistral or Anthropic outright, and commit $100 billion annually to data center infrastructure to enter the frontier AI race. The hosts agree the impulse is understandable—AI is existential to computing platforms—but disagree sharply on whether Apple should swing hard.

Tim Cook addressed the question directly on the earnings call. Apple has acquired around seven companies this year and remains open to M&A that accelerates the roadmap, but "the ones that we've acquired thus far this year are small in nature." Cook's comfort with smaller, talent-focused deals reflects Apple's historical M&A discipline. The company rarely pays VC-scale valuations for companies, let alone unicorns.

Distribution

Instead of buying a foundation model lab, the hosts sketch a more elegant path. Apple could monetize its position as the default device for AI consumption. Google pays Apple roughly $20 billion annually for search default status. A similar dynamic could emerge with large language models. One scenario: a foundation model company pays Apple a massive fee to own the Siri button, while Google retains search. ChatGPT is so dominant that most iPhone users don't care which model runs behind the scenes. Apple could route billions of queries to a paid LLM and capture a cut without owning the model itself.

Currently, zero dollars flow between Apple and OpenAI because OpenAI pushes users to web subscriptions to avoid the App Store's 30% take. But that economic dynamic could shift.

Vertical integration thought experiment

The hosts run through implausible but mechanically plausible uses of $100 billion: buying Samsung (market cap $330 billion, leveraged 2:1 with debt), opening 6,000 new retail stores to bring Apple's footprint to 7,000 locations (matching Subway, CVS, and 7-Eleven in density across America), buying Corning, Interdigital, rare earth mines, battery gigafactories, and sapphire crystal supply chains. One proposition: send 200 million free iPhones to American consumers as a competitive response to OpenAI's hardware push. The hosts acknowledge these moves would be wildly anti-competitive. The math works anyway.

Cook's restraint

Apple will not compete with OpenAI, Anthropic, Meta, Google, and Xai on frontier AI superintelligence. Those are five well-funded, serious teams. Apple will center strategy on its devices and direct customer relationships, at least under Cook. Anyone arguing Cook should step aside faces a difficult argument: he has been on a generational run and solves the most important problem Apple faces. Not AI applications, but supply chain execution and keeping iPhones flowing to market. Cook is now the longest-tenured CEO in Apple history, longer than Steve Jobs.

Apple's moat is not compute or model weights, but distribution and the device. Every new computing platform failed to dislodge Apple because the company was never meaningfully behind. The one crack in the foundation would be an AI-first phone designed from the ground up around on-device AI and chatbots. No folding phone, design refresh, or spec improvement has managed that threat. But that risk does not necessarily justify Apple burning capital on a frontier AI arms race it cannot win and does not need to own.