Apple Will 'Watch Everything Burn' When AI Bubble Bursts - Ed Zitron

Ed Zitron, host of the Better Offline podcast, argues that the current AI boom is a massive, unsustainable bubble. His recent analysis suggests that the economics of Large Language Models are fundamentally broken because companies are forced to burn through compute tokens that cost far more than the subscription fees users pay. Major tech firms are currently subsidizing these services to gain market share, but industry financials show that primary players like OpenAI lost over 20 billion dollars in 2025 alone.

This infrastructure spending is causing widespread ripple effects in the hardware market. As hyperscalers prioritize data center expansion, memory prices have spiked, leading to higher costs for consumer devices like Macs, iPads, and iPhones. Zitron contends that consumers are effectively paying a premium to subsidize data centers that lack a clear path to profitability. The demand story for these massive computing facilities is largely driven by the AI companies themselves rather than actual enterprise utility.

According to Zitron, Apple is taking a different approach compared to its peers. While competitors spend hundreds of billions on speculative infrastructure, Apple has kept its capital expenditure significantly lower, opting to treat AI as a commodity rather than the core of its business. He suggests that if the AI market experiences a sharp correction, Apple is positioned to remain on the sidelines. The company's focus remains on its hardware ecosystem rather than deep, costly integration with unstable AI providers.

The potential fallout from a bubble burst could impact private credit funds and pension funds that currently finance these data center projects. Investors in semiconductor firms and hardware providers could see significant stock price corrections as the artificial demand for AI-specific servers subsides. Despite the pressure to keep pace with industry trends, Apple appears to be betting on its established product lines while avoiding the debt-fueled race that defines the current AI landscape.