Ed Zitron: Apple Rides Out the AI Bubble While Everyone Else Burns

Critic Ed Zitron argues the generative-AI buildout is economically unsound and headed for a costly unwind. His core claim: large language models break the software business model because they meter usage at a per-million-token rate regardless of output quality, and providers can’t charge customers what inference actually costs. To hide this, firms sell flat subscriptions that let users burn far more in tokens than they pay, and enterprise token-based billing has already produced sticker shock — Uber reportedly exhausted its annual AI budget in a single quarter. He points to OpenAI losing roughly $21 billion on $13 billion in revenue in 2025, and notes that nearly all AI startups reliant on those token rates are unprofitable and largely undifferentiated.

Zitron’s bigger worry is systemic. AI data centers cost billions and take years to build, are financed largely through project debt and private credit funds backed by pension money, and have few real tenants beyond OpenAI and Anthropic themselves — which he says masks weak enterprise demand. Because the financing runs through special-purpose vehicles, he doubts a clean bailout is possible. He predicts contagion rippling from Taiwanese and Korean hardware makers up through the hyperscalers and chipmakers whose valuations depend on continued AI spending, and singles out Oracle’s debt-heavy data-center bets as dependent on OpenAI becoming the world’s most profitable company.

The consumer angle is where it hits Apple users: hyperscalers hoarding memory for AI infrastructure have roughly doubled DRAM prices this year, pushing up Mac and iPad prices with iPhones expected to follow. Tim Cook calls the increases ‘unavoidable.’ Zitron frames Apple as the odd company out — spending about $14 billion versus the hyperscalers’ $650 billion-plus, renting Google’s Gemini for around $1 billion a year to power Siri and leaning on-device rather than building. He’s skeptical that depending on vendors he expects to collapse is truly savvy, but suggests Apple’s restraint may leave it far better positioned than rivals when the spending unwinds.