Apple's Price Hikes Linked to AI Bubble Burst – Expert Warns

The AI Bubble: Why It's Unsustainable

According to Ed Zitron, the economics of Large Language Models (LLMs) are fundamentally broken. He argues that the cost of AI tokens—the units of computation used by models like ChatGPT—far exceeds what consumers pay. Companies like OpenAI lose money on every subscription, burning through tokens at a rate that makes profitability impossible. Zitron claims OpenAI lost $20.9 billion on $13.07 billion in revenue in 2025.

• OpenAI lost $20.9 billion on $13.07 billion revenue in 2025.
• Uber spent its annual token budget in one quarter.
• 89% of AI revenues go to just two companies.
• Hyperscalers need $1.5 trillion in new profit to justify AI spending.

Token-Based Billing: A Hidden Cost

Enterprise customers are being moved to token-based billing, leading to shocking overspending. Uber, for example, burned its entire annual token budget in one quarter. This model makes it difficult for businesses to justify AI costs when the benefits are hard to measure.

Why AI Services Aren't Differentiated

Zitron notes that most AI services offer similar capabilities—generation, summarization, search—making them commoditized. This lack of differentiation means 89% of AI revenue goes to just two companies: Anthropic and OpenAI. The rest struggle to reach $100 million in annualized revenue.

Data Center Costs and Memory Prices

The massive buildout of AI data centers has driven up DRAM and memory prices worldwide. Hyperscalers have spent over $1 trillion on capital expenditures since 2022, with much of that going to AI infrastructure. This has caused memory prices to roughly double, directly impacting Apple's hardware costs.

Apple's Unavoidable Price Increases

Tim Cook has called Apple's recent price hikes "unavoidable." Macs and iPads have already gone up, and iPhones are expected to follow. Consumers are effectively subsidizing data centers that may never become profitable.

The Scale of Hyperscaler Spending

Hyperscalers are spending north of $650 billion on AI infrastructure this year alone, compared to Apple's $14 billion. Apple treats AI as a commodity, paying Google for Gemini and focusing on on-device processing. This conservative approach may protect Apple from the worst of the fallout.

Who Bears the Cost When the Bubble Bursts?

Zitron warns that private credit funds, pension funds, and even individual investors will suffer. Data centers are financed through project financing, and if they fail, there's no easy bailout. Companies like Oracle have bet heavily on OpenAI's success, with billions in debt tied to AI data centers.

Oracle's High-Stakes Bet

Zitron predicts that Oracle will be among the biggest losers. With $340 billion in data center commitments and hundreds of billions in debt, Oracle's survival depends on OpenAI becoming the most profitable company in the world by 2030—an outcome he considers unlikely.

Global Market Contagion

The collapse could affect Taiwanese ODMs like Quanta and Hon Hai (Foxconn), which have seen stock boosts from AI server sales. Korean and American investors in hyperscalers and semiconductor companies are also at risk. Private credit funds, backed by pension funds like CalPERS, could trigger systemic contagion.

What This Means for Apple Users

If the AI bubble bursts, memory prices might eventually stabilize, but the immediate impact is higher hardware costs. Apple's reliance on commodity AI services could position it better than its peers when the reckoning comes. In summary, the AI bubble is raising costs for everyone, especially Apple users. Understanding these dynamics can help consumers make informed decisions about their next hardware purchase.