Apple's Low-Capital AI Strategy: Not Building All Models, Investing Less in Infrastructure, Profiting from Interface Position
Apple's latest earnings report shows its quarterly capital expenditure was only $2.2 billion, far below Amazon's $77.3 billion and Alphabet's $51.4 billion. The article explains how Apple is generating multiple revenue streams from the AI boom without building large-scale AI infrastructure, through hardware upgrade incentives, App Store commissions, and planned usage-based subscription tiers.
During Apple's fiscal year 2026 fourth-quarter earnings call, AI and Apple Intelligence were mentioned 45 times, setting a historical record. But what truly matters is not the frequency of mentions, but the fact that the company demonstrated an AI economic model distinct from its peers: it does not need to train all large models itself, nor spend tens of billions of dollars building data centers; instead, it profits from AI by selling devices, collecting platform fees, and implementing planned subscription tiers.
The comparative figures are striking. Apple's capital expenditure for the last quarter was $2.2 billion, while Amazon reported $77.3 billion during the same period, and Alphabet (Google) reported $51.4 billion. These are company-wide capital expenditure figures, not limited to AI, and they exclude some AI-related expenses, yet the magnitude of the gap remains significant. Apple Chief Financial Officer Kevan Parekh acknowledged on the call that Siri AI's cloud costs will be reflected in service costs and may expand as usage grows—he also stated that it is currently impossible to predict how many requests will require server processing rather than on-device computation. This means Apple's low capital expenditure is not a permanent commitment, but rather a phase-specific cost structure choice.
Apple has three revenue pathways. The first is hardware: iPhone revenue grew 17.5% year-over-year to $57.6 billion, setting a September-quarter record, while Mac revenue grew 24% to $10.8 billion, its highest level in four years. Parekh described AI as "a compelling reason for customers to upgrade and switch to get the latest features." The second is platform commissions: according to data cited by The Wall Street Journal from analytics firm AppMagic, Apple is expected to collect over $1 billion in App Store fees from generative AI apps in 2026. The third is upcoming usage-based subscriptions: newly appointed CEO John Ternus stated that features relying on server-side models (including Siri AI and intelligent photo editing) will have usage limits, and customers needing more capacity "can purchase expanded access in the future." Ternus illustrated this logic's evolution using iCloud as an example—"Initially, iCloud tiers were primarily about storage, but we have been adding more capabilities and features."