March 15, 2026
AI or just iCheapskate?
The most brilliant move in corporate history?
Apple just watched Big Tech burn $650B and said “nah we’re good”
TLDR: Apple is skipping the $650 billion data-center spending spree and instead renting AI tech while using its existing devices as a giant distributed computer. Commenters are wildly split between calling it 4D chess genius and accusing Apple of laziness and free-riding, making this the latest big-tech flame war to watch.
Comment sections erupted after analyst Horace Dediu called Apple’s AI strategy possibly “the most brilliant move in corporate history.” While Amazon, Google, Microsoft, and Meta are spending hundreds of billions building giant AI computer farms, Apple basically said, “Why buy the factory when we can just rent the machines?” and the internet lost it.
One camp is screaming that Apple just pulled off the ultimate “galaxy brain” move: avoiding massive debt, licensing Google’s Gemini AI for pocket change, and turning 2 billion iPhones and Macs into a giant “invisible supercomputer” without building a single mega–data center. These folks are posting memes of Tim Cook sitting calmly while the other tech CEOs toss cash into a bonfire.
But the other side is calling this cope. They argue Apple missed the AI wave, is freeloading off everyone else’s hard work, and will panic-build data centers later at an even higher cost. Some say Apple is smart; others say they’re the kid who didn’t study and is hoping to copy off Google’s homework.
Jokes are everywhere: people comparing Nvidia to a “digital oil cartel,” calling the data centers “crypto mining but with extra steps,” and dubbing Apple’s plan “AI, but make it frugal.” The community vibe: half admiration, half chaos, 100% drama.
Key Points
- •Apple’s capital expenditure remains around $14 billion focused on hardware tooling and related investments, while major hyperscalers are projected to spend about $650 billion in 2025 on AI data centers.
- •Amazon, Google (Alphabet), Microsoft, and Meta are estimated to devote about 94% of their operating cash flow to AI infrastructure, contributing to rising debt levels and reduced free cash flow.
- •AI services currently generate roughly $35 billion in revenue, about 5% of the estimated AI infrastructure spending, while AI models are rapidly commoditizing and becoming cheaper to build and deploy.
- •Apple has chosen to license Google’s Gemini model for about $1 billion per year instead of heavily investing in its own large AI model training infrastructure.
- •Apple’s M5 chip, with a 16-core Neural Engine and neural accelerators in every GPU core, is claimed to run 70‑billion‑parameter AI models locally with four times the AI performance of the M4, enabling distributed on-device AI across about 2 billion Apple devices.