Milk Road AI How Apple Could Beat Amazon, Google & Microsoft in AI w/ Ejaaz
9 snips
Feb 18, 2026 They unpack the $645B AI capex surge from hyperscalers and why markets are jittery. They compare two AI races: massive cloud model scale versus device-driven user experience. They explore Apple’s different play—local models, sensors, wearables, and an AI OS for personalized agents. They debate hardware-first strategies, Tesla parallels, and which AI layers capture the most value.
AI Snips
Chapters
Transcript
Episode notes
Demand, Not Waste, Fuels Hyperscaler CapEx
- Hyperscalers are overspending because demand already exceeds capacity across clouds and chips.
- Ejaaz argues the $645B CapEx is justified because GPUs/TPUs and instances are sold out and driving immediate usage.
Two AI Races: Models Versus Distribution
- Apple and hyperscalers race different layers: hyperscalers build the models, Apple builds distribution and UX.
- Ejaaz frames Apple's strength as an 'AI operating system' with massive device reach, not model training.
Personalized Agents Are The Next Value Layer
- Personalized intelligence (agents) will be the next most valuable AI layer and consumers will pay subscription fees.
- Ejaaz predicts agents that monitor health, automate work, and proactively act will command meaningful monthly payments.