- Markets - Around The World
- Posts
- Apple's AI Bet Is Powered by Google
Apple's AI Bet Is Powered by Google
& ChatGPT Couldn't Kill Google Search
Apple Just Built the Best AI Agent on a Phone - Powered by Google
Apple unveiled "Siri AI" at WWDC 2026 - a rebuilt-from-scratch personal assistant that runs on Google Gemini under a multi-year pact the two companies struck in January 2026 - alongside an AI-powered Shortcuts app in iOS 27 that lets users build multi-app workflows by writing a natural-language prompt. Together they form the most integrated AI agent experience on a smartphone today: camera-aware (a new Siri Mode lets you point the iPhone camera at food for nutrition info or split a bill via Apple Cash), all-app access on iPhone with on-device privacy controls, and continuity across iPad, Apple Watch, and Vision Pro.
The integration moat matters more than raw model capability - Gemini on Android exists, but only Apple ties the model into every system app, the camera, Apple Cash, Vision Pro continuity, and a Shortcuts engine that builds workflows from plain English. The price of admission is a forced hardware cycle: Siri AI requires iPhone Air, iPhone 17 Pro, iPhone 17 Pro Max, iPad M4+, or Mac M3+, all with 12GB+ of unified memory, meaning the long tail of older devices is locked out of the agent entirely.
The strategic implication is unmistakable - Apple has subcontracted the model layer (the most expensive part of AI) to Google while keeping the interface layer (the moat) for itself, the same way it once outsourced LCD panels to Samsung while keeping the iPhone brand premium. Tim Cook will hand the CEO chair to hardware lead John Ternus in September, just as Apple bets the next decade on agent hardware - and notably, Siri AI is not launching in the EU or China at all, where the regulatory questions around agentic AI remain unresolved.
Amazon Wants to Do Print-on-Demand With a Prompt
Amazon launched on June 8 a feature inside Alexa for Shopping that lets US customers describe any design via text prompt, see it generated in seconds, edit it, and order it printed on a T-shirt, polo, hoodie, jersey, sweatshirt, tank top, or tumbler - all routed through Amazon's existing Merch on Demand print network. The feature is free to use (customers only pay for the product itself), and designs can be shared as links so other people can buy the same custom item from the original prompt.
The global print-on-demand market is roughly $10 billion and growing 25%+ annually, with Etsy alone clearing $11–12 billion in GMV last year on a long tail of small-batch designers selling exactly this type of custom apparel. Printful, Printify, Spreadshirt, Redbubble - the entire mid-funnel ecosystem of "I have an idea for a T-shirt" services - now sits below Amazon's voice assistant, with no design step, no upload friction, and no separate checkout.
The implication is that design as labor - the actual human work of mocking up a T-shirt graphic in Photoshop or Canva - has become a commodity input that AI handles for free, while the high-margin work (manufacture, fulfilment, customer relationship) consolidates further into Amazon's hands. This is the same flywheel TikTok GO and Snapchat's conversational ads built around discovery - the AI commerce loop of prompt, generate, share, ship - now living end-to-end inside one app, and the long tail of independent creators that powered the 2010s print-on-demand boom is the first casualty.
Google Search Just Hit an All-Time High of Queries - What AI?
Google's search market share still sits at roughly 80–90% globally three years into the ChatGPT era, with Q4 2025 Search revenue of $63 billion (up 17% YoY) and CEO Sundar Pichai citing an "all-time high" of queries last quarter. AI Overviews now appear on about 48% of tracked queries and reach more than 2 billion monthly users, AI Mode crossed 1 billion monthly users just one year after launch, and total global search usage grew 26% - meaning AI took a slice of a bigger pie rather than Google's existing slice.
ChatGPT itself holds roughly 17% of digital queries by Q2 2026 estimates, but its share has stopped growing as Google integrated generative answers directly into the surface a billion people open every day - the same playbook Facebook used to absorb Snapchat Stories in 2016, and Microsoft used to survive Google Docs (Microsoft 365 still holds ~87% of the productivity market in 2025). The pattern is consistent: when incumbents with distribution integrate disruptive technology silently, the "killer app" thesis breaks down - better technology loses to better distribution.
The strategic implication for the AI industry is uncomfortable: OpenAI's $25 billion ARR is real, but a standalone agent app now competes inside a saturated market against integration plays from Google, Apple (via the new Siri AI), and Microsoft - all of whom embed AI inside the surfaces consumers already open daily. The 2023 venture thesis that ChatGPT would dethrone Google funded a generation of "AI-first" apps; the 2026 reality is that the most valuable AI businesses will be the ones that hide inside the search bar, the inbox, and the OS - not the ones that ask users to learn a new product.