Expo, which makes an open-source framework and cloud services for building iOS, Android and web apps, launched Expo Simulators on 12 November: “cloud iOS simulators and Android emulators built for AI coding agents”, its press release said. Pricing is usage-based at $0.10 per minute for iOS and Android, which comes to $6 per simulator-hour by our arithmetic.
Expo names Claude Code, Codex and Cursor as agents it works with, and says it “adds no AI model of its own”. The service is available on all paid plans, and the launch blog says the Starter, Production and Enterprise plans include 60 minutes per month until 2027.
Business Pill · AI Agents Explained
An evergreen Business Engineer explainer on AI agents that act rather than answer. It teaches the general idea only and does not discuss Expo or its simulators.
The key insight: As we read the launch, Expo is pricing the step after the code: a cloud simulator an agent can drive and record, at $0.10 a minute, so a reviewer watches a run instead of building the branch.
Each session runs on its own virtual machine, “so parallel agents never share a simulator or clobber each other’s work”, the blog says. “The agent installs your build, drives the app the same way it drives a local simulator, and brings back proof.”
The proof is a record. The blog says each session has a “full screen recording, screenshots, and logs” of everything the agent did, which the agent can attach to a pull request. Anyone on the team can open a running session from a link in a browser, with nothing to install, Expo says.
A session starts with one command, “npx eas-cli@latest simulator –platform ios”. The agent drives it with the same agent-device, Argent or Appium commands it uses locally, and Expo says any agent that can run CLI commands works with it.
Expo Simulators’ list price of $0.10 a minute applied to common runs. The dollar figures are our arithmetic; the 115-minute cap and 60 included minutes are Expo’s.
The Problem Expo Is Pricing
Expo frames verification as the new bottleneck. “iOS simulators run only on Apple hardware, while cloud agents typically run on Linux”, the release says. With Expo Simulators, teams can run the iOS simulator on Linux, on Windows through WSL, or in a cloud sandbox, it says.
The release also puts a number on the laptop limit: “running seven agents in parallel on a laptop can require seven simulators or emulators, and Android emulators can use 4+ GB of RAM each”. That is 28 GB or more of memory for seven Android emulators, by our arithmetic.
Expo’s CTO and co-founder James Ide put the pitch in one line: “The agent was the fast part.” He said a fix “arrives for review already tested on several simulators, and the only decision left is whether it’s the right code”.
Memory for parallel Android emulators on one machine at Expo’s figure of 4+ GB each; the seven-emulator case is Expo’s example. Totals are our arithmetic.
What It Costs
At $0.10 a minute, an hour on one simulator costs $6 by our arithmetic. Expo’s billing documentation caps a paid-plan session at 115 minutes, so one session at the cap costs $11.50. Seven agents running for an hour on separate sessions would cost $42, and the 60 included minutes are worth $6 a month at the list rate, by the same arithmetic.
Sessions use the account’s existing job concurrency and queue when none is free, the documentation says. Teams can set a maximum duration and an idle timeout so “a forgotten session stops on its own”, the blog says.
Expo says it will follow with Agent Sessions, “a secure layer that lets enterprises trigger agents from sources such as customer feedback”, and “expects to share details by the end of November”. The blog says a group of design partners used the service for the past few months before launch.
The Structural Read
As we read the release, the product answers a hardware gap more than a model gap: iOS simulators need Apple hardware and cloud agents typically run on Linux, the release says, and Expo adds no AI model of its own.
As we read the billing page, the meter and the caps are the design: a 115-minute session limit, idle timeouts and queues on existing job concurrency keep agent runs from billing indefinitely.
James Ide, co-founder and CTO of Expo, press release, 12 November 2026
“The agent was the fast part.”
Three Implications
FOR MOBILE TEAMS USING CODING AGENTS Expo says sessions run from Linux, Windows through WSL or a cloud sandbox, so iOS checks no longer need a Mac on the agent’s side.
FOR AGENT TOOL MAKERS Expo names Claude Code, Codex and Cursor and says any agent that can run CLI commands works with the service.
FOR OUR COVERAGE We will add customer or usage figures, and details of Agent Sessions, if Expo publishes them.
The framework puts it this way: “Universal harnesses are likely to dominate highly digital, modular, and verifiable work.”
As we read it, a simulator that records each run makes mobile app work more verifiable for whichever agent drives it. The framework predates the launch and does not discuss Expo.
What Is Not Established
Expo gives no customer count, session count or revenue figure for the service, and names no design partners. It publishes no comparison with other cloud device services, and we make none.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
Expo now rents AI coding agents a cloud iPhone simulator or Android emulator at $0.10 a minute, $6 an hour by our arithmetic, so agents such as Claude Code, Codex and Cursor can install a build, tap through it and hand back a recording. Expo frames checking an agent’s work, not writing the code, as the slow step.
A note on sourcing. We describe Expo Simulators from Expo’s launch blog and press release of 12 November 2026 and its usage, limits and billing documentation. Prices, limits and the 4+ GB figure are Expo’s. The hourly, per-session, seven-agent and RAM totals are our arithmetic on those figures.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
