A test suite that passes in CI says little about how an app behaves on a three-year-old Android phone with a weak signal. Mobile teams know the gap, and more of them are closing it on real hardware. Sauce Labs Real Device Cloud just completed over 5 million tests on real iOS and Android devices in a single month, a first for the platform.
What does real device usage data show about mobile testing in 2026?
With 10,000+ real iOS and Android devices in its fleet, Real Device Cloud offers a broad view of how mobile teams test. Usage from the past 12 months shows which devices teams test on most and how quickly their device matrices change. Use it as a benchmark for your own.
Which devices get tested the most?
The five most-tested devices on Real Device Cloud over the past year:
- iPhone 15
- iPhone 16
- iPhone 14
- iPhone 16 Pro
- iPhone 13
The iPhone 15 alone saw nearly three times as many tests as the No. 2 device. Android accounts for close to half of all tests, but the market is more diverse, so that volume spreads more evenly across a wider range of models, and no single Android device reaches the top five.
The list also skews older. Every device in the top five launched between 2021 and 2024, a sign teams weight their testing toward the phones their users already own.
Which devices are growing fastest?
The iPhone 17 Pro posted the biggest jump in test volume, growing more than sevenfold between its first winter on the market and this fall. The iPhone 17 more than quadrupled. On Android, the Google Pixel 9 Pro and Pixel 10 each nearly doubled, and the Redmi Note 14 Pro 5G grew 303%, the fastest Android growth among devices with steady volume.
How fast do teams adopt new devices and OS versions?
Fast. Apple released iOS 27.0.1 on September 28, 2026, and Real Device Cloud customers were testing on it within a day. Samsung’s Galaxy S26 went on sale in March, and by fall it ranked among the 10 biggest volume gainers on the platform.
How long do teams keep testing older devices?
For years. The iPhone 13, released in 2021, is still the fifth most-tested device, and its test volume grew more than a third over the past year. The iPhone 12, iPhone 13 Pro, Galaxy S21 5G, and Pixel 6, all released in 2020 or 2021, remain in the top 30. Older OS versions stick around too: Teams were starting new test runs on Android 8, a 2017 release, as recently as August 2026.
How does your matrix compare? If your most-tested devices skew newer than this list, or new OS releases sit untested for weeks, your coverage may not match the devices your users carry.
What can teams do on Real Device Cloud that they couldn’t a year ago?
Teams can now program real devices directly, test with real screen readers, connect device results to Sauce Labs AURA’s autonomous learning loop, and let AI agents drive devices through MCP.
Program devices with the Real Device Access API
The Real Device Access API turns private devices into programmable infrastructure. Teams control a device over HTTP and WebSockets to install apps, run ADB shell commands, proxy network traffic, and pull logs, with no test framework in the middle. Platform and AI teams use it to build custom orchestration and give agents device-level control.
Test accessibility with real screen readers
iOS live testing now includes a built-in Accessibility Inspector for VoiceOver validation, and Android live testing turns on TalkBack with one click. Automated scans run inside Appium suites through Deque axe DevTools Mobile. With European Accessibility Act enforcement underway, teams can confirm screen reader behavior on real hardware before an audit does.
Close the loop with AURA
Since AURA launched in July, Real Device Cloud has served as its test stage. Agents author tests from business intent, Real Device Cloud runs them on real devices, AI-driven analysis surfaces root causes, and production errors feed the next round of tests, with humans in control.
Bring real devices into agentic workflows
The Sauce Labs hosted MCP server connects Claude, Cursor, Windsurf, and other MCP-compatible clients to Real Device Cloud with nothing to install. A developer can ask an agent to find a device, install a build, start a session, and collect logs without leaving the AI assistant or IDE environment.
Why does real device testing keep growing in the AI era?
Verification hasn’t kept pace with AI coding assistants. Sauce Labs calls that shortfall the AI code verification gap. Release speed is its own bottleneck: Only 3% of organizations in the State of Mobile App Quality 2026 report release updates multiple times per week, and nearly half release monthly. Leaders are responding, with more than 60% planning to integrate AI to predict defects and generate tests.
Real devices are where much of that verification happens. Network conditions, memory limits, OS fragmentation, and hardware quirks surface on physical devices first, and AI-written code doesn’t get a pass on any of them.
What’s next for Real Device Cloud?
With AI generating more code and teams adding quality gates to verify it, Sauce Labs expects tests per app to double, pushing Real Device Cloud past 10 million tests a month within the next 12 months. Credit for the first 5 million goes to the customers who made real devices part of their daily workflow and to the Sauce Labs teams keeping the platform fast and stable.
Want to see how your suite performs on real hardware? Start a free trial or request a demo.




.png)
