Analysis

Rust embedded development matures as teams weigh safety and control

Rust embedded has crossed into ship-now territory on STM32 and ESP32, but probe setup, wireless gaps, and safety certification still draw a hard line.

Nina Kowalski··4 min read
Published
Listen to this article0:00 min
Rust embedded development matures as teams weigh safety and control
Photo illustration
This article contains affiliate links, marked with a blue dot. We may earn a small commission at no extra cost to you.

probe-rs now targets ARM and RISC-V devices from a host machine and can stream output through RTT and defmt when code runs on the board. In Derek Molloy’s July 15 survey, embedded Rust looks ready for mid-2026: bare-metal async on microcontrollers is real, embedded Linux is in the picture too, and the remaining friction sits in the toolchain, safety story, and wireless gaps.

What has changed since the last wave of optimism

Debugging is the biggest practical upgrade. The old embedded-Rust problem was never only syntax or borrowing; it was whether the flashing and logging path felt like a daily tool instead of a lab exercise.

The remaining rough edges are narrower and more recognizable. The Embedded Rust Book still treats microcontroller work as remote-device programming, which is exactly the right mental model, and cargo-embed can still fail to find a supported probe, so setup can still depend on the exact board, chip ID, and target configuration. In other words, the happy path is much better than it was, but “plug in and it just knows” is still not the default everywhere.

The ecosystem now has a governance spine

Rust embedded is not being held together by scattered enthusiasm alone. The Rust Embedded Devices Working Group, first announced on the Rust Internals forum in 2018, exists to improve the embedded ecosystem, maintain and develop core crates, and coordinate community communication. It defines Embedded Rust broadly as Rust for resource-constrained environments and hardware-level work, which matches how people actually use it now, from tiny MCUs to edge systems.

Embedded teams care about boring stability more than abstract momentum. heapless v0.9.1 landed on August 20, 2025, nearly two years after the previous release, which is a good reminder that key no-alloc crates are alive but not always fast-moving. The Rust Blog posted a crates.io development update on July 13, 2026.

Where Rust is ready to ship now

STM32: the clearest bare-metal lane

If you want the least surprising path for a fresh microcontroller project, STM32 is still one of the strongest choices. embassy-stm32 is a no_std HAL for ST’s STM32 series, and it has been moving in 2026 with releases 0.5.0 on January 4 and 0.6.0 on March 20, while also adopting Rust 2024.

Pair that with the Embassy framework and you get the async story embedded Rust spent years promising. In a February talk on device-envoy at the Seattle Rust User Group, Carl Kadie described an open-source Rust library built on Embassy for application-level programming on bare-metal microcontrollers, with demos including LED panels, Wi-Fi auto-configuration, and composable device abstractions.

ESP32: the vendor-backed route is now obvious

Espressif’s Rust stack looks substantially more mature than it did when esp32_hal was the headline. The old crate is deprecated in favor of esp-hal, and esp-idf-svc adds type-safe wrappers for Wi-Fi, networking, httpd, and logging. That combination gives ESP32 projects a more credible path from hobby board to product firmware, especially when connectivity matters.

The ecosystem around ESP32 is also visibly active in public. FOSDEM 2026 included “From C to Rust on the ESP32: A Developer’s Journey into no_std,” which is the sort of migration talk that only happens once people are actually shipping work on the platform. The Espressif ESP32-C3-DevKit-RUST-1 board is among Ariel OS’s supported boards.

Other supported targets

Ariel OS currently provides drivers for ESP32, nRF, RP, and STM32.

What is still painful

  • Wireless is still a gap worth planning around if your project depends on radios, pairing flows, or vendor-specific networking glue.
  • Safety certification is now part of the conversation, but not a solved problem. The Rust Blog published “What does it take to ship Rust in safety-critical?” on January 14, 2026.
  • Tooling is much better, but board-specific friction still appears. The cargo-embed probe-detection problem on an Adafruit ItsyBity M0 Express is exactly the kind of setup snag that can still eat an evening before your first successful flash.
  • Crate cadence is uneven in the places people rely on most. heapless finally moved again after a long gap, but it also shows that core embedded crates can still advance on a slower rhythm than mainstream desktop libraries.

The production backdrop is more reassuring than it was a year ago. Nearly half of all companies now use Rust in production, by The New Stack’s March 13, 2026 count, and that broader adoption makes embedded evaluation easier to justify inside real teams.

This article was produced by Prism’s automated news system from verified source data, official records, and press releases, then run through automated quality and moderation checks before publishing. The system is built and supervised by the people who set the standards it runs under. Read our full AI policy.

Did this article answer your question?

Discussion

More Rust Programming News