>
Open Source

Fedora on a Pixel, no Red Hat strings attached

I spent a long weekend flashing PocketBlue onto a spare Pixel 7a. PocketBlue is a community project that builds immutable Fedora images (an “immutable” image is one where the root filesystem is read-only at runtime, so updates are atomic and you cannot break the install by editing system files) for phones and tablets, and it is the first Fedora-on-phone project I have used that did not feel like a science experiment. It is not a daily driver yet. It is also not vaporware, which puts it ahead of most things labeled “Linux phone” in 2026.

The project, briefly

PocketBlue is run by a handful of developers on the Fediverse and the Fedora discussion boards. They build OSTree-based Fedora Atomic images (Fedora’s “immutable” variant, which uses OSTree for versioning and rpm-ostree for layering) for supported devices, flashable through the standard Android fastboot path. The list of devices is short, but the Pixel 6 and 7 series, the OnePlus 9, and a handful of Pixels are all there with working images. The project is not affiliated with Red Hat. The “(Without Red Hat)” framing in the title is the headline, not a marketing line: the maintainers are volunteers, the build pipeline is community infra, and Red Hat’s involvement is the same as any other Fedora contributor.

That distinction matters. Red Hat has been conservative about mobile for years, and for sensible reasons: the support cost of a phone is not the same as a server, and the user base is tiny. A community fork can move faster because it is not waiting on a corporate roadmap.

What works

I flashed PocketBlue to a Pixel 7a using the instructions on the project wiki. The whole process took about twenty minutes, including a wrong USB cable the first time. Fedora’s GNOME Mobile shell (a touch-optimized build of the GNOME desktop, designed for tablet and phone form factors) booted cleanly, the modem worked, Wi-Fi worked, and the touchscreen was responsive enough that I did not feel like I was debugging input latency. Calls worked. SMS worked. The camera did not work, which is the single biggest gap and the first thing on the project’s roadmap.

For the first two hours I had a Linux phone that I could actually call people from. That has not been true of any phone I have tried since the Nokia N900 in 2010.

A short list of what is real today:

  • Phone calls and SMS through the modem, with the standard GNOME Calls app.
  • Wi-Fi and Bluetooth working out of the box, no manual driver work.
  • Fedora’s package ecosystem via rpm-ostree (the tool that layers RPM packages onto an OSTree base image), with the same repos as desktop Fedora.
  • GNOME apps (Files, Calculator, Web, Maps) all running in their touch builds.
  • A real terminal in case you want to remember the 1990s.

The list of what does not work is longer, and I will get to it.

What does not work

The camera is the headline. The image pipeline is half-written, and the apps that ship with PocketBlue do not yet talk to the hardware correctly. You can take pictures with the default GNOME Camera app, but they are usually a blank frame or a green rectangle. For a phone in 2026, that is a deal-breaker for daily use.

Other gaps I hit in my weekend test:

  • Cellular data is flaky. SMS and calls worked on my carrier (T-Mobile), but LTE/5G data dropped to 2G half the time. The maintainers blame the modem firmware, which is binary-only and not under their control.
  • Most Android apps do not run. There is no Waydroid preinstalled (Waydroid is a container that runs a full Android image as a Linux app, letting you install Android apps from a normal package manager). If you need WhatsApp, you do not have a real solution.
  • Battery life is short. Roughly six hours of screen-on time on the 7a, which is about half what stock Android manages.
  • Updates are not signed with a hardware key. This is fine for tinkerers and bad for anyone who cares about verified boot.

The first three are fixable with engineering time. The fourth is a longer-term question about the project’s threat model.

Why this is interesting anyway

The interesting thing about PocketBlue is not that you should use it as your phone. It is that it exists at all, and that the Fedora Atomic model maps surprisingly well to phones. An immutable image plus an OSTree-based update path is exactly the architecture Android uses under the hood, minus the Google services layer. PocketBlue proves that you can build a Linux phone with the same update semantics as a Chromebook or a Steam Deck, and the result is a system that boots cleanly and does not rot over time.

A second interesting thing: the project is not trying to be a competitor to Android. The maintainers are explicit that this is a community project for people who want a Linux phone, not a mass-market push. That framing has kept the scope small and the morale high. There is no business plan. There is no Series A pitch. There are five or six maintainers and a working image.

What I would tell past me

If I could send a message back to the version of me that wiped a Nexus 5 trying to install Ubuntu Touch in 2015, I would say four things:

  • The tooling has finally caught up. Fastboot and OSTree do the heavy lifting. You are not building a kernel from source.
  • Flash a spare device, not your main phone. PocketBlue is not a daily driver. Treat it like a Raspberry Pi with a touchscreen, and your expectations will be calibrated correctly.
  • The community is small but responsive. I filed a bug about the camera on the project tracker and got a reply in three hours from someone who actually understood the issue.
  • Wait one more release cycle for the camera. The roadmap has camera as the next milestone. If the next release lands with a working image pipeline, this becomes a real daily-driver option.

Trade-offs

PocketBlue is not a daily driver. The camera does not work. The cellular data path is unreliable. Most of the apps you rely on do not have a Linux equivalent. The project is small enough that a single maintainer burning out could stall it for months. The image is not signed, so a malicious update server could in theory flash anything to your phone. None of this is a reason not to use PocketBlue if you are the kind of person who keeps a drawer of old phones for exactly this. All of it is a reason not to use it as your only phone.

In our case, I am keeping the 7a with PocketBlue as a travel phone for situations where I do not want to bring a corporate device. The battery life is fine for that, the camera does not matter for navigation and messaging, and I have a real Linux terminal in my pocket for SSH emergencies. For a person who wants a phone that works for everything, this is a year away from being a serious option. For a person who wants to be on the Linux phone train early, this is the train to be on.

Bottom line

PocketBlue is the most usable Fedora phone I have tried, and I have tried them all. It is not a Slack-and-Spotify daily driver, and the maintainers will tell you that themselves. If you want a Linux phone that boots cleanly, takes calls, and gives you a real terminal, this is the project to watch in 2026. Flash a spare device, file some bugs, and check back in three months. The camera is coming.

Leave a comment