Ubuntu 26.10 ships a not-yet-final 7.3 kernel and completes the Rust coreutils migration

October 15, 2026. Ubuntu 26.10, codenamed Stonking Stingray, reaches general availability. Nine months. It is an interim release, supported until July 2027. October 18, 2026. The Linux 7.3 kernel it ships is not out yet: the GA image runs package 7.3.0-9.9, built from 7.3-rc6. Why it matters: this release closes three bets Canonical opened across the cycle — a kernel ahead of mainline, the end of the coreutils to Rust migration, and the arrival of Sequoia in the main archive.

The kernel first: Ubuntu jumps ahead of the official release

The most disorienting fact about 26.10 shows up in uname -r: a 7.3.0-9-generic kernel while Linux 7.3 is not yet tagged final. Canonical originally targeted 7.2, then raised the target to 7.3 mid-cycle, betting on the calendar: the final 7.3 release is expected on October 18, three days after Ubuntu’s GA.

The trade-off is deliberate. The release image carries a kernel built from 7.3-rc6 — in effect “the 7.3 series without the final tag,” as release-day upgrade testing put it. Canonical will fold the final fixes into later updates once the 7.3 tag lands. This is a scheduling choice, not an oversight: for an interim release aimed at early adopters, having the newest kernel matters more than a tag’s guarantee. But it is also a signal worth reading twice — the release-candidate window caught a silent data-loss bug earlier this month, fixed in later RCs, and a GA image built from an RC carries mechanically less hindsight than the final kernel.

The Rust coreutils migration is complete

The real under-the-hood event is elsewhere. Ubuntu 26.10 is the first release where the entire coreutils set, including the destructive trio cp, mv and rm, runs on the Rust implementations from the uutils project. On a fresh install, cp is a symlink into uutils’ cargo tree:

ls -l $(which cp)
# /usr/bin/cp -> ../lib/cargo/bin/coreutils/cp
cp --version | head -1
# cp (uutils coreutils) 0.12.0

The migration began in 25.10, and 26.04 LTS deliberately deferred it by keeping GNU coreutils. Stonking Stingray completes the arc, but reversibly: the rust-coreutils (0.12.0) and gnu-coreutils (9.10) packages are co-installed, and the coreutils meta-package flips the default toward the Rust build. It is a default flip rather than a forced removal, which is why the transition has been quiet in practice.

The ordering is not accidental: the lower-risk tools moved first, and the destructive trio moved last, precisely because they demanded the most testing. A badly behaved rm deletes data; a cp that changes behavior around attributes or links breaks scripts. That Canonical chose this interim release — not the LTS — for the default flip is the right engineering call: you experiment on an early-adopter audience, not the production fleet.

Sequoia enters main, and RISC-V becomes official

The third bet is cryptographic. Sequoia, the Rust-based OpenPGP toolkit, lands in the main archive for the first time: sq 1.4.0, backed by sequoia-openpgp 2.4.0 and the OpenSSL 4.0.1 backend, coexists with GnuPG rather than replacing it. It is a slow, cautious migration step, in line with the vendor’s Rust strategy — gradually swapping historical C components for memory-safe implementations without breaking existing workflows.

Alongside it, Canonical publishes official RISC-V desktop images for the first time, conforming to the RVA23 profile. That is an industrialization signal for an architecture long confined to development boards: an official desktop image is a reproducible test and deployment path, not symbolic support.

The rest of the cycle is more conventional: GNOME 51, still Wayland-only, drops WEP Wi-Fi configuration and makes open kernel modules the only NVIDIA path for Turing and newer cards; Myna, the speech-to-text tool, arrives as a preview without being preinstalled. One detail deserves a caveat: the release notes list dbus-broker as the default D-Bus message bus, but a machine upgraded from the beta still runs the classic dbus-daemon — the result depends on the install path.

What this means for your machines

The practical read depends on the fleet. For a production server, nothing here justifies leaving 26.04 LTS: 26.10 support ends in July 2027, and the LTS keeps receiving HWE kernel updates. For an adopter desktop or a homelab, it is an ideal test release: Rust coreutils in real conditions, a 7.3 kernel ahead of the pack, and Sequoia within reach.

Two precautions before migrating. First, if you script against exact GNU coreutils behavior, keep an eye on cp, mv and rm: the residual behavior differences are documented by the uutils project, and the gnu-coreutils package remains available as a fallback. Second, accept that the kernel is an rc6 dressed as a stable series: the final 7.3 fixes will arrive through normal updates, and a 26.10 machine is, by construction, one phase ahead of the official tag.

Verdict

If you run production, stay on 26.04 LTS and leave 26.10 to the adopters: an interim release is not worth a migration for nine months of support. If you are a Linux desktop or homelab enthusiast, jump on 26.10: it is the first chance to validate Rust coreutils by default and a 7.3 kernel before everyone else, with Sequoia on top — provided you treat that kernel as a reprieved RC, not a frozen release.

References

• Ubuntu Announce — Ubuntu 26.10 (Stonking Stingray) released (October 15, 2026)
• FOSS Linux — Ubuntu 26.10 Stonking Stingray: Every Major Change Explained (October 15, 2026)
• 9to5Linux — Ubuntu 26.10 (Stonking Stingray) Is Now Available for Download
• Ubuntu Discourse — Announcing 7.2 Kernel for Ubuntu 26.10 Stonking Stingray (updated to 7.3)