According to Tom’s Guide, Mullvad VPN has announced a new in-house WireGuard implementation called GotaTun, now available on its Android app. The core was rewritten in Rust to replace the previous wireguard-go code, which was responsible for over 85% of the app’s crashes reported via the Google Play Store. The Android rollout happened with version 2025.10 in late November, and the impact has been immediate with no crashes attributed to the new engine. Mullvad plans to introduce GotaTun on desktop and iOS platforms in 2026. Furthermore, the company will completely remove support for the older OpenVPN protocol on January 15, 2026. A third-party security audit for GotaTun is scheduled for early next year.
Rust vs Go: The backend battle
This move is way more technical than your average app update. It’s a fundamental shift in how Mullvad builds its core networking tech. The old wireguard-go was a userspace implementation in Go, a language known for its simplicity and concurrency. But here’s the thing: most of Mullvad’s other apps are written in Rust, a language prized for its speed and memory safety. Trying to make Rust and Go play nicely together was, in Mullvad’s own words, creating “limited visibility” into crashes and making long-term maintenance “tedious.” Basically, when your app crashes in Go-land, it’s harder to figure out why from your Rust-based control center. By rewriting WireGuard in Rust, they’re unifying their tech stack. That’s a huge win for stability and for the developers who have to maintain this code for years to come.
Why Android first, and who benefits?
So why launch on Android before the other platforms? It’s a pretty smart tactical move. Android’s ecosystem, with its vast array of devices and chipsets, is a brutal stress test. If you can make your VPN tunnel stable there, you’re in good shape. The fact that over 85% of their Play Store crashes came from the old Go code meant Android users were feeling the pain the most. Fixing it there delivers the most immediate user satisfaction. But look, this isn’t just about fixing bugs. Mullvad is betting big that a faster, more reliable Rust core will enhance performance across the board. They’re already one of the fastest VPNs out there in tests—imagine what a streamlined, native-Rust engine could do. This puts pressure on other privacy-focused VPNs. If you’re not investing in this level of low-level engineering, are you really in the performance game?
The OpenVPN sunset and what it means
The announcement to kill OpenVPN support in 2026 is a massive statement. OpenVPN is the old reliable warhorse of the VPN world. It’s everywhere, it’s trusted, but let’s be honest—it’s not exactly sleek and modern. WireGuard is the new kid, designed from the ground up to be simpler and faster. By fully committing to WireGuard (and their own superior implementation of it), Mullvad is ditching legacy baggage to focus all their energy on one modern protocol. For users on older devices or in highly restrictive networks that only allow OpenVPN’s specific ports, this could be a problem. But for the vast majority? It means a leaner, more focused app. It’s a gamble on the future, and it signals that WireGuard has well and truly won the protocol war in the commercial VPN space.
A tunnel-vision future
Mullvad’s entire strategy here is about control. They’re moving away from a forked third-party Go implementation to their own Rust engine. They’re ditching a legacy protocol they don’t control. Every part of the stack is being brought in-house. In a world where VPNs are often just reskinned versions of the same white-label software, this kind of deep technical investment is what separates the privacy leaders from the marketing-driven pack. The upcoming security audit is crucial for trust, but the early results on Android are promising. No crashes? That’s the kind of boring, reliable performance users actually want. If they can deliver that same rock-solid experience on desktop and iOS in 2026, they’ll have built a formidable technical moat. Other VPNs will be watching closely, probably wondering if they need to start their own Rust rewrite projects too.
