Nintendo Switch 2 reportedly jailbroken via old SmileBASIC 4 game and OpenAI's Codex

What happened

The Nintendo Switch 2 has reportedly been jailbroken, and the modder behind the exploit says an AI coding assistant helped make it happen.

According to The Verge, which reported the news on October 22, 2026, the jailbreak — dubbed "ns2jb" — was built with help from OpenAI's Codex. That credit comes from Sammy Hajhamid, a developer who goes by "pepsipu" online.

How the exploit works

The Verge describes the jailbreak as software-only and tethered. Tethered, in this case, means the exploit does not persist across reboots: anyone using it has to launch the hack from within a game every time they power on the console for it to take effect.

The entry point is SmileBASIC 4, an older title that lets players write and run their own programs. That programmability appears to be what makes the game a usable foothold into the system. The Verge adds a caveat, though: another modder claims the game is not strictly required, which suggests the underlying vulnerability might be reachable through other means.

Early proof of concept

The jailbreak already has a showpiece. The Verge reports that a modder demonstrated the console running a modified version of Mario Kart World with unofficial playable characters, including Cinnamoroll, the red Angry Bird, Peter Griffin, and Steve from Minecraft. Custom characters injected into a first-party Nintendo racing game are a clear signal that code outside Nintendo's control is executing on the hardware.

How significant is it

For now, the practical reach looks limited. The Verge notes that the process still seems far too complicated for most users, and that the full consequences of the exploit are not yet clear. Members of the Switch 2 hacks subreddit are actively debating how much of a milestone this actually represents.

That said, The Verge points to the possibility that this is only the first of many exploits for the console — particularly if modders keep using AI tools to assist with development.

Why it matters

Three things make this story worth following.

First, it is a genuine security incident on a flagship console. A software-only break means a fixable vulnerability exists in shipped code, and it kicks off the familiar cat-and-mouse cycle between Nintendo's firmware updates and the modding scene. Historically, once one foothold is found on a console, refinement and additional exploits tend to follow.

Second, there is the AI angle. If Hajhamid's account is accurate, Codex played a meaningful role in developing a real-world console exploit. That points to a broader shift in security research: the same AI coding tools that accelerate legitimate software development can also accelerate the discovery and weaponization of flaws. That dynamic applies well beyond one gaming platform, and it raises questions for every vendor shipping software that attackers — or researchers — can now probe with machine assistance.

Third, the tethered design cuts both ways. Requiring the exploit to be re-run at every boot limits casual adoption and softens immediate risks such as large-scale piracy or cheating. But it does not eliminate them, and a tethered break is often just an early step toward something more convenient and persistent.

The practical takeaway for now: the Switch 2 has its first public crack, the modding community has its proof that custom code runs, and the variable to watch is not this exploit alone but how quickly others — possibly AI-assisted — arrive behind it.