Why VR's Next Big Games Are Modded, Not Published

The most wanted VR port of the week was made by one person and an AI agent

Astro Bot Rescue Mission is one of the best reviewed VR games ever made, a 2018 PlayStation VR exclusive that Sony never ported to PSVR 2 and never brought anywhere else. This week it runs natively on Quest 3. Modder Bigmak94 got it there by pointing an AI coding agent at shadPS4, the open source PlayStation 4 emulator, and asking for two builds: a standalone Quest 3 version you sideload on the device, and a PC VR version you stream to any headset. The port has 6DOF headset tracking, native controller mapping, and immersive spatial audio, and it runs the whole game, not a demo.

Nobody licensed it. Sony did not ask for it. And that is the point.

Read the week’s news as a whole, and the shape of VR’s content supply becomes visible. On one side, the studios that are supposed to build this market are contracting: Mighty Coconut, the studio behind the genre-defining Walkabout Mini Golf, announced a second round of layoffs in nine months, about 35 percent of staff, following a 25 percent cut in January and a DLC price rise from $4 to $5. On the other side, individual modders are shipping full conversions of games that publishers have abandoned, and they are getting there faster than usual because AI coding agents have collapsed the cost of reverse engineering. Astro Bot on Quest 3 is not a curiosity. It is the clearest signal yet that the demand for VR games and the ability to supply them have decoupled from the companies that own the rights.

Why an emulator can be a better porting strategy than a port

The interesting thing about Astro Bot running on a Quest 3 is not that it happens. It is that it happens on a mobile chip.

Quest 3 runs a Snapdragon XR2 Gen 2 with 8GB of RAM, which is a phone-class part doing a job that a PlayStation 4 did with a much larger thermal envelope. Running a PS4 game is worse than running a Quest-native game on that hardware, because an emulator adds a translation layer on top. shadPS4 is a C++ emulator for Windows, Linux, macOS and FreeBSD that recompiles the PS4’s AMD GPU shader programs onto Vulkan and reimplements the console’s libraries and OS surface. The project is GPL-2.0 licensed, has 33.1k GitHub stars, and its shader compiler was built with yuzu’s Hades compiler as a blueprint, which tells you where the difficulty lives: emulating a modern AMD GPU’s instruction set and memory model correctly enough that a shipping game’s own optimizations still work.

Now stack VR on top. Every frame has to be rendered twice, once per eye, with a slightly different camera transform, which either doubles the draw work or forces a single-pass stereo trick the original renderer was never written for. The emulator has to hold a stable frame time under a reprojection deadline, because in VR a dropped frame is not a stutter, it is nausea. And the whole thing has to fit in a mobile GPU’s tile-based rendering pipeline, which behaves very differently from the PS4’s desktop-style immediate-mode GPU when it comes to render target switches and bandwidth. Bigmak94’s own report is blunt about the result: the Quest 3 build is fully functional, but the headset’s mobile GPU is “pushed to its limits,” and the standalone version runs at lower resolution and fidelity than either the original PS4 or the PC VR build. That is the honest trade. The translation tax is real, and on a standalone headset it is paid in pixels. Pixel budget is the resource that decides how a headset feels, which is why Quest 3 versus Steam Frame comparisons come down to optics and memory bandwidth as much as resolution.

The PC VR side is different. There, the same mod streams to a headset from a desktop GPU, and the emulation tax disappears into headroom. That is why these mods matter disproportionately for PC VR owners, and why Valve’s open platform is the natural host for them. Nobody is sideloading an emulator onto a locked console.

The mod scene has better distribution tooling than some stores

The Astro Bot release landed days after a second, equally instructive project: a full VR conversion of Titanfall 2 from modder CircuitLord, distributed through a one-click installer that also handles Big Walk VR. The installer finds your copy of Titanfall 2 across Steam, EA or Xbox installs, checks the EA app dependency, and installs the mod plus the Northstar client into a separate profile directory so that launching the game normally stays unmodded. Updates are driven by a signed-off manifest that lists each release with a download URL and a SHA-256 hash; the installer compares that against what you have installed and offers Install or Update. It records every file it writes, deletes files that a newer version no longer ships, and uninstalls exactly what it put down. There are optional beta channels.

That is release engineering. A commercial store would call it a patching system with rollback hygiene. What the mod itself contains is the other half of the story: stereo rendering, full body IK, manual reloads across the full weapon roster, Titan controls, and head tracking layered onto a shooter whose entire design language is high speed movement and wall running. The installer is MIT licensed, but the mods themselves are closed source and all rights reserved, and the project page states plainly that these are community projects, “not affiliated with or endorsed by the game developers or publishers.”

Put the two releases next to each other and the pattern is hard to miss. The unpaid side is producing ports of beloved games with versioned installers, hash verification and clean uninstalls, while the paid side consolidates the pitch: the pipeline argument was made at length in Flat2VR’s Spark initiative, which set out to institutionalize conversion work rather than leave it to lone hobbyists.

The studios are not losing to modders. They are losing the economics

It would be tidy to end there, with scrappy modders beating slow publishers. The rest of the week’s news refuses that ending.

Mighty Coconut’s announcement is the number to sit with. The studio cut 25 percent of staff in January, which left it at around 27 people, then cut roughly 35 percent more in late September, with CEO Lucas Martell writing that ten people would be leaving and that “without a steady stream of new players coming into VR and Walkabout, the scale at which we’ve been operating is no longer sustainable.” Note what he is not saying. He is not saying Walkabout Mini Golf stopped selling, or that players stopped playing. He is saying the player base is not growing fast enough to fund the size of the studio that the game’s production values require. Walkabout is one of the most successful VR-native franchises in existence, and it is still shrinking toward the size it started at. Its revenue diversification is instructive too: one hybrid VR and flatscreen roguelike, and one flatscreen co-op game about lawn care.

This is not an isolated case, it is the environment. Meta’s Reality Labs division cut about 10 percent of staff this year, taking Sanzaru Games, Armature Studio and Twisted Pixel with it. A Harry Potter VR game for Quest was cancelled, as was a follow-up to Batman: Arkham Shadow. Polyarc, whose Moss games are among the most decorated in VR, announced it was shutting down after cancelling a major project. The longer argument about platform economics and who absorbs that risk is the one I made in why VR’s best studios are shutting down, and nothing this week contradicts it.

Against that backdrop, a volunteer port of a Sony exclusive is not a distribution channel. It is a symptom that reads as a gift.

What official VR content looks like when it is built for the platform

There is a live counterexample in the same week, and it is worth studying because it shows the shape that funded VR content is taking.

EVANGELION: Δ CROSS REFLECTIONS is an XR game from Pixelity built on the Neon Genesis Evangelion franchise, demoed on Quest 3 and slated to arrive in installments next year, with the whole thing following the 26-episode series. Its combat is entirely hand tracked: you punch your hands forward to dash, raise a thumb to reverse, hold your palm open to grip an invisible rifle, pull an air trigger with a curled index finger to fire, and raise both hands to bring up a shield. Road to VR’s hands-on calls it surprising and effective, with the caveat that the gesture set is still under research. The interesting part is not the gestures. It is that the game was designed around hands, seated scale and directed cinematic framing from the first frame rather than retrofitted at the end.

That is the same bet Meta is making with the VR Glasses era, which I wrote about in Meta’s controller-free VR Glasses: hand tracking as the default input, a $1 million developer competition whose only hard requirement is that an entry be playable end to end with hands, and a launch slate weighted toward puzzle, tabletop, management and narrative genres where coarse input and a narrow field of view cost the least. The mod scene and the official slate are converging on the same design constraints from opposite directions.

The official release calendar, meanwhile, looks thin. October’s slate on Quest, PC VR and PS VR2 is dominated by small indie releases and ports (a BMX game, a horror puzzler, a birdwatching game, a prison break comedy), alongside a PS VR2 version of Into the Radius 2 that arrives with eye-tracked dynamic foveated rendering at a native 90fps to make the port viable. Those are respectable releases. They are not the volume you would need to grow a platform, which is the underlying problem: the roundup covering the VR mod wave led with the same two mods and the same Evangelion demo, and the reason they dominate the conversation is that there is not much else in the queue.

The catch: mods are a demand signal, not a supply chain

The temptation is to conclude that modding solves VR’s content problem. It does not, for four reasons that are worth stating plainly.

The first is legality and durability. Every one of these mods requires you to own the original game and dump your own copy. They exist in a grey zone that a single legal letter can end, and the Astro Bot port leans on an emulator that reimplements Sony’s console libraries. The project authors say out loud that they are unaffiliated and unsanctioned. Nothing about that is a foundation.

The second is fidelity and performance. The Quest 3 build of Astro Bot is fully playable, but it is running at reduced resolution and visual quality because the emulation tax has to be paid somewhere, and in VR the bill always lands on frame time. That means the standalone version is not the definitive way to play the game, it is the convenient one, and the experience differs by hardware in a way that a native port would never allow.

The third is that mods do not add players. This is the part the Mighty Coconut numbers make unanswerable. If the constraint were a shortage of games, a wave of free high quality conversions would grow the market. But Martell’s diagnosis is the opposite: the problem is “a steady stream of new players coming into VR.” A perfect Astro Bot port played by existing Quest 3 owners does not bring in anyone who was not already there, and it earns nobody a living. It is evidence of unmet demand, not a way of meeting it.

The fourth is what happens when an AI agent lowers the cost of doing the work. The reason this wave is happening now is that Claude could do in hours what used to take a skilled modder months of binary archaeology. That same cost collapse applies to studios, and it applies to the old argument that porting a flatscreen game to VR is too expensive. If a solo modder can retrofit 6DOF tracking, stereo rendering and manual reloads onto a shipped shooter, a studio with the source code has no technical excuse left, only a business one. None of that is unique to modding. Tooling keeps getting cheaper faster than business models adapt, the same pattern visible in the XR rendering tooling and Godot roundup from three weeks earlier.

Where this goes

Three things to watch from here, in order of how much they matter.

Whether any publisher treats modding as a market signal instead of a nuisance. Sony has a critically adored VR exclusive that the mod scene just proved is playable on Meta’s hardware. That is a porting decision sitting on the table with a free proof of concept attached.

Whether the hybrid release path becomes the norm. Walkabout’s own survival plan points the same way as Into the Radius 2 on PS VR2: build once, ship flatscreen and VR, and let the VR mode ride on the same asset and code budget. The studios that survive this contraction will be the ones that stop treating VR as a separate market to fund alone.

And whether the platforms keep the door open. Steam Frame’s openness is what makes modded VR practical, since side loading emulators and unsigned builds onto a phone-class headset is the only reason any of this works. Every headset that closes that path gives up the free content pipeline that the industry is currently relying on without admitting it.

The most interesting VR release of October has no publisher, no budget and no launch date. That is not a healthy sign for the industry. It is a very clear one about where its players are.

---

Sources

• Road to VR: PSVR Exclusive ‘Astro Bot Rescue Mission’ Running on Quest 3 is the Latest in VR Mod Explosion
• Road to VR: ‘Titanfall 2’ VR Mod Brings the Legendary Movement Shooter to VR
• GitHub: CircuitLord’s VR Mod Installer
• GitHub: shadPS4 PlayStation 4 emulator
• Road to VR: ‘Walkabout Mini Golf’ Studio Announces Another Round of Layoffs Following DLC Price Hike
• Mighty Coconut: Mighty Coconut Update
• Road to VR: ‘Neon Genesis Evangelion’ VR Hands-on: Hand-Tracked Combat That Actually Works
• UploadVR: New VR Games & Releases October 2026
• Road to VR: Meta VR Glasses, Quest 3 & Steam Frame Compared