Garry Newman has been building game creation tools long enough to know that day-one chaos isn't necessarily a bad sign. Still, the reception to S&box's recent launch must have stung a little.
Within 24 hours of going live on Steam, the spiritual successor to Garry's Mod—the decade-old sandbox that launched a thousand YouTube careers—had generated nearly a million dollars in revenue. Facepunch Studios, Newman's outfit, shared their financials publicly through a dashboard—an uncommon practice in the industry. The numbers looked good. Better than good, really, for a $19.99 user-generated content platform entering a market already dominated by Roblox's billions and Epic's Fortnite Creative ambitions.
But Steam's review section told a messier story. "Mixed" reviews. Complaints about discovery. And a flood of what Newman himself called "AI slop" clogging the platform's front page.
That tension—between immediate commercial success and community skepticism—may define whether S&box becomes a genuine alternative in the increasingly crowded space where platforms court creators, or just another cautionary tale about what happens when you open the floodgates without enough guardrails.
A Different Bet Entirely
The pitch Facepunch is making differs sharply from the Roblox playbook. While Roblox paid creators $923 million in 2024 alone (and crossed the billion-dollar threshold in the first nine months of 2025, according to industry analysis), it keeps those creators firmly inside its walled garden. Everything lives on Roblox. Everything runs through Robux.
S&box offers something closer to liberation, at least in theory. Creators can export their projects as standalone executables and publish them directly on Steam. No engine royalties to Facepunch. The company secured a license with Valve specifically to enable this pathway, though it launched in preview mode and still requires case-by-case approval while they scale the system.
It's a compelling pitch for developers exhausted by platform revenue shares. Build your game in S&box, take it to Steam, keep more of what you earn. The engine itself went open source under an MIT license late last year. For creators who've chafed under Roblox's economics or Epic's Unreal Engine fees for Fortnite Creative, it represents a fundamentally different arrangement.
Whether it works is another question.
The Numbers Game
Facepunch's unusual commitment to financial transparency offers a rare window into the economics of launching a UGC platform. Their public metrics page showed their launch month brought in $2.6 million in gross revenue—split roughly evenly between game sales and in-game item purchases. After refunds, the studio cleared just over $200,000 in profit.
That margin might seem thin for a million-dollar opening day, but context matters. Facepunch had spent five years developing the platform and run an 800,000-person developer preview that winnowed down to 40,000 retained accounts by launch. The studio responded to early traction by doubling its Play Fund—the pool of money distributed to creators based on engagement—to $1 million annually.
The payout system tracks player hours, with earnings "roughly clamped" per player according to the platform's monetization documentation. (The language there is deliberately vague, perhaps understandably.) As of early returns, the live dashboard showed nearly $500,000 distributed to the community. Top earners included titles like "Mow the lawn," "DXRP," and "Case Opening Simulator"—a range that suggests, well, exactly the kind of ecosystem you'd expect from a platform anyone can publish on.
Pre-launch, Facepunch had already pushed over half a million dollars to S&box creators, funded initially through Garry's Mod revenue. The plan moving forward hinges on growing that fund through S&box sales and its cosmetic economy, where creators sell items through Steam Workshop.
The financial model appears sustainable, at least on paper. Facepunch made a point of noting—in what felt like a pointed jab at the broader games industry—that "we don't have to fire 1000 people to keep it working." Whether sustainability translates to scale is a different matter entirely.
The Technical Stack
S&box arrives with a comprehensive technical toolkit: C# scripting with hot-reload, built-in multiplayer and party systems, VR support through OpenXR, Valve's Hammer level editor. There's visual scripting through something called Action Graph, shader tools, and a system for mounting and converting assets at runtime from other Steam-owned games—though those assets can't be republished, an important limitation.
No subscriptions required. No separate installations to play community games. Just one-click access, in theory.
The studio made deliberate choices on rendering philosophy that sparked immediate debate. They're avoiding temporal upscaling and anti-aliasing in favor of forward rendering with MSAA—a technical decision that led to performance complaints but reflects, Newman would argue, a particular vision for visual quality. "We love it when things break because then we get to fix them," he said when addressing launch issues, a comment that sounds either refreshingly candid or worryingly cavalier depending on your perspective.
When the Slop Hits
And then there was the AI problem.
Steam reviews in the early window landed at 46% positive among English-language users—2,969 total reviews across all languages as of early May 2026. The complaints centered on discovery: a front page flooded with AI-generated thumbnails and low-quality content that made finding the good stuff nearly impossible.
PC Gamer's assessment was particularly harsh, calling the platform "tragically more Roblox than anything else" despite Facepunch's different technical approach. The criticism highlighted creator payout models that resembled Roblox's engagement-based system and discovery UI issues that felt all too familiar.
Newman addressed the issue in a post-release update, framing it primarily as a discovery challenge rather than taking a hard anti-AI stance. "Eventually the slop will just fall to the bottom," he suggested, though that kind of organic filtering assumes a level of algorithmic sophistication the platform clearly hadn't achieved at launch.
The official position tried to split the difference: "We're not an AI-based product... not anti-AI either." Fair enough. The platform even included a philosophical aside about "AI is the teacher now," which feels like the kind of comment that will age in unpredictable ways.
Facepunch's strategy appears focused on curation rather than prohibition. They committed to promoting human creativity on the main page through improved discovery algorithms, letting engagement metrics and community behavior create natural filters while low-effort AI content sinks. Weekly updates promised ongoing adjustments.
It's a bet that engagement data and time will solve what moderation policies can't. Whether that faith is warranted remains very much an open question.
The Roblox Problem
The challenge S&box faces isn't really about revenue models or engine licensing. It's about scale and network effects.
Roblox doesn't just pay creators over a billion dollars a year—it's built an ecosystem where kids expect to find other kids, where social dynamics drive discovery as much as algorithms do, where the platform itself becomes the destination. Epic has invested heavily in building similar momentum around Fortnite Creative.
S&box is playing a different game: one where standalone Steam releases matter more than on-platform economies, where creator freedom theoretically trumps platform lock-in. But you can't compete with Roblox by simply offering better economics if nobody's there to play the games.
The early numbers suggest Facepunch has built something that can survive—immediate profitability after a five-year development cycle is nothing to dismiss. But survival and relevance operate on different timelines.
The AI content problem, more than anything, reveals the central tension in democratizing game creation. Lower the barriers enough and you get exactly what Facepunch got: an explosion of quantity that threatens to bury quality. The Steam review scores suggest they hadn't solved that discovery problem at launch, which may be the more existential threat than any concern about revenue splits or platform royalties.
Newman has weathered community storms before. Garry's Mod endured for years in part because the community policed itself, because YouTube creators and server operators created their own curation layers. Whether S&box can replicate that kind of organic quality filtering in an era of AI-generated content and algorithmic feeds—well, that's the real test.
The million-dollar opening day bought Facepunch time and attention. What they do with both will determine whether S&box becomes a genuine alternative or just an interesting footnote in the ongoing platform wars over user-generated content.
