Here’s a contradiction worth sitting with. Earlier this year, Epic Games — the studio that spent five years and north of $100 million breaking the platforms’ 30% tax — quietly cut the value of V-Bucks by 20% to, in its own words, “pay the bills.” The company that won the fee war still had to raise prices on its own players, amid layoffs, in the same news cycle.
The reflex was to call it a betrayal. I think that misses the point entirely. Epic’s fight was real, and every developer benefits from how it ended: Google settled at 20%, Apple cut fees in China, and link-outs are finally on the table. The more useful reaction isn’t outrage — it’s the question almost nobody asked. If the biggest winner in the West, with its own store and its own engine, still can’t escape the squeeze, what chance does anyone else have?
The V-Bucks cut was never an Epic story. It was a margin story, and the entire industry is living in it.
This piece is about where margin actually comes from once content is infinite and attention is finite — because it isn’t the fee percentage, and it isn’t even distribution the way most studios picture it. It’s the player relationship. And a sprawling, nearly invisible economy of community-run game servers has spent fifteen years quietly proving exactly what that relationship is worth in dollars. To get a real number, I sat down with Liam Wiltshire, GM of Tebex — the company sitting on the most underreported figure in our business.
The math is brutal, and it’s first-principles simple
On one side of the equation: player attention is finite, and the competition for it isn’t just other games anymore. It’s TikTok, YouTube, Netflix — everything with a screen. I’ve written about attention compression before: players now evaluate a game in roughly 16 seconds.
On the other side: supply is exploding. iOS App Store submissions jumped over 60% year-over-year this past December, after three years of basically flat growth. Why? AI is collapsing the cost of building software toward zero. People who couldn’t ship apps are shipping apps, and games are next.
Infinite shelf. Finite attention. When supply explodes and attention doesn’t, two things happen mechanically: the cost of acquiring a player goes up, and your pricing power goes down. That’s margin compression. That’s the V-Bucks story, the layoff story, the whole story of the industry right now.
So ask the first-principles question: in that world, what’s actually scarce?
Not content — AI just made content abundant. The obvious next answer is distribution. But getting on the shelf isn’t distribution. Steam will list you. The App Store will host you. Getting a player to walk down your aisle is distribution, and that’s the most expensive it has ever been.
Here’s the move most people miss: a player relationship is distribution. A player who knows you, trusts you, and comes back without being paid to come back — that’s not a brand metric. That’s a distribution channel you own instead of rent. Every other channel — paid UA, store featuring, the algorithm — you rent by the impression, and the rent is going up. The player relationship is the only channel whose cost goes down the stronger it gets.
And here’s how you know that’s true: watch what the platforms refuse to give up. They’ll negotiate the percentage. Google lowered its headline fee — then introduced a charge for steering and quietly moved chargeback costs onto developers. Valve is in federal court right now facing a certified class action of some 32,000 developers over rules that stop you from selling your own game cheaper elsewhere; their position is that no such written policy exists, while the emails in evidence suggest otherwise.
The platforms will give you margin before they give you the relationship. The login, the payment method, the purchase history — that’s the part they keep, because they’ve understood for fifteen years what most studios are only starting to see.
The relationship is the moat.
Which is exactly what I wanted to put a number on. Tebex sits on the data to do it.
🎧 Listen on Spotify, Apple Podcasts, or Anchor
The shadow games industry
While everyone argued about percentages, somebody already solved this — not in theory, but in production, at scale, for fifteen years. Just not anywhere the industry press ever looks.
There’s a Minecraft server called Hypixel that holds the Guinness World Record for concurrent players and is closing in on year 15. Microsoft doesn’t run it; a community does. There are FiveM role-play servers — the multiplayer layer on top of GTA V — where players have held the same job, in the same city, with the same friends, for years. Thai communities. Portuguese economy servers. German servers. No publisher pays for the hardware. No platform processes the payments. The people who run these worlds own the entire player relationship, and players pay them directly, happily.
In a brutal attention economy, these communities have held people for a decade-plus.
And the company that built the payment rails underneath this world — a company most people in this industry couldn’t have named six months ago — has processed $1.5 billion in lifetime payments. These days that includes direct-to-consumer for studios you’ve heard of: Rockstar, Take-Two, Hytale, FiveM.
So what’s the whole UGC server economy worth? Honest answer: nobody knows. No earnings call reports it. No analyst covers it. The best window anyone has is one company’s books — and through that one window, a billion and a half dollars. The market is big enough that the number should exist, and it doesn’t. That’s how deep in the shadows this thing runs.
That company is Tebex. It was started by a 16-year-old named Lee McNeil in his childhood bedroom in Nottingham, England, with £20 — the price of a domain name. (More on that story in a minute, because it’s worth it.)
What Minecraft kids know that billion-dollar studios don’t
Why should you care what Minecraft kids and GTA roleplayers are doing? Because their data tells you what a real player relationship is worth in dollars. Three numbers from Liam stopped me cold.
1. Trust is worth 50–227% more spend. When a purchase goes through a creator code — a streamer the player actually trusts — average spend jumps from $24 to $41. In Rust, from $18 to $60, a 227% lift. And before you call it a promo effect: 84% of players only ever use one code. One. That’s not a coupon. That’s loyalty.
2. How players pay reveals how much they’ll spend. The same player, buying the same content, spends 50–70% more with buy-now-pay-later than with a card. Tebex controlled for this with cohort analysis — same users, card vs. BNPL — and the lift held. Crypto showed the same pattern, running as high as ~150% more in parts of LATAM. Players are budgeting entertainment like an investment, spreading the cost of a game across the months they expect to play it.
3. Players are fixing a checkout nobody designed for them. On desktop games — Minecraft, Rust, Ark — 35% of purchases happen on a phone. Players keep one screen in the game and buy on the other, optimizing their own flow because no one built one for them.
Put it together and the conclusion writes itself: in a world of infinite entertainment, players pay more — meaningfully, measurably more — when they trust who they’re paying. The relationship isn’t a soft metric. It’s the margin you’ve been looking for.
“Every purchase a gamer makes these days is a CapEx decision. They’re choosing to invest their time and their money, and they want to know they’re going to get a return — enjoyment, entertainment — that justifies it.” — Liam Wiltshire
“Is the 30% dead?” Yes — with a giant asterisk
Liam’s read on the fee wars is more sober than the headlines. Yes, studios now have alternatives. But the platforms aren’t giving up margin; they’re relocating it.
Google drops the headline fee, then adds a steering fee and pushes chargeback risk onto developers. Steam may permit link-outs — but the same anti-steering mechanics that govern mobile are now playing out in PC. “Is 30% dead?” Liam’s answer: “Yes, to a point. But no, in other ways.”
The smarter framing he offered is to treat the platform as an acquisition channel, not a tax. Pay the steering fee on the first purchase and book it as cost of acquisition. Then — exactly like the little “come to our store next time” card eBay sellers tuck into the box — convert that buyer to your own webshop, where you keep 90%+ instead of 70%.
“I don’t mind if that first purchase is only $5. If I can get them to come back and make twenty $10 purchases over the next year that go straight through my webshop, that initial acquisition cost suddenly becomes the best business decision you ever made.”
The mindset shift he keeps fighting: studios still ask “How do I maximize the one-time spend?” The whole model breaks unless you’re optimizing for the relationship, not the transaction.
What a “merchant of record” actually does
A lot of payments companies throw the term around, so Liam defined it plainly. As merchant of record, Tebex takes full selling responsibility: it buys the content from the studio and resells it to the player under the studio’s own branding. The studio still owns the relationship — but Tebex becomes the principal in the transaction. Which means Tebex eats the parts of the business that quietly kill small teams:
Sales tax across 120+ jurisdictions (one sale into the wrong U.S. state can trigger an audit).
Chargebacks and fraud. If a $5 sale gets charged back and Tebex loses $20 fighting it, the studio still keeps its $5. “You will never pay a cent.”
First-level billing support, so a server owner doesn’t have to staff a support desk.
Global payment methods most studios have never heard of — iDEAL (70% of Dutch purchases), Pix in Brazil (which erased the cash-over-the-counter boleto in five flat years), RuPay in India. Throw friction at a German player’s checkout and you lose them; the fix is local.
The throughline: free the people who are brilliant at building worlds from the “crappy business stuff” they’re not experts in. As Liam puts it, the goal is to spend as little time paying and as much time playing as possible.
The Hytale proof point
If you want a single case study for everything above, it’s Hytale.
Hytale launched off Steam entirely — web-only — and with 0% server commission for two years. No library management, no one-click Steam convenience, none of it. By every conventional rule, that should have failed.
Instead, pre-orders alone secured two years of runway for the studio before launch sales even started. Players were willing to sacrifice convenience because they were fans, not just users. Hytale leaned into that with creator codes (20% of pre-launch purchases came through a code — three times Tebex’s platform average) and with BNPL and crypto, which let supporters back the most expensive “Cursebreaker” tier by spreading the cost over the months they knew they’d be playing.
There’s a poetry to it: Hytale was born out of Hypixel Studios, which was born out of running the Hypixel Minecraft server. The shadow economy didn’t just inform the playbook — it funded the game.
The £20 origin story
I’ll close where Liam did, because the founding story is the thesis in miniature.
Lee McNeil started what became Tebex (then “Buycraft”) at 16, in his bedroom, with the £20 he needed to buy a domain. The early Minecraft economy was a Wild West: you’d send money to a random PayPal address and maybe your in-game reward showed up two days later, maybe you’d already filed a dispute. Lee built the layer that fulfilled the purchase instantly. Every time someone subscribed, he ran downstairs to tell his parents. His dad: “Tomorrow it’ll be even more, and one day you’ll be a millionaire.”
For the first five or six years it was a flat $9.99-a-month subscription — and that was the ceiling. The inflection point wasn’t a new market; it was the team finally recognizing its own value. They were doubling some servers’ revenue through local payments and absorbing all the chargeback risk for free. When they launched the merchant-of-record product (with FiveM as the first partner) and started taking a cut of each transaction, the business changed shape.
Tebex was acquired by Overwolf in a reported ~$29 million deal — with a headcount of eight, having never taken outside investment or debt. It’s now closing in on 60 people. Liam, who took a pay cut to join two kids at a kitchen table because he believed UGC and modding were what make games live for decades, says he still wakes up excited.
The whole arc — bedroom project to $1.5 billion in rails — runs on one bet: that the player relationship is the most valuable asset in games. The platforms have known it for fifteen years. The Minecraft kids figured it out by accident. The only people still catching up are the studios.
The industry spent fifteen years engineering pressure. The data says we should have been engineering trust.
Liam Wiltshire is GM of Tebex, the merchant-of-record and D2C platform behind direct payments for Rockstar, Take-Two, Hytale, and FiveM. Find him on LinkedIn.
If this was useful, the best thing you can do is share it with one person who’s still arguing about percentages. And if you’re not subscribed yet — join us at gamemakers.com.
