The Friction That Created Two Models
India's gaming split isn't philosophical. It's mechanical.
A user in Bengaluru downloading a AAA console port faces three gates. Download size kills data budgets. Payment requires credit card or NetBanking friction. Server location means sub-100ms latency costs infrastructure. Each gate is a 30-50% dropout.
The same user playing Ludo on MPL faces zero gates. UPI is one tap. Download is 50MB. Server latency matters less. Conversion funnel stays open.
Native Stack Architecture
Stack-native gaming means: multiplayer, social, skill-based, low-data, UPI-first.
MPL's model: 50 rupees entry, instant UPI settlement, no payment gatekeeping. Scaled to 2.5M+ DAU in 36 months. Winzo followed similar rails.
Both work because they're not fighting the stack. They're using it.
UPI reduces payment failure from 8-12% to 0.3-0.8%. That delta alone swings unit economics. At 50-rupee entry point with 40% rake, one payment failure kills margin. UPI changes the math entirely.
Vernacular is easier here too. Ludo, Rummy, Carrom don't need English text. MPL supports 9 languages. Localization cost drops. Retention lifts because grandmother can play.
Think of it as using existing infrastructure you don't own. Like e-commerce using ONDC rails instead of building logistics.
Wrapper Economics
Wrapper games mean: ported AAA titles, console-style mechanics, premium pricing, global studios.
Gautam Adani's Nodwin Gaming invested in premium titles. EA and Activision ported games to India. They all faced the same wall: users here don't pay 500+ rupees upfront.
India's ARPU for premium games is $0.80-$1.20 annually. North America's is $12-18. The 10x gap isn't snobbery. It's payment infrastructure.
Wrappers work in three scenarios: (1) you subsidize from elsewhere, (2) you build IP moats (Nintendo effect), (3) you get VC to burn cash.
Beyond Play raised $12M for console-style games in India. That capital is working backward. Building audience before monetization. Only works if VC runway extends far enough.
Compare: MPL's model generates cash from month 2. Wrapper model generates losses until year 3-4.
The Margin Inversion
Here's the structural gap most miss.
MPL's 40% rake on UPI (1% commission) nets ~39% to platform. Payment risk is near zero. Chargeback rate is <0.1%. Customer acquisition cost is 30 rupees per user in Year 1 because viral loop works.
A wrapper's 30% margin on premium game sale (after platform, payment, tax): 26% nets to studio. But chargeback rate is 4-6%. CAC is 200+ rupees because organic adoption is thin.
Net: Stack-native margins are 200-300% higher before scaling efforts.
Timing and Saturation
This gap is compressing, but not closing.
Winzo, MPL, India Games are now saturating skill-gaming TAM. Prize pools ballooned. User acquisition costs are creeping toward 100+ rupees. The easy margin days ended in 2021.
Meanwhile, broadband speeds (average 18 Mbps) and device quality (OnePlus, Pixel under 30K) are making wrapper games viable. Royale Chaos (Solitaire Royale) hit 5M downloads. That wasn't possible in 2020.
But wrappers still need capital. They're slower. They're less efficient.
The Non-Obvious Part
The real advantage of wrappers isn't games. It's user data.
MPL knows who plays Rummy at 3 AM. Wrapper studios will eventually own which 25-year-old males prefer turn-based over real-time. That data feeds into advertising, credit, loyalty. The moat isn't the game. It's the behavior profile.
MPL is already moving here (credit, fantasy). Wrappers are 18 months behind in seeing it.
Implication
If you're founding in gaming, pick one:
Stack-native if you want to scale fast and cash-positive by Year 1. You'll fight saturation early but you'll have capital.
Wrapper if you're willing to burn VC for 24-36 months and you have IP or technological moat. You'll build slower but defensible.
Hybrid (skill-based + premium cosmetics) is the current gold zone. But it requires both capital and UPI ruthlessness. Few teams have both.
The game isn't which model wins. It's that both models exist because infrastructure constraints are that different in India. Build for constraints first. Scale second.