We tried. In 2025 we deployed full on-chain settlement on Polygon, then Sei. Thousands of rounds on mainnet. Hit three structural limits:

• 1. Commit-reveal needed two transactions per round.
• 2. Gas and nonce management required a dedicated backend.
• 3. Selective reveal was a real attack vector.

Randomness has to live inside consensus, not in an oracle callback. Solvency has to be a protocol rule, not a smart contract that can be paused or upgraded.

That's why it's a sovereign Cosmos SDK chain — we control the state machine.

Exohash is a sovereign chain — it has its own validator set, independent of Cosmos Hub. Validators run our custom binary (exohashd), not the Cosmos Hub binary.

Think of it like Osmosis or dYdX — built with Cosmos SDK, but a fully independent network with its own consensus, its own validators, and its own governance.

The Cosmos SDK is the framework (like Rails or Django). The chain is the application. We use CometBFT for consensus and IBC for cross-chain communication, but we are not part of the Cosmos Hub validator set.

Genesis validator set: 32 validators at launch, scaling to independent operators as the network matures. Validators are incentivized with a share of protocol revenue, paid in USDC.

Validators have equal voting power — one validator, one vote. Bond size doesn't matter. The BLS threshold requires 2/3 of validators to produce valid randomness. With 32 validators, an attacker needs to compromise 22 of them.

At genesis, validators are invited by the foundation — similar to how Noble (Cosmos USDC issuer) launched with a permissioned proof-of-authority set. As the network grows, new validators join through governance — existing validators vote to admit new ones.

This is a fundamentally different model from stake-weighted systems where a single entity with enough capital can dominate the validator set.

Validators run a Distributed Key Generation (DKG) protocol — they collectively create a shared BLS key. No single validator holds the full key.

Every block, each validator signs the block height with their key share. The block proposer combines any 2/3 of these shares using Lagrange interpolation into one aggregate signature. That signature is verified against the group public key.

Randomness = SHA256(aggregate signature). 32 bytes. Deterministic, unpredictable, bias-resistant, and verifiable on-chain.

Partial signatures are propagated via ABCI++ vote extensions and aggregated by the block proposer before finalization.

Same cryptography as drand — used by Cloudflare and Protocol Labs. The difference: drand is an external service. Ours is embedded in consensus. No oracle call. No latency. No external oracle trust assumption.

The protocol ships with the most popular casino game formats: crash, dice, mines, plinko, limbo, keno, plus a slot engine framework where each slot game is a config file.

The engines are native chain modules, not smart contracts. Deterministic, open-source, non-upgradeable without governance.

Three engine types cover virtually all formats a crypto casino needs:

• Arena — multiplayer rounds (crash, roulette, baccarat)
• Instant — single bet, single outcome (dice, slots, keno, plinko)
• Step — sequential decisions with cashout (mines, blackjack, tower)

Protocol captures a configurable share of the house edge on every bet settled on-chain (initially 25%, adjustable via governance). Remaining 75% goes to the bankroll owner — either an operator running a private bankroll, or distributed to LPs in a public bankroll.

A portion of protocol revenue goes to validators securing consensus and generating randomness.

All fees denominated and settled in USDC. Zero inflationary token emissions. EXOH holders earn real yield from real wagering activity, not from printing tokens.

Revenue is purely volume-driven with zero exposure to game outcome volatility.

The house takes game-outcome risk — if players win more than they lose, the bankroll shrinks. That's normal and expected. Over volume, the house edge ensures the bankroll grows.

What the protocol guarantees: every bet is fully collateralized before it settles. Before a bet enters the ledger, the protocol computes the worst-case payout and reserves that amount from the bankroll escrow. If the bankroll can't cover it, the bet is rejected.

Bankroll funds are held in on-chain escrow — not in an operator's wallet, not in a multisig, not in a database. Payouts are capped to the reserved amount. The bankroll balance can never go negative.

No undercollateralized bets. No IOUs. No "pending balance." Every payout is backed by locked funds before the bet even starts.

Exohash provides infrastructure — randomness and settlement. Operators handle their own licensing, KYC, and compliance.

The iGaming industry has two layers of trust:

Technical certification — GLI-19 proves the RNG meets mathematical and security standards. Recognized in 710 jurisdictions. Zero crypto projects have it today.

Business licensing — Malta's B2B Critical Gaming Supply license authorizes a company to sell gaming infrastructure to operators. 170+ companies hold this license. Zero are blockchain protocols.

Whether a blockchain protocol needs either is an open question. A Nevada court just ruled that state gambling laws apply to Polymarket despite being CFTC-regulated and running on smart contracts. AWS hosts FanDuel without a gaming license. The line between infrastructure and operation is still being drawn.

If ExoHash needs them, we'll get them. The architecture is designed to pass both.

When an operator applies for a Malta license and the regulator asks "how is your RNG certified?" — they point to our GLI certificate. First bridge between "provably fair" and "regulator-approved."

The iGaming infrastructure market has two layers: operators and the platforms they run on.

Operators — Stake.com, BC.Game, Rollbit — serve players. They buy infrastructure from platform providers.

Platform providers — SoftSwiss, EveryMatrix, BetConstruct — provide the backend: RNG, settlement, game engines. All server-based, audited periodically, trust-based in between.

The one on-chain attempt — Luck.io on Solana — uses 4 team-controlled VRF keys with bets happening off-chain. It's a server with blockchain receipts.

No one has built real iGaming infrastructure on-chain. ExoHash is the first.

Why we win: verifiability. Every player can verify their bet outcome on-chain — no trust required, no disputes, no "contact support." The proof is in the block. Once players experience that, they don't go back to trusting a server.

The industry argues about who owns the server. We removed the server.

EXOH has zero inflation. No emissions. No staking rewards paid in newly minted tokens.

Protocol revenue is distributed in USDC to stakers. EXOH also governs protocol parameters — fee splits, game configurations, and upgrades. The token value comes from real yield and governance rights, not speculation.

Listing is a post-mainnet conversation. We want the token to have real utility and real revenue behind it before it trades.

USDC via IBC. Native Cosmos interoperability. Deposit from any connected chain, one transfer.

No card processors, no chargebacks, no PSP stack. Native crypto payments from day one.

Solo founder. PhD in Nuclear Physics from TU Delft. Started in high-frequency trading in 2007 — statistical arbitrage, market making.

Then institutional risk management — Partners Group at $1.5B AUM, Assura at $3.2B AUM under FINMA regulation.

Since 2021, full-time blockchain — MEV strategies, DEX arbitrage, execution-layer protocols. Built and deployed on Polygon and Sei before building Exohash from scratch.

Every role was about the same thing: systems that handle money at speed under adversarial conditions. HFT taught me execution. Risk management taught me solvency. Both taught me that when real money moves in real time, you can't rely on audits — you need the math to enforce the rules. That's what ExoHash does.

Post-funding: 2 frontend engineers, iGaming consultant, community manager.

Three parallel workstreams from day one:

Protocol & Engines (Founder)

• Month 1-2: Remaining game engines, protocol hardening
• Month 3-5: Security hardening, stress testing, guarded mainnet launch

Reference UI & Wallet SDK (Frontend hires)

• Month 1: Scaffold reference UI, embedded browser wallet (no extensions, no popups)
• Month 2-3: Game UIs, bankroll dashboard, USDC deposits via IBC
• Month 4-5: Polish, player onboarding flow, LP deposit flow

Ecosystem & Community (Community hire)

• Month 1: Discord, Telegram, X — community setup and engagement
• Month 2-3: Testnet launch, player engagement, validator infrastructure setup
• Month 4-5: Mainnet campaign, LP bankroll onboarding, testnet → mainnet migration

All three tracks run in parallel with minimal dependencies. The protocol is already built and running — this isn't starting from zero.

Mainnet launches in guarded mode: capped wagering limits, whitelisted LPs and players, progressive removal of guardrails as the protocol proves itself in production.

Swiss foundation in Zug. Founder is already based there. The foundation stewards the protocol, manages grants, funds GLI certification, and coordinates validator onboarding.

Operators are separate entities with their own licenses.

Standard model — same as Ethereum Foundation, Interchain Foundation, Polkadot Web3 Foundation. All in Zug.