Cardano's Ouroboros consensus and the Chang governance upgrade (parts 1 and 2 of this series) were about making the base chain work better. Midnight is a different kind of bet: a dedicated network for computing on data you don't want to put on a public ledger, tethered to Cardano but running its own rules. It went live on mainnet on 30 March 2026. It's also, by its own node-operator structure, still not finished.
What Midnight Actually Is
Midnight is Cardano's first "partner chain" — not a competitor to Cardano, and not a new general-purpose layer-1 chasing Ethereum or Solana's use cases, but a dedicated confidential-computing network plugged into the Cardano ecosystem. IOG (Input Output Global, the engineering company behind Cardano) built it to solve one problem: letting smart contracts process private data without putting that data on a public chain.
The trick is a split. Public data — the stuff that needs to be verifiable by anyone — sits on a public UTXO (unspent transaction output, the accounting model Cardano and Bitcoin both use) ledger. Private data — account balances, counterparty identities, contract terms — is processed locally, on the user's own machine, and never leaves it. What gets submitted to the chain is a zero-knowledge proof: cryptographic evidence that a valid computation happened, without revealing the inputs that produced it.
The model underneath this is called Kachina: it proves a set of private inputs justifies a public state change, so the chain can update its records without ever seeing what caused the update. Only the proof touches the chain; everything else stays off it. IOG's own framing is "programmable privacy" — developers choose what's shielded per application, rather than the network imposing blanket anonymity on every transaction the way Monero does. The two chains are also firewalled from each other: a failure on Midnight can't endanger Cardano, and vice versa.
That distinction — optional, developer-defined privacy instead of mandatory network-wide privacy — is Midnight's whole thesis, and everything else here follows from it.
Who Is Building It
The protocol engineering is IOG's — the same company, run by Charles Hoskinson, that built Cardano. A separate Midnight Foundation, led by president Fahmi Syed, handles governance, funding, and branding, mirroring the IOG/Cardano Foundation split. Cardano stake pool operators can also run Midnight infrastructure.
Hoskinson has been explicit that Midnight isn't trying to peel users away from Monero or Zcash. He's described that audience as having "hardline expectations around privacy" — anonymity as a non-negotiable default, not a feature a developer can switch on or off. Midnight is aimed elsewhere: institutions that want a public blockchain but can't put customer balances or transaction details in the open.
NIGHT and DUST: The Dual-Token Model
Midnight runs on two tokens that do opposite jobs, not two flavors of the same thing — and this split is what actually enforces the network's thesis.
NIGHT is the network's ordinary, liquid asset. It's unshielded (visible on-chain, like any transparent token), tradeable, and stakeable, with a fixed maximum supply of 24 billion. It carries governance rights and underwrites network security. You can buy it, sell it, send it to anyone — it behaves like any other crypto asset.
DUST is the resource that pays for shielded (privacy-preserving) transactions and private contract execution on Midnight. It's deliberately crippled in one way: it cannot be transferred — not to another wallet, not to settle a debt, not sold on an exchange. It has no market, because it isn't allowed to have one.
You don't buy DUST — you generate it by holding NIGHT. A network parameter called the night_dust_ratio governs the rate: roughly 5 DUST per NIGHT at full capacity, reached after about a week of continuous holding. Spend or move the backing NIGHT, and the DUST balance decays back to zero. It has no existence independent of NIGHT you're currently holding — it isn't earned once and banked forever.
| NIGHT | DUST | |
|---|---|---|
| Transferable | Yes | No, under any circumstance |
| Tradeable on exchanges | Yes | No |
| How you get it | Buy, earn, claim | Generated passively by holding NIGHT |
| Supply | Fixed at 24 billion | Uncapped, decays without backing NIGHT |
| Purpose | Governance, staking, security | Pays for shielded transactions and contract execution |
| Can it move value between people | Yes | No |
Why build it this way? Because the token that pays for privacy can never itself become an anonymous payment rail. If DUST were transferable, you could sell someone "privacy fuel" like any other asset, and Midnight would effectively be a shielded value-transfer network with extra steps — closer to the networks it's explicitly trying not to be. Non-transferable, self-decaying DUST ties the right to transact privately to holding NIGHT, not to a market for anonymity itself.
There's a second, practical benefit: it decouples usage cost from NIGHT's market price. An enterprise running payroll on Midnight doesn't want its fees to double because a token pumped on an exchange. DUST generation is set by protocol parameters, not supply and demand, so fees stay predictable regardless of what NIGHT is trading at.
The Glacier Drop
Midnight's initial NIGHT distribution — the Glacier Drop — was unusually broad. The snapshot covered eight separate ecosystems: Cardano, Bitcoin, XRPL, Ethereum, Solana, BNB Chain, Avalanche, and Basic Attention Token. Nearly 34 million wallet addresses were eligible.
The claim window opened 5 August 2025, originally for 60 days, and was extended by 16 days to close on 20 October 2025. Uptake was fast: over 1 billion NIGHT — more than 4% of total supply — was claimed in the first week alone. By the time the window closed, more than 3.5 billion NIGHT had been claimed by upwards of 170,000 wallets.
Tokens nobody claimed didn't disappear. They rolled into a 30-day "Scavenger Mine," a proof-of-work-style redistribution round, followed by a four-year "Lost-and-Found" window for late claimants. Supply breakdown: Glacier Drop 3.547 billion, Scavenger Mine 1 billion, Lost-and-Found 252 million, Midnight Foundation 8.4 billion, TGE (token generation event) 3.66 billion, Reserve 6 billion, Treasury 1.141 billion.
Claiming NIGHT didn't mean getting it immediately, either. Claimed tokens are subject to a 360-day thaw starting 10 December 2025, released in four equal 25% installments, 90 days apart, with each address's start date randomized so unlocks don't all land on the same day and hit the market as one supply cliff.
Mainnet Status: Live, But Federated
Midnight's mainnet genesis block landed on 30 March 2026, opening what the project calls the "Kukolu" phase. Be clear about what that means: this is not full decentralization. Block production currently runs through nine named federated node operators — Google Cloud, Worldpay, Bullish, MoneyGram, Pairpoint (Vodafone), eToro, AlphaTON Capital, Blockdaemon, and Shielded Technologies. That's a list of large, mostly regulated institutions, not a permissionless set of validators anyone can join.
The chain has run steadily for several months. What hasn't happened yet is permissionless block production — letting the wider community, not nine named operators, actually produce blocks. That's the hard test still ahead, with no public date for when it clears.
One quirk worth flagging: NIGHT the token predates Midnight the chain. NIGHT launched as a Cardano native asset in December 2025, months before Midnight's own mainnet went live in March 2026 — worth knowing if you're reasoning about "when did Midnight launch" from token price history alone.
As of mid-July 2026, NIGHT trades around $0.03, with roughly 16.6 billion circulating and a market cap in the $490-500 million range.
Compact and the Developer Stack
Midnight's smart-contract language is Compact — a TypeScript-based DSL (domain-specific language, built for one job rather than general-purpose programming) that compiles application logic directly into zero-knowledge circuits. The point is to hide the cryptography: you write contract logic in something that looks like familiar code, and the compiler handles witness generation and proof construction behind the scenes.
As of the February 2026 network update, the stack stood at Compact compiler v0.28.0, midnight-js 3.0.0, wallet-sdk 1.0.0, and Proof Server 7.0.0 (the component that generates proofs for the client). It's a maturing toolchain, not a finished one — several core pieces are still versioned below 1.0.
What It's Actually For
The clearest live example is Monument Bank, a Bank of England-regulated UK digital bank with roughly £7 billion in deposits and more than 100,000 customers. Monument is tokenizing up to £250 million of retail deposits on Midnight, reportedly the first deployment of its kind by a UK bank on a public blockchain. The pitch: represent regulated bank liabilities on a public network while keeping account balances and transaction details shielded from everyone except the bank and its customers.
That's the shape of use case Midnight is chasing — not anonymous peer-to-peer cash, but compliance-heavy applications wanting a public chain's auditability without broadcasting customer data. Other plausible categories: private payroll and B2B settlement, KYC/AML-compliant DeFi, cross-border payments (MoneyGram, notably, is one of the nine node operators), and real-world asset tokenization needing an audit trail without full public exposure of underlying positions.
Beyond Monument Bank, there isn't yet a large-scale, live consumer application to point to — early infrastructure with one credible flagship deployment, not proven demand across many use cases.
The Honest Risk Section
Midnight invites comparison to every other privacy project in crypto, and the comparisons cut both ways.
Against Monero and Zcash: Monero shields every transaction by default, with no transparent mode at all — a design that's structurally hard to reconcile with exchange KYC (know-your-customer) checks or the FATF Travel Rule, the international requirement that transfers above a threshold carry sender and receiver identity information. Zcash already supports selective disclosure through viewing keys, and roughly 31% of its supply sits in shielded pools as of early 2026, so selective disclosure isn't an idea Midnight invented. What's different, Midnight argues, is that selective disclosure is the default assumption for every contract, not an opt-in layer bolted onto a transparent base chain. That's a real architectural difference. It is not, by itself, proof regulators will treat it differently.
Against Aztec, the closest thing to a direct competitor: both chase privacy for programmable applications, but Aztec is an Ethereum layer-2 rollup (a scaling chain that settles back to Ethereum) using a language called Noir, while Midnight is its own partner chain using Compact. Different ecosystems, different trust and settlement assumptions.
The bigger risk is regulatory, and it isn't hypothetical. The EU's Anti-Money Laundering Regulation bars EU-licensed exchanges and custodians from listing, holding, or servicing "privacy-preserving" crypto-assets starting 10 July 2027 — under a year from now. Self-custodied ownership stays legal, but that's cold comfort for a project whose institutional pitch depends on regulated banks and payment companies being comfortable touching it. It's genuinely not established that regulators will look at Midnight's selective-disclosure design and conclude it's meaningfully different from Zcash or Monero for classification purposes. Midnight's whole technical thesis is that the distinction matters. The law that takes effect in under a year may simply not draw the line where Midnight needs it drawn.
There's also a trust trade-off worth naming plainly: selective disclosure means someone — the application, the issuer, a regulator-facing key holder — decides what gets revealed and to whom. That's a fundamentally different property from Monero-style privacy, where no party can be compelled to disclose anything at the protocol level. Programmable privacy isn't private-by-default with an escape hatch; it's disclosed-by-someone's-decision, with shielding as the starting state. Depending on who you are, that's either the whole point — an institution wants exactly this kind of controlled disclosure — or the reason you'd never treat it as a substitute for actual anonymity. It's a real trade-off, and Midnight doesn't get the benefits of both models at once.
Put together: this is a serious technical wager, backed by real engineering and one credible institutional deployment, running on infrastructure that's still federated rather than decentralized, walking into a regulatory environment that may not care about the distinction its entire design is built on. How that regulatory line gets drawn is arguably a bigger risk to Midnight's institutional pitch than any technical hurdle left in front of it.
Part 4 of this series covers Hydra and Mithril, Cardano's answer to the scaling problem — a different bet, on a different layer of the same stack.
