WTF is ZK (Zero-knowledge) and why is everyone talking about it now?
In the simplest of terms, Zero-knowledge (ZK) = Cryptographic proof that something is true with zero leakage of why it's true.
In the simplest of terms, Zero-knowledge (ZK) = Cryptographic proof that something is true with zero leakage of why it's true.
In blockchains = privacy + scale, without trusting anyone. You have private data (e.g., your bank balance, a password, or transaction history).
- You generate a tiny cryptographic "proof" that says: "This statement is true." - John has 4.2 Bitcoin or is over 21 yrs old, etc
- Anyone can verify that proof in seconds.
- But zero underlying data leaks—not even a hint.
ZK tech is in the spotlight right now with @Zcash having a big run, @Starknet advancing, @NEARProtocol intents, @0xMert_ influencing teams on Solana like @vanishTrade, EVM privacy like @RAILGUN_Project and @Lighter_xyz proved ZK in action on the L2 for perps. Even @VitalikButerin and @drakefjustin are floating EVM going to a zkEVM...privacy is a focus again.
Why does it matter?
Simply - ZK allows provable ways to verify information and maintain privacy.
Before it gets too confusing, I'll hit on what ZK is, a ZK-SNARK and a ZK-STARK and try to give some real life uses to make it more digestible. It's tough because it can all seem like a jumble of acronyms...which it is.
Why now?
2025 transformed zk-SNARKs from research curiosities to infrastructure with smaller/faster proofs, quantum prep, and $300M+ DeFi impact via SoKs (Systematization of Knowledge), dynamic updates, and L2 integrations. Privacy is back in style.
How does Zero-knowledge work in blockchains?
Privacy Send $100 privately -> prove it happened -> hide sender, receiver, amount (e.g., @Zcash @RAILGUN_Project)
Scalability Batch 10,000 transactions off-chain -> submit one ZK proof to the blockchain(ex: L2 submitting to the ETH L1) -> verified in ~100ms (@zksync, @Starknet or a perps chain like @Lighter_xyz)
Trustless Identity Prove "I'm a U.S. citizen over 21" -> without showing ID or birthdate (@ZKPassport)
That's the TLDR. Now time to dig a level deeper.
Snarks vs Starks
What is a ZK-SNARK?
ZK-SNARK = Zero-Knowledge Succinct Non-Interactive Argument of Knowledge
You Lock Your Secret in a Math Box
- Your birthdate -> turned into a secret equation.
- Only you know the "key" to solve it.
Create a Tiny "Proof Ticket"
- Using a public template (the "trusted setup"), you generate a 200-byte proof.
- It says: "I solved the age equation correctly."
Hand Over Just the Ticket
- Bartender gets only the proof — no birthdate, no name.
Bartender Scans It in 1ms
- "Valid!" -> You're in.
- Zero info leaked
Why snarks?
- Tiny proof -> fits in one Ethereum block.
- Instant finality -> no 7-day fraud window.
- No interaction -> works in smart contracts.
What are some real life applications of a SNARK?
- Fully Private Crypto Payments (@Zcash $ZEC) - "I sent X amount from A to B" -> hides sender, receiver, amount
- Scalable Ethereum L2 (ZK-Rollups) - "10,000 txs were executed correctly" -> 1 SNARK (~288 bytes)
- Private DeFi Trading - "I swapped 100 USDC -> DAI at fair price" -> hides amounts & wallet
- Anonymous Identity & Credentials - (Worldcoin) "I'm over 18 / unique human" -> no name, no ID
- Compliance Without Exposure - (Railgun/Tornado Cash) - "Funds are not from blacklisted addresses" -> hides full history
- Enterprise & Supply Chain - "Invoice paid, goods shipped" -> hides pricing & partners
- Voting - "I voted once, valid option" -> hides vote choice
What is a ZK-STARK?
ZK-STARK = Zero-Knowledge Scalable Transparent Argument of Knowledge
It's the "no-trust, future-proof" upgrade to ZK-SNARKs.
According to the Starknet FAQs Page:
STARKs are defined as Validity proofs, meaning they prove computational integrity. With Validity proofs (STARKs included), one party can prove to another that a certain computation was performed correctly. Zero-Knowledge is an additional property asserting that the prover does not have to reveal any information included in the computation.
Most ZK-Rollups are not really Zero-Knowledge. The mistake comes from the confusion between Validity proofs (assures computational integrity) and Zero-Knowledge proofs (assures no information is revealed).
Starknet uses STARK Validity proofs. It is worth noting that extracting information from the proof is extremely hard — almost impossible in practice.
ZK-STARK = ZK proof with no trusted setup, quantum-safe, and scalable — but proofs are bigger.
SNARK = smaller, faster, needs a ceremony.
STARK = safer, transparent, future-proof.
What are some real life applications of a STARK?
- Starknet (Ethereum L2): Full app logic off-chain -> 1 STARK proof -> cheap & secure
- Immutable X (NFTs/gaming): Prove game moves -> no cheating
- Future DeFi, DAOs or possibly Government: Prove complex rules (e.g., voting, lending) -> fully private & trustless
One more...What is a Validity Proof?
A validity proof is a cryptographic certificate that says: "This batch of work (e.g., 10,000 transactions) was done 100% correctly — according to the rules."
In simple terms, don't trust me, trust/verify the proof. This has been implemented on the L2 level with Ethereum because it solves some L1 issues.
On Ethereum:
- Layer 1 is secure but slow & expensive (~15 TPS, $1-$10 per tx).
- Layer 2 (Rollups) process thousands of txs off-chain -> bundle them -> post one proof to L1.
But how does Ethereum trust the L2 didn't cheat?
A Validity Proof = the answer.
Validity Proof vs. Fraud Proof
| | Validity Proof | Fraud Proof | |---|---|---| | How it works | Proves correctness upfront | Assumes correct, challenged later | | Finality | Instant | 7-day dispute window | | Used by | ZK-Rollups (zkSync, Starknet) | Optimistic Rollups (Arbitrum, Optimism) |
