Cash has a location
Hand someone a banknote and you no longer hold it. One note cannot sit in two pockets at once, so scarcity comes free with the paper.
Start with the problem
Every crypto design answers one question: what breaks when you take away the single trusted operator?
A bank holds your balance, checks each change, and cleans up the mess when something goes wrong. Take the bank out and that work does not vanish. Keys, agreement rules, rewards, and open markets are what grows back in its place, and each one arrives with a bill.
The starting problem
Why owning something digital is hard
Files copy for nothing
Copy a file, edit a row in a database, screenshot a picture. Nothing stops you. Scarcity does not survive the jump to bits, so digital money has to build it back out of rules.
Most systems solve that by naming an operator. A bank, an exchange, or a payment app runs the database and decides which changes count. One authority fixing mistakes, blocking fraud, reversing payments, and pushing upgrades is the cheaper design, right up to the day you need something the operator will not give you.
| Who runs the record | What it gives you | What it can cost |
|---|---|---|
| Bank or payment app | Customer support, reversals, fraud handling, legal recourse. | Account freezes, censorship, operating hours, jurisdiction limits, counterparty risk. |
| Exchange or platform | Fast internal balances and an easy sign-up. | Custody risk, hidden liabilities, withdrawal risk, policy changes. |
| Public crypto network | Rules that any participant can check alone. | More moving parts, mistakes nobody can undo, public data, slower coordination. |
Summary: coins never move anywhere. A valid transaction changes an accepted record, and that record is the whole asset.
Proof without a login desk
Keys take over from the account
No operator means no password reset and no support desk, so math stands in for both. A transaction is a signed instruction that a stranger's node can check alone: spend this value, create that value, pay this fee. A block is a batch of those instructions tied to the history behind it, and a node takes the whole batch or none of it. Without keys, the network falls back on a central login desk or an ID registry to decide who may move a balance.
Private key
The secret behind spending authority. Steal one and the network cannot tell the thief from the owner, because there is no ID check behind the math.
Public key or address
Anyone can confirm that a signature is valid without ever seeing the secret that made it.
Signature
Proof that the key holder approved this exact transaction. It does the job a username, a password, and an account admin would do on a normal platform.
Hash
A fingerprint for data. Change one byte of the input and the fingerprint changes completely, which is how each block commits to the history behind it and how tampering becomes visible.
Agreement
A record only counts if strangers agree on it
Show Bob one ledger and Carol another and the same coin has just been spent twice. Thousands of independent computers see network messages land in different orders, and they still have to end up on one shared history. Consensus is the set of rules that gets them there. What those rules have to stop is a short list, and every design choice below traces back to it.
| Attack | Plain meaning | Why users care |
|---|---|---|
| Theft | Spend value without the rightful key holder's consent. | Ownership is meaningless if signatures can be forged or keys are stolen. |
| Double-spend | Try to spend the same value in two conflicting ways. | Recipients need to know which payment counts. |
| Rewrite | Replace accepted history with a different history. | Settlement needs confidence that old payments stay settled. |
| Censorship | Keep valid transactions out of history. | Open access fails if one control point can cheaply block users. |
| Hidden inflation | Create value outside the rules. | Supply rules work only if users can reject bad supply themselves. |
Mining, staking, tokens
Agreement has to cost something
Proof of Work
Fake history gets priced out with a cost from outside the system: hardware, power, time. Miners propose blocks and get paid only if that work sticks in accepted history. The bill is energy use, ASICs, mining pools, and steady pressure toward industrial scale, and Kaspa pays it.
Proof of Stake
Collateral does the same job from inside. Validators post the native asset and can lose it for bad behavior the protocol can prove, which trades the power bill for wealth concentration, delegation, and staking providers with real say over who validates. Ethereum, Solana, and most newer networks run some version of it.
Either way the network needs an internal unit of value, and that is the job the token does. It is the thing the ledger tracks, the fee asset that keeps spam costly, the reward that pays for security, and the way people get in and out without an issuer approving each participant. Speed, privacy, security, node cost, and decentralization all pull against each other here, which the how Kaspa compares lays out choice by choice.
Simple check: most tokens do not need to exist. Ask what this one does that BTC, ETH, USDC, or a plain database does not already do.
Two ways to track state
Balances or notes
Account model
A balance table. Alice has 10, sends 3, and now Alice has 7 while Bob has 3. It reads well, and it is the default for smart-contract platforms, until two transactions touch the same balance at once and order starts to matter.
UTXO model
Closer to cash notes than to a table. Alice spends a 10-coin output, Bob receives a 3-coin output, Alice receives a 7-coin change output, and the old 10-coin note is gone for good.
Kaspa uses the UTXO model, same as Bitcoin. That one choice shapes how a node checks a transaction, how much work can run side by side, and what shape later programmability can take.
How to read it: the old output is consumed once. New outputs carry the next spendable state.
Where Kaspa enters
Keep the mining, fix the ordering
Bitcoin-style chains pick one block path at a time and throw the rest away. That is conservative and it works, and it caps how fast blocks can arrive: two honest miners who solve a block seconds apart end up competing, and only one of them counts. Kaspa's bet is that a mined UTXO network can keep both blocks, order them with GHOSTDAG, and bring payments closer to real time without changing what the mining secures. Payments, receipts, wallets, proofs, custody, and later rule-based apps all fit that shape, which how Kaspa compares works through against Bitcoin, Ethereum, and Solana.
Mainnet runs Proof of Work, UTXOs, GHOSTDAG ordering, pruning-oriented nodes, the 10 BPS Crescendo-era network, and Toccata's covenant and proof-check primitives.
A live protocol rule and a shipped product that uses it are separate events. Wallet, explorer, and SDK support for Toccata's new rules each carry a status of their own.
Nobody should read this site as claiming live native DeFi, live DAGKnight, live vProgs, instant finality, or easy adoption. vProgs and native DeFi are roadmap; DAGKnight and RTD-derived app systems are research or future architecture.
Kaspa is testing whether Bitcoin-style mining can feel fast enough for daily use while the base layer sticks to ownership, ordering, and proofs.
BPS means blocks per second, and it is not TPS. Crescendo's 10 BPS era is live; Toccata, KIPs, covenants, and sequencing commitments each carry their own label on the status page, with the primary evidence collected in sources.
Quick corrections
Four things beginners get backwards
A low unit price usually just means more supply
Market cap, float, and real demand carry more signal than the price of one unit.
Speed splits into separate stages
A network can be quick at inclusion and slow at settlement. Confirmation depth, throughput, and how a wallet feels are measured apart.
Influence piles up in more than one place
Mining is the obvious one. Nodes, exchanges, wallets, and RPC providers each form a choke point of their own.
Crypto is several markets sharing one ticker list
Stablecoins, meme coins, governance tokens, and mined money price on completely different things.