What Is Kaspa? A Plain-English Technical Overview

Why waiting ten minutes for a block feels so old

If you have ever sent Bitcoin and watched the transaction sit there, you already know the trade-off at the heart of proof-of-work. Bitcoin produces one block about every ten minutes, and that slow rhythm is not an accident. It is what keeps the network safe when miners on opposite sides of the planet find blocks at almost the same moment.

Kaspa starts from a simple question: what if a proof-of-work network could make blocks every 100 milliseconds without throwing that safety away? In this article we will look at what Kaspa is, how its BlockDAG and the GHOSTDAG protocol make fast blocks possible, which parts of the roadmap are live in October 2026, and where the design still has limits.

Kaspa in one paragraph

Kaspa is a proof-of-work cryptocurrency that launched in November 2021 with no premine, no presale and no coins set aside for the team. Miners secure it with the kHeavyHash algorithm, and its native coin is KAS, with a maximum supply of about 28.7 billion. What makes it unusual is the shape of its ledger: instead of a single chain of blocks, Kaspa keeps a BlockDAG, a structure in which blocks can be created in parallel and still all count. Since May 2025 the network produces 10 blocks per second.

You can follow the live numbers, from price to hashrate, on the KaspHub dashboard.

Why fast blocks are hard on a normal blockchain

On a traditional blockchain, every new block points to exactly one parent. That works well while blocks are rare. When two miners find a block at nearly the same time, though, the network has to pick one and discard the other. The losing block is called an orphan: its proof-of-work was real, but it ends up wasted.

Now imagine shortening the block time from ten minutes to one second. Blocks would collide all the time, because a block needs a few hundred milliseconds just to travel around the world. More orphans mean more wasted work, and it hurts small miners most, since large, well-connected pools hear about new blocks first. That is the real reason Bitcoin is slow: speeding it up naively would make it less secure and more centralized.

Kaspa's answer is not to make blocks collide less. It is to stop throwing the colliding blocks away.

How the BlockDAG keeps every block

A DAG, or directed acyclic graph, is a set of points connected by one-way arrows that never loop back on themselves. In Kaspa, each point is a block, and each arrow says "this block saw that earlier block". When a miner creates a block, it references every recent block it knows about, not just one.

Think of a single-lane road where cars must line up behind each other, compared with a braided river where several streams flow side by side and keep rejoining. Blocks mined at the same time on opposite sides of the world simply become siblings in the DAG. Their work is kept, and so are their transactions.

Keeping everything raises an obvious question, though. If two parallel blocks contain conflicting transactions, which one wins? That is the job of GHOSTDAG.

GHOSTDAG: turning a tangle of blocks into one order

GHOSTDAG is the consensus protocol that gives the BlockDAG a single, agreed order. It was developed from the PHANTOM research paper by Yonatan Sompolinsky and colleagues, and it works roughly like this:

• 1Each block looks at the blocks it references and colors them: blocks that are well connected to the rest of the DAG are 'blue', blocks that look isolated or withheld are 'red'.
• 2A parameter called k sets how many parallel blocks honest miners are expected to produce, so the protocol knows how much parallelism is normal.
• 3Every block picks a 'selected parent', the one with the most blue work behind it. Following selected parents from block to block gives a main spine called the selected chain.
• 4Each block on that chain then orders the other blocks it merged in a deterministic way, so every node arrives at exactly the same sequence of transactions.

When two transactions conflict, the one that comes first in this order is valid and the other is rejected. Red blocks still stay in the DAG, but they have less influence on the order. If you want the longer version, with the trade-offs of both designs side by side, see our article on BlockDAG vs blockchain.

The numbers that matter today

Here is where Kaspa stands in October 2026. The figures change slowly, except the reward, which drops a little every month.

A note on throughput: since the move to 10 blocks per second, secondary sources have reported sustained capacity in the low thousands of transactions per second, with peaks around 5,500. The Kaspa team itself expected an 8 to 9 times improvement over the old one-block-per-second network, rather than a full 10 times, because parallel blocks sometimes include the same transactions.

What changed in 2025 and 2026

Older articles about Kaspa often describe one block per second and smart contracts "on the roadmap". Both statements are now out of date.

• May 2025, Crescendo: the hard fork that raised the block rate from 1 to 10 blocks per second. It was made possible by Rusty Kaspa, a full rewrite of the node software in Rust.
• March 2026, Igra: an EVM-compatible network built on top of Kaspa launched its public mainnet, so Solidity smart contracts can now settle on Kaspa.
• June 2026, Toccata: a hard fork that added covenants (rules that limit how coins can be spent in the future) and native verification of zero-knowledge proofs to the base layer.
• Still in progress: DAGKnight, a successor to GHOSTDAG that adapts to network latency, and vProgs, a model for fully programmable applications. A move toward 100 blocks per second is targeted for 2027.

KRC-20 tokens, a token standard similar to Ethereum's ERC-20, have existed since 2024 through an indexer called Kasplex. They work, but they are not part of the base protocol the way Toccata's covenants are.

How new KAS are created

Kaspa has no team allocation, so every KAS in circulation came from mining. The emission follows a fixed schedule. After a short initial phase in 2021 and 2022, the network entered what the community calls the chromatic phase: the block reward falls by about 5.6% every month, which works out to exactly half every twelve months.

In practice, this means issuance shrinks quickly. By October 2026, around 96.7% of the maximum supply had already been mined, and it should pass 99% around mid-2028. Our tokenomics page shows the full emission curve and how the inflation rate falls over time.

The trade-offs worth knowing

Fast blocks and a fair launch are real strengths, but they are not the whole picture. These are the limits a careful reader should keep in mind.

Mining is industrial. kHeavyHash has been dominated by specialized ASIC machines since 2023, and most hashrate flows through a handful of large pools. That is normal for a mature proof-of-work coin, but it means "anyone can mine with a graphics card" is no longer true.

Nodes work harder. Ten blocks per second means more headers and more data to move around. Kaspa keeps disk usage under control by pruning old block data, but bandwidth and processing needs are higher than on a slow chain.

The security budget is smaller. Kaspa's total hashrate, and the value of the rewards paying for it, are a small fraction of Bitcoin's. Attacking the network is still expensive, but the absolute cost is lower.

Programmability is young. Covenants and the Igra network are recent, and the tooling around them is still maturing. An ecosystem the size of Ethereum's does not appear in a few months.

Access is uneven. KAS trades on many exchanges, including Kraken, KuCoin, Bybit, Gate and MEXC, but not on the spot markets of Coinbase or Binance. Our guide on where to buy Kaspa covers the options in detail.

So, what is Kaspa good for?

Back to the opening question: can proof-of-work be fast without giving up its security model? Kaspa's answer, so far, is a qualified yes. By keeping parallel blocks instead of discarding them, it reaches 10 blocks per second and confirmations in seconds, while miners still secure the network the same way they secure Bitcoin.

The open question is no longer whether the BlockDAG works, but what people will build on it. That depends on how covenants, the Igra network and the coming DAGKnight upgrade are adopted over the next few years. If you want to go deeper, our comparisons with Bitcoin, Ethereum and Solana are a good next step.