How Does Kaspa Work?

GHOSTDAG — a blockDAG consensus protocol that extends Nakamoto consensus by allowing parallel blocks to coexist rather than discarding orphans. The protocol orders blocks within a directed acyclic graph while maintaining the security guarantees of single-chain PoW, enabling 10 blocks per second with a target of 100 BPS

GHOSTDAG is a genuine academic innovation by Dr. Yonatan Sompolinsky. Unlike traditional blockchains that select one block per height, GHOSTDAG includes all blocks and orders them, increasing throughput without sacrificing security. This approach has not been deployed at scale by any other major protocol.

kHeavyHash mining — a proof-of-work algorithm combining SHA-256 with a heavy matrix multiplication step, designed to be ASIC-friendly while maintaining energy efficiency relative to pure SHA-256 mining

While kHeavyHash has specific properties, PoW mining is a well-understood mechanism. The transition from GPU to ASIC mining follows the standard PoW maturation pattern seen with Bitcoin and Litecoin.

Chromatic emission schedule — block rewards decrease geometrically based on a musical 12-note chromatic scale, halving approximately annually but smoothed so that rewards decrease by a factor of (1/2)^(1/12) each month. Started at 440 KAS per block in May 2022

The chromatic emission schedule is a novel approach to supply reduction. While halving schedules are standard (Bitcoin), Kaspa's monthly smooth reduction based on musical intervals is unique. Nearly 95% of the 29B total supply will be mined by July 2026.

UTXO transaction model — Kaspa uses the Bitcoin-style UTXO model for transactions, providing parallel transaction processing capabilities and simple verification of transaction validity

Standard UTXO model as used by Bitcoin. Provides good parallelization for payment transactions but limits programmability compared to account-based models.

High-frequency block propagation — at 10 blocks per second, Kaspa requires efficient block propagation across the mining network. The GHOSTDAG protocol tolerates delayed blocks by including them in the DAG rather than orphaning them

The tolerance for parallel blocks reduces the orphan rate penalty, but the fundamental block propagation challenges at high frequency are addressed through the GHOSTDAG protocol itself.

BlockDAG security at high throughput — the GHOSTDAG protocol's security proofs assume honest majority hashrate, but at 10+ blocks per second, the attack surface expands because an attacker can create more parallel blocks within the DAG. The interplay between block rate and security margin is less empirically tested than Bitcoin's single-block model

Mining economics post-emission — with 95% of supply mined by July 2026, miner revenue will transition from block rewards to transaction fees. If Kaspa's UTXO-only design generates insufficient fee revenue (no smart contracts, limited DeFi), mining profitability could drop sharply, potentially reducing hashrate and security

Network partition amplification — at 10 blocks per second, network partitions create rapid DAG divergence. While GHOSTDAG handles this gracefully in theory, real-world network fragmentation could create extended periods of uncertainty about transaction finality

Security budget cliff — the chromatic emission schedule aggressively reduces block rewards, and without smart contract fee revenue, the UTXO-only transaction model may not generate sufficient fees to sustain mining security long-term

ASIC centralization — as kHeavyHash ASICs become dominant, mining may concentrate among large operators with preferential ASIC access, reducing the decentralization that GPU mining initially provided. At 10 BPS, miners with lower latency connections gain a structural advantage in block inclusion