DemoCrypto

Crypto, explained simply, without the jargon

Kaspa: the BlockDAG that wants to speed up proof of work in 2026

Daniel Platteau crypto

An introduction to Kaspa: beyond the traditional blockchain

In the bubbling, ever-changing world of cryptocurrencies, a genuine technological shift has taken place before our eyes in recent years. While Bitcoin paved the way in 2009 by introducing the concept of decentralised digital money, it also laid the groundwork for a fundamental problem known as the "blockchain trilemma". This concept states that a decentralised network can only optimise two of the following three characteristics at the same time: security, scalability and decentralisation.

This is where Kaspa (KAS) comes in, a very ambitious project that aims to push back these limits. In 2026, Kaspa is among the most closely watched layer-1 networks in the ecosystem. Rather than resorting to complex workarounds or sacrificing the security of proof of work in favour of proof of stake, Kaspa has rethought the very architecture of the distributed ledger. Goodbye to the traditional linear chain of blocks, hello BlockDAG (a directed acyclic graph of blocks).

This article takes you deep inside Kaspa. Whether you are passionate about new financial technologies, a curious investor or simply keen to understand how computer science pushes back the limits of what is possible, we will explore together the definition, history, advantages, drawbacks and scientific advances that get people talking about Kaspa.

What is Kaspa? Definition and core concept

To understand Kaspa, you first need to understand the limits of classic blockchains such as Bitcoin. In the Bitcoin network, miners compete to solve a cryptographic puzzle. The first to find the solution creates a block that is added to a single, linear chain. If two miners find a block at almost the same time, the network must reject one of them (what is called an "orphan" block). To prevent this from happening too often, Bitcoin artificially imposes a slow block time (about 10 minutes), which drastically limits the number of transactions per second (about 7).

Kaspa breaks this limitation by using a BlockDAG. In a directed acyclic graph (DAG) of blocks, the network does not reject blocks created simultaneously. On the contrary, it accepts them all and integrates them into a networked geometric structure, weaving a genuine web of transactions.

The GHOSTDAG protocol

At the heart of Kaspa is a consensus protocol called GHOSTDAG (Greedy Heaviest Observed Subtree Directed Acyclic Graph). This ingenious mechanism, which is a generalisation of Nakamoto consensus (Bitcoin's), makes it possible to order these parallel blocks mathematically and securely. Thanks to GHOSTDAG, Kaspa can process multiple blocks per second without compromising the integrity or chronology of transactions.

Today, Kaspa no longer produces one block every ten minutes, but several blocks per second (currently 10 blocks per second since the 2025 Crescendo upgrade), offering near-instant confirmations while preserving the robustness of proof of work (PoW).

An academic history and a fair launch

Unlike many modern crypto projects born of aggressive marketing campaigns or private fundraising (ICOs), Kaspa has its roots in pure academic research.

The founding fathers

Kaspa was founded by Yonatan Sompolinsky, an Israeli computer science researcher at the Hebrew University of Jerusalem, under the supervision of Professor Aviv Zohar. In 2013, Sompolinsky and Zohar had already made their mark on the history of cryptography by co-creating the GHOST protocol, a technology that went on to strongly influence Ethereum's initial architecture.

Continuing his research into solving proof of work's scalability problem, Sompolinsky published in 2021 (with Zohar and other researchers) a founding scientific paper entitled "PHANTOM GHOSTDAG: A Scalable Generalization of Nakamoto Consensus". This paper proved mathematically that it was possible to increase transaction throughput without sacrificing Bitcoin's security assumptions.

A "fair launch"

On 7 November 2021, Kaspa's main network (mainnet) went live. The launch followed the "fair launch" principle. There was no pre-mine (coins created before the public launch), no private sale to venture capital investors, and no allocation reserved for the founding team. Every KAS token in circulation had to be mined transparently.

Kaspa was initially supported by DAGLabs, a start-up co-founded by Sompolinsky, but once the network had launched, the company was dissolved to make way for fully open-source, community-led governance.

A musical monetary policy

A fascinating detail about Kaspa is its issuance policy, known as the "chromatic" phase. Unlike Bitcoin, which halves miners' rewards every four years (the halving), Kaspa reduces its rewards smoothly every month, according to a ratio inspired by the chromatic musical scale (the 12 semitones of an octave). This approach guarantees a predictable, stable supply curve. In summer 2026, an important milestone was reached: around 95% of the maximum total supply of 28.7 billion KAS had already been mined, reinforcing the token's scarcity for decades to come.

The main advantages: speed, security and scalability

Kaspa's design gives it real advantages, which explain its place among the market's large capitalisations.

1. An answer to the trilemma

Kaspa's main strength is its attempt to reconcile security, scalability and decentralisation.

  • Speed: with 10 blocks per second, a transaction on Kaspa appears almost instantly and receives its first confirmations within seconds, an experience close to a card payment.
  • Decentralisation maintained: anyone can run a node or take part in mining, unlike delegated systems (DPoS) where only a few entities control the network.
  • Proof-of-work security: the network is protected by the kHeavyHash hashing algorithm, designed to be efficient on specialised hardware. Like any proof-of-work network, its security against a 51% attack (an entity controlling the majority of the computing power) depends on the total computing power committed: the higher it is, the more expensive the attack.

2. A very lean base layer

Until recently, Kaspa focused on one thing only: being a very fast value-transfer network. By avoiding complex smart contracts from day one, the code stayed lightweight and easier to audit.

The drawbacks and technical challenges to overcome

Despite a very positive picture, Kaspa is not without challenges. The ambition to push back the limits of network physics carries inherent risks, highlighted in particular by experts in 2025 and 2026.

1. The hardware burden and the risk of centralisation

To process 10 blocks per second (with the goal of reaching 100 blocks per second in future), every node on the network has to download, verify and store a huge amount of data. Although Kaspa uses pruning techniques to delete old, unneeded data and lighten the blockchain, the internet bandwidth and computing power required to keep a node up to date are far higher than for Bitcoin. That is where the danger lies: if the required hardware becomes too expensive, only large data centres will be able to run nodes or mine profitably, threatening the decentralisation that is the project's strength.

2. Adoption still young compared with the titans

Although much faster than Bitcoin for transactions, Kaspa has to contend with the "network effect". Bitcoin benefits from a decade's head start, worldwide recognition and billions of dollars of institutional infrastructure (such as ETFs). Kaspa also competes with ultra-fast proof-of-stake networks such as Solana or Sui. Establishing itself as the "new peer-to-peer electronic cash" or as a global settlement layer requires convincing people beyond the circle of tech enthusiasts.

Current scientific research: the arrival of DAGKnight

If you thought the GHOSTDAG protocol was the end of the road, think again. Scientific research around Kaspa has never stopped. The big technical project of 2026 is called DAGKnight.

The problem with GHOSTDAG

The current protocol, GHOSTDAG, has a subtle theoretical weakness: it depends on a hard-coded constant (the parameter k), which represents the maximum time assumed for a block to propagate from one end of the global network to the other. If the global internet suddenly slows down (for example, after a major undersea cable is cut), propagation times could exceed this constant, weakening consensus and lowering the amount of computing power needed to carry out an attack.

A consensus without a fixed parameter

Published by Michael Sutton and Yonatan Sompolinsky, the DAGKnight protocol is the first "parameterless" evolution of Nakamoto consensus. Instead of relying on a pessimistic, fixed estimate of network latency, DAGKnight can analyse the geometric structure of the BlockDAG in real time to deduce the network's actual latency at any given moment.

What does this mean in practice?

  1. The network adapts dynamically: if global internet infrastructure improves, Kaspa automatically becomes faster. If there is a slowdown, the network adjusts to stay safe.
  2. The security threshold remains close to 50% of the computing power, even during periods of high latency, where GHOSTDAG becomes weaker.

This consensus upgrade, proposed through "KIP-2" (Kaspa Improvement Proposal), is being prepared for later deployment. It would be a notable advance for distributed systems.

Recent developments in 2026: the Toccata upgrade

Beyond research, summer 2026 marks an important milestone for the network. Until now regarded solely as a transactional currency, Kaspa becomes programmable with the activation, on 30 June 2026, of the Toccata hard fork.

This major upgrade introduced:

  • Covenants (native smart contracts): making it possible to set strict rules on how funds can be spent, opening the way to decentralised finance (DeFi) applications directly on proof of work.
  • Verification of zero-knowledge proofs (ZK proofs): Kaspa can verify on-chain complex computations carried out off-chain, which lightens the load on the network.
  • The groundwork for native assets: Toccata lays down the rules needed for tokens issued directly at protocol level; the corresponding standard is still being drafted. KRC-20 tokens, for their part, have existed since 2024 through an inscription system.

These changes redefine Kaspa. The network is no longer just "fast money", but aims to become the universal, decentralised settlement layer of the future Web3.

Conclusion: can Kaspa dethrone the giants?

In 2026, Kaspa (KAS) is one of the most original approaches to have emerged since Bitcoin and Ethereum. By combining genuine proof of work with a BlockDAG architecture and the GHOSTDAG and then DAGKnight protocols, it offers a serious answer to the speed and congestion problems of classic blockchains.

However, the road to mass adoption is strewn with obstacles. Maintaining hardware decentralisation in the face of frantic block rates will require constant vigilance, as will the ecosystem's ability to attract developers to the new features introduced by the Toccata fork.

This project, born of Israeli university research and driven by a very active community, shows that innovation in consensus protocols is far from over. Kaspa presents itself as a generalisation of Nakamoto consensus: it remains to be seen whether adoption will follow the technology.

Sources

Disclaimer: this article is provided for information and educational purposes. It is not investment advice. Cryptocurrencies are highly volatile: do your own research and only invest what you can afford to lose.

Related articles

← Back to articles