The Blockchain Trilemma

A few weeks ago I was at an Ethereum event recapping the activities at ETHCC. At one point, one of the speakers said something along the lines of ‘..today we have a new blockchain popup every other day’. The crowd laughed. I chuckled.

Later on, I wondered why this was the case, and came across a foundational computer science problem called the blockchain trilemma.

Blockchains rely on thousands of machines reaching consensus on whether a transaction is valid before the state of the chain can be updated. This creates a natural engineering constraint, and chains are ultimately defined by how they balance three competing priorities:

Decentralisation — distributing control across many independent nodes to ensure the network is censorship resistant and has no single point of failure

Security — ability of the network to defend against attacks and ensure that data (like transaction history) remains irreversible and tamper-proof

Scalability — the capacity to process a high volume of transactions quickly and cheaply

The trade-off

The rule states that increasing one property forces a trade off in at least one of the other two:

Decentralisation + Security = Slow speed
Maximise decentralisation and security with thousands of nodes reaching consensus, however this means transaction throughput remains limited (about seven transactions per second) and congestion fees high. Scalability is sacrificed.

Scalability + Security = Centralisation
If you want to maximise scalability (50,000 TPS), you can’t wait for thousands of global nodes to agree. Instead you limit the work to a few hand picked, powerful servers. This achieves both high security and scalability — but decentralisation is sacrificed.

Decentralisation + Scalability = Insecurity
A network that attempts to process a massive amount of transactions across a group of loosely coordinated nodes becomes highly vulnerable — attackers can more easily overwhelm consensus and rewrite or manipulate history.

How chains address the trilemma

‘The store of value specialist’: Bitcoin
Maximises decentralisation and security, relying on thousands of independent miners and an immutable Proof of Work protocol. Transaction throughput remains limited (around 5-7 transactions per second), leading to slow confirmations and high fees during congestion.

‘The world computer’: Ethereum
Originally focused on decentralisation and security but encountered scalability challenges as it became more popular. Transitioned to a Proof of Stake mechanism, reducing energy consumption and laying the groundwork for scaling — mostly through L2s, with further protocol upgrades still in progress.

‘The speed demon’: Solana
Puts scalability front and center, making it one of the fastest blockchains for throughput and speed. Its focus on performance has attracted many DeFi, NFT, and gaming applications, but its validator structure has resulted in occasional outages, as well as ongoing debates about whether it is as decentralised as other blockchains.

In addition to being subject to the blockchain trilemma, each blockchain is trying to optimise for a certain use-case — e.g. DeFi, supply chain, identity, gaming — while having their own take on how best to reach consensus (the agreed state of truth on the chain, e.g. PoW, PoS) and what access level the chain should have (public vs private). Which explains why there’s a new one every other day.