Crypto Gloss

TPS

Transactions Per Second (TPS) measures the number of transactions a blockchain network can finalize in one second. It is the most widely cited metric for comparing blockchain throughput and evaluating scalability.

How It Works

TPS is calculated as: TPS = Number of Transactions in a Block / Block Time in Seconds.

For Bitcoin, with an average of roughly 2,500 transactions per block and a 10-minute block time, TPS is approximately 4-7. For Ethereum, with ~15-second block times and variable transactions per block, TPS is roughly 15-30 on the base layer.

Higher TPS generally means a network can accommodate more users and applications without congestion. When transaction demand exceeds TPS capacity, users experience rising gas fees and longer confirmation times.

TPS Across Networks

Network Approximate TPS Consensus Type
Bitcoin 4-7 Proof of Work
Ethereum (L1) 15-30 Proof of Stake
Solana 400-700 (observed) Proof of History + PoS
Visa (for reference) ~1,700 average / 65,000 peak Centralized
Ethereum L2 rollups 1,000-4,000+ Various

The Blockchain Trilemma

TPS is central to the blockchain trilemma, which states that a blockchain can optimize for only two of three properties: decentralization, security, and scalability (throughput). Chains that achieve very high TPS often make trade-offs in decentralization—running fewer validators or requiring high-spec hardware—while highly decentralized networks like Bitcoin prioritize security at the cost of lower throughput.

Layer-2 solutions, including rollups and state channels, address this by processing transactions off the main chain and settling results back to the base layer. This allows ecosystems like Ethereum to achieve high effective TPS without sacrificing base-layer decentralization.

Theoretical vs. Observed TPS

Many projects advertise theoretical TPS based on testnet conditions or maximum block capacity. Real-world observed TPS is usually lower due to variable transaction sizes, network latency, and actual demand patterns. Always compare observed mainnet TPS rather than theoretical claims.

History

  • 2009 — Bitcoin launched with an effective ceiling of ~7 TPS, sufficient for its early use case as peer-to-peer electronic cash.
  • 2017 — The CryptoKitties congestion event on Ethereum highlighted the practical consequences of low TPS, with gas fees spiking and transactions delayed for hours.
  • 2020 — Solana launched mainnet with claims of 50,000+ theoretical TPS, sparking debate about methodology behind TPS benchmarks.
  • 2021 — Ethereum Layer-2 rollups (Arbitrum, Optimism) went live, dramatically increasing effective ecosystem TPS.
  • 2024 — Ethereum’s Dencun upgrade reduced L2 costs via proto-danksharding, boosting aggregate L2 throughput.

Common Misconceptions

“Higher TPS always means a better blockchain.”

TPS is one dimension of performance. A chain with 100,000 TPS but 10 validators is far more centralized than Bitcoin at 7 TPS with thousands of nodes worldwide. Security, decentralization, finality time, and cost per transaction all matter alongside raw throughput.

“Visa does 65,000 TPS, so blockchains need to match that.”

Visa’s 65,000 figure is peak theoretical capacity. Its average is roughly 1,700 TPS. Blockchains serve fundamentally different functions—settlement, programmability, censorship resistance—and direct TPS comparisons with payment networks can be misleading.

Social Media Sentiment

TPS is one of the most debated metrics in crypto communities. Layer-1 projects frequently tout TPS numbers in marketing, leading to “TPS wars” on Twitter and Reddit. Skeptics point out that theoretical TPS is meaningless without real-world usage, while proponents of high-throughput chains argue scalability is a prerequisite for mainstream adoption. The conversation has matured since 2022, with more emphasis on observed TPS and finality time rather than raw benchmarks.

Last updated: 2026-04

Related Terms

Sources