For years, transactions per second (TPS) has been the headline metric used by blockchain developers to showcase the speed and scalability of their networks. From Bitcoin’s modest 7 TPS to Solana’s claimed 65,000 TPS in benchmark tests, the race to top the TPS leaderboard has fueled marketing campaigns, investor interest, and heated debates across the crypto community.
But in 2024, with the rise of smart contracts, layer-2 solutions, and account abstraction, is TPS still a meaningful measure of blockchain performance—or has it become a misleading, easily manipulated number that tells us more about marketing than real-world utility?
The Rise and Fall of TPS as a Key Metric
In the early days of blockchain, when networks like Bitcoin and Litecoin were primarily used for peer-to-peer payments, TPS made perfect sense. It was a straightforward way to compare how fast a network could process simple transactions—like when Laszlo Hanyecz famously bought two pizzas for 10,000 BTC in 2010.
At that time, "scaling" meant increasing block size or optimizing consensus mechanisms to handle more transactions per second. If crypto were to replace traditional payment systems like Visa—which processes around 1,700 TPS on average—then high TPS was essential.
But as blockchains evolved, so did transaction complexity. Ethereum introduced smart contracts, enabling decentralized finance (DeFi), NFTs, and automated protocols. These operations require far more computational power than a simple wallet-to-wallet transfer.
👉 Discover how modern blockchain networks are redefining performance beyond TPS.
This shift exposed a major flaw in TPS: it treats a $1 token swap and a complex DeFi arbitrage trade as equal—one transaction each—despite vastly different resource demands.
As Offchain Labs co-founder Steven Goldfeder put it:
“Counting TPS is like counting the number of bills in your wallet but ignoring that some are singles, some are twenties, and some are hundreds.”
How Transaction Bundling Breaks the TPS Model
The introduction of account abstraction and User Operations (UserOps) has further undermined TPS as a reliable metric. On Ethereum’s ERC-4337 standard, users can bundle multiple actions—such as approving a token, swapping it, and staking the output—into a single UserOp.
While this improves user experience and reduces friction, it confuses TPS tracking. A single UserOp may replace three or more traditional transactions, yet TPS counts it as just one. Conversely, networks that don’t support batching may register higher TPS numbers despite offering inferior functionality.
Solana’s head of strategy, Austin Federa, notes that transaction complexity has increased dramatically even if raw TPS hasn’t:
“Solana today is powering many more complex transactions than it was back in 2021—even though the number of transactions per second has not astronomically gone up.”
A consensus vote might cost little compute power, while an NFT mint or arbitrage trade could be 100 times more intensive. Yet under TPS, both count as one.
The Problem with Inflated TPS Claims
Criticism of TPS often centers on how it’s measured. Solana, for example, claims up to 65,000 TPS in benchmarks and even 710,000 TPS theoretically. However, real-time data shows closer to 3,000 TPS—and much of that consists of non-user transactions like validator votes.
Solana Compass tracks a “true TPS” metric that excludes these internal operations, reporting around 704 actual user-facing transactions per second. While Federa argues this is still 10x higher than competing chains, it highlights the gap between marketing numbers and real usage.
Other networks have been accused of inflating TPS by counting internal node communications or executing “no-op” transactions—transfers that do nothing but boost throughput stats.
Neil Davies from Input Output (Cardano’s parent company) calls this statistical manipulation:
“Some chains count extensive inter-node messaging as transactions. These aren’t proxies for end-user activity—they’re making a virtue out of overhead.”
Are There Better Alternatives to TPS?
If TPS is flawed, what should replace it? Industry experts are exploring several alternatives:
User Operations Per Second (UOPS)
Proposed by Matter Labs (zkSync), UOPS measures how many user-level actions a network can process per second. Since one UserOp can include multiple steps (e.g., swap + stake), it better reflects actual user activity.
However, UOPS is largely limited to Ethereum-compatible chains using account abstraction. As Federa points out:
“It’s like measuring website traffic by HTTP requests—more requests don’t mean better performance. In fact, efficient sites make fewer.”
👉 See how next-gen blockchains are adopting smarter performance metrics.
Gas Per Second (GPS)
Avihu Levy of StarkWare suggests gas per second (GPS) as a more accurate metric. Since gas represents computational effort on EVM chains, GPS accounts for both transaction volume and complexity.
On Ethereum, every operation—from storing data to executing logic—has a defined gas cost. GPS aggregates these costs per second, giving a clearer picture of network capacity.
The challenge? Not all blockchains use gas. Solana uses Compute Units, Starknet uses Cairo steps, and Aptos uses its own gas model. Without standardized computation units, cross-chain GPS comparisons remain difficult.
Levy argues that a canonical benchmark is needed—a universal test that converts different computation models into comparable units.
Why TPS Still Dominates (For Now)
Despite its flaws, TPS remains the most widely understood metric. As Anthony Rose of Matter Labs admits:
“People understand TPS in a way they don’t understand more meaningful metrics. There’s a trade-off between simplicity and accuracy.”
Most retail investors and casual users don’t grasp gas models or computational complexity. They want a simple number to compare networks—and TPS fits the bill.
Moreover, internal teams still use TPS for development testing. It’s useful for measuring the impact of code changes or network upgrades in controlled environments.
But as Davies notes, TPS matters most when blockchains are treated as speculative assets:
“If you’re shilling your investment, high TPS helps. But if you’re building real applications, you care about latency, cost, and reliability—not bragging rights.”
The Future: Performance Beyond Numbers
Austin Federa believes we’re moving toward a future where speed becomes invisible—just like with smartphones and laptops today.
“The success of the PC industry is that you don’t have to care about the hardware. We’re not there with blockchains yet—but we should be.”
When all major networks are fast enough for seamless DeFi, gaming, and social apps, the focus will shift from raw speed to developer experience, security, and decentralization.
TPS won’t disappear overnight. But as adoption grows and user expectations evolve, the industry will need better ways to measure what truly matters: real utility.
Frequently Asked Questions (FAQ)
Q: What is TPS in blockchain?
A: Transactions per second (TPS) measures how many transactions a blockchain can process in one second. It’s commonly used to compare network speed and scalability.
Q: Why is TPS criticized as a blockchain metric?
A: Because it treats all transactions equally regardless of complexity. A simple transfer and a multi-step DeFi trade both count as “1” transaction, making TPS misleading for modern smart contract platforms.
Q: Can TPS be manipulated by blockchain networks?
A: Yes. Some networks inflate TPS by counting internal operations (like validator votes) or running meaningless “no-op” transactions during benchmarking.
Q: What are better alternatives to TPS?
A: Metrics like UserOps per second (UOPS) and gas per second (GPS) offer more accurate insights by accounting for transaction complexity and computational load.
Q: Does Solana really achieve 65,000 TPS?
A: Only in lab conditions. Real-world throughput is closer to 3,000 TPS, with “true” user transactions estimated around 700–800 TPS after filtering out non-user activity.
Q: Will TPS remain relevant in the future?
A: Likely as a baseline metric, but its importance will decline as account abstraction, layer-2 scaling, and better performance indicators gain adoption.
👉 Explore OKX to stay ahead of blockchain performance trends and innovations.