For years, Transactions Per Second (TPS) has been the go-to metric for blockchain developers eager to promote their latest high-speed networks. Bitcoin, with its modest capacity of around 4–7 TPS, has long served as the baseline. In contrast, networks like Ripple once claimed 1,500 TPS—rivaling Visa’s throughput—though its CTO later admitted those figures were never achieved in practice.
Solana, a major player in the high-performance blockchain space, boasts benchmark numbers as high as 65,000 TPS, though real-world usage hovers around 3,000 TPS. Even more ambitious claims have surfaced from newer blockchains, with some touting theoretical throughputs exceeding 297,000 TPS. But does any of this matter?
As blockchain technology evolves, industry experts are increasingly questioning whether TPS remains a meaningful measure of performance—or if it's become a misleading vanity metric shaped more by marketing than reality.
The Rise and Limits of the TPS Metric
In the early days of cryptocurrency, TPS was the dominant benchmark. Blockchains like Bitcoin and Litecoin were primarily used for simple peer-to-peer transactions—like Laszlo Hanyecz famously spending 10,000 BTC on two pizzas in 2010. At that time, TPS offered a straightforward way to compare how different blockchains handled transaction volume based on block size or consensus mechanisms.
Scaling debates centered on how many transactions a network could process per second, a crucial factor if crypto were to replace traditional financial systems.
But everything changed with the arrival of Ethereum in July 2015. With support for smart contracts and decentralized applications (DApps), transactions became far more complex. A single “transaction” might now involve multiple operations, contract calls, or state changes—none of which TPS could accurately reflect.
👉 Discover how modern blockchains are redefining performance beyond TPS
Account Abstraction and Transaction Bundling: The End of Simple TPS?
One of the most significant advancements undermining TPS is account abstraction, which allows users to bundle multiple actions—called UserOps—into a single transaction.
As Steven Goldfeder, co-founder of Offchain Labs, puts it:
“Measuring raw TPS is like counting the number of bills in a wallet without considering whether they’re singles, twenties, or hundreds.”
Some transactions require minimal computation—like sending SOL or ETH—while others, such as arbitrage trades or NFT mints, can be hundreds of times more resource-intensive. Yet under TPS, they all count as “one transaction.”
Offchain Labs’ Arbitrum One, for example, averages just 9.95 TPS but claims a peak capacity of 40,000 TPS. This discrepancy highlights how real-world usage differs from theoretical benchmarks—and why raw transaction counts misrepresent actual network utility.
Austin Federa, Solana’s Head of Strategy, echoes this sentiment:
“Solana is now at least five times faster than when I joined. But you don’t see it because transaction complexity has increased dramatically.”
Simple actions like consensus voting are low-cost, while advanced DeFi operations can consume vastly more computational resources. Yet both are counted equally in TPS metrics.
“Today’s Solana handles far more complex workloads than in 2021—even if the TPS number hasn’t skyrocketed.”
Solana’s TPS Claims: Misleading or Misunderstood?
Solana has faced intense scrutiny over its TPS figures. Critics argue that much of its reported 3,000+ real-time TPS consists of non-user activities—especially consensus voting transactions, which make up an estimated 80–90% of total volume.
While Solana’s official website displays live transaction counts near 3,000 TPS, third-party tools like Solana Compass track “real TPS”—excluding votes—and report only 704 TPS.
Federa defends the inclusion of votes:
“Voting transactions are real—they pay fees and secure the network. If you exclude them, fine. But even then, Solana’s real TPS is still 10x higher than most other blockchains.”
He adds that dismissing all high-volume activity as spam reflects a lack of serious engagement:
“Once people say it’s all bot traffic from X or Y, we’re no longer talking to serious analysts.”
Can We Trust Any Blockchain Benchmark?
Neil Davies, Systems Performance Scientist at Input Output (Cardano), draws on Mark Twain:
“There are lies, damned lies, and statistics.”
He argues that many blockchain teams manipulate benchmarks to showcase peak performance under ideal conditions—often using trivial transactions like “no-op” calls that do nothing but inflate numbers.
“TPS becomes meaningless when the benchmark doesn’t reflect real user behavior,” Davies says. “It’s less about utility and more about bragging rights.”
Goldfeder agrees:
“They’ll say their chain does 1,000 TPS—but with asterisks. Those are basic transfers or empty operations. Not real-world use cases.”
Federa acknowledges manipulation is widespread:
“Every metric is gamed to some extent. Be skeptical of peak numbers.”
But he stresses that Solana’s traffic reflects genuine network maturity:
“No one’s just spamming transactions to make the numbers look good.”
Beyond TPS: Emerging Performance Metrics
If TPS is flawed, what should replace it?
1. User Operations Per Second (UserOps)
With Ethereum’s ERC-4337 standard enabling account abstraction, UserOps offer a more granular view of user activity. A single transaction might batch multiple UserOps—such as approving a token, swapping assets, and staking rewards—all in one step.
While this improves UX, it further distorts TPS: one transaction now represents several meaningful actions.
Anthony Rose, SVP of Technology at Matter Labs (zkSync), calls UserOps a “more meaningful metric,” but admits adoption is still limited:
“The community hasn’t calibrated around it yet.”
👉 Explore next-gen blockchain metrics shaping the future of Web3
2. Gas Per Second (GPS)
Avihu Levy, Chief Product Officer at StarkWare, proposes Gas Per Second (GPS) as a superior alternative.
Since gas measures computational effort on Ethereum-based chains, GPS accounts for both transaction volume and complexity. A high GPS indicates not just speed but actual processing power.
“If I know my network can process X gas per second, I can compare it meaningfully to others—even if they use different architectures.”
But challenges remain. Non-EVM chains use different units:
- Solana: Compute Units
- Aptos: Gas Units
- Starknet: Cairo Steps
Levy notes the need for a standardized benchmark to translate these metrics across ecosystems—a major hurdle preventing GPS from replacing TPS.
The Future of Blockchain Performance Metrics
Despite its flaws, TPS remains entrenched due to its simplicity. As Rose admits:
“It’s a flawed metric—no strong defense exists. But it’s what most people still rely on.”
There’s a trade-off between clarity, interpretability, and accuracy. While UserOps or GPS offer deeper insights, they’re harder for average users to grasp.
Davies believes TPS will persist until blockchains move beyond speculation and serve real-world applications:
“If you’re just holding crypto as a store of value, TPS only matters if it helps sell your investment.”
Federa envisions a future where speed becomes invisible—like modern smartphones or laptops:
“The PC industry succeeded because you don’t have to care about the specs. We haven’t reached that point with blockchains yet—but I hope we will.”
Until then, developers and users alike must look beyond headline numbers and ask: What kind of transactions are being counted? Who benefits? And what does this really say about usability?
👉 Stay ahead with insights into blockchain innovation and performance evolution
Frequently Asked Questions (FAQ)
Q: What is TPS in blockchain?
A: TPS stands for Transactions Per Second—a measure of how many transactions a blockchain can process in one second. It's commonly used to compare network speed and scalability.
Q: Why is Solana’s real TPS lower than its advertised number?
A: Solana’s peak claims (up to 65,000 TPS) come from lab benchmarks. Real-world throughput (~3,000 TPS) includes many consensus votes. When filtered for user activity only (“real TPS”), the number drops significantly—around 704 according to Solana Compass.
Q: Is high TPS always better?
A: Not necessarily. High TPS achieved through simple or artificial transactions doesn’t reflect real utility. Network security, decentralization, and transaction complexity matter just as much as speed.
Q: What is UserOps in blockchain?
A: UserOperations (UserOps) are bundled actions enabled by account abstraction (e.g., ERC-4337). They allow multiple operations—like token swaps or approvals—to be grouped into a single transaction, improving user experience but complicating traditional metrics like TPS.
Q: Can we compare GPS across different blockchains?
A: Not directly. While GPS (Gas Per Second) better reflects computational load on EVM chains, non-Ethereum networks use different units (e.g., Compute Units on Solana). Cross-chain comparison requires standardized benchmarks still under development.
Q: Will TPS become obsolete?
A: Likely—but not yet. As account abstraction and Layer 2 solutions grow, metrics like UserOps and GPS may replace TPS. For now, TPS persists due to familiarity and lack of widely accepted alternatives.
Core Keywords: blockchain TPS, transactions per second, Solana performance, UserOps, account abstraction, gas per second, blockchain benchmarks, network throughput