Solana TPS Explained: What the Numbers Actually Mean
If you have spent any time looking at Solana presales, you have seen the pitch: “65,000 TPS, near-zero fees, the Visa killer.” Solana TPS explained properly is a more boring story than that, and one worth understanding before you put money into anything that markets itself as “Solana-native” or “built on the fastest chain.” This guide walks through what those throughput numbers actually measure, what they leave out, and how to read them as someone who has probably already been burned once.
What “TPS” actually counts
TPS means “transactions per second,” but a transaction on Solana is not the same unit as a transaction on Ethereum or Bitcoin. On Solana, validators broadcast a vote transaction roughly every slot to confirm the chain. These vote transactions are protocol overhead. They do nothing for the user. They are not swaps, transfers, mints, or contract calls.
Most public dashboards historically reported a single TPS number that bundled vote transactions and user transactions together. That is where the headline figures of 3,000-5,000 TPS came from. Once you strip out votes, the picture is very different. According to Solana Compass and the Solana Foundation’s own performance documentation, non-vote TPS on mainnet through 2024 and into 2025 has typically run in the 1,000 to 4,000 range during normal load, with brief spikes higher during meme coin launches or NFT mints.
The 65,000 figure, which still appears in pitch decks for projects building on Solana, comes from internal benchmarks on optimised hardware under controlled conditions. It is a theoretical ceiling, not an observed throughput. Helius, an infrastructure provider with no incentive to talk the chain down, has been fairly direct about this gap in their own technical writing.
Theoretical TPS vs. effective TPS vs. user TPS
Three numbers worth keeping separate in your head:
- Theoretical TPS. What the protocol could do under lab conditions. This is the marketing number.
- Effective TPS. What the network actually processes when you include all transactions, votes included. This is what most explorers show by default.
- User TPS (non-vote). What real users and applications are actually doing. This is the number that matters when you are trying to figure out if a chain can handle a presale launch, a token claim event, or a mint.
When a presale page tells you “Solana handles 65,000 TPS so the launch will be smooth,” the honest translation is “we are quoting the theoretical ceiling and hoping you do not check.” We have written more about how to spot that kind of language in our presale red flags checklist.
What TPS does not tell you
Throughput is one variable. It is not the variable, and a presale that leans heavily on it as a selling point is usually avoiding harder questions.
Finality. TPS measures how many transactions are admitted into blocks per second. It does not tell you when those transactions become economically irreversible. Solana’s optimistic confirmation lands in roughly 400ms per slot, with full finality typically inside 13 seconds, but during periods of network stress these numbers degrade.
Liveness. A chain that processes a lot of transactions when it is up is not the same as a chain that stays up. Solana’s mainnet has had multiple full or partial outages since 2022, documented on the status page archives. Some lasted hours. If you are buying into a token claim that has to happen on a specific day at a specific block, that history matters.
Decentralisation cost. Higher throughput requires beefier validator hardware. Solana’s recommended specs (high core counts, NVMe storage, large RAM, gigabit-plus bandwidth) mean fewer entities can run a validator economically. Throughput and validator decentralisation pull in opposite directions, and no amount of marketing makes that trade-off vanish.
Fee market behaviour. Solana’s local fee markets and priority fees mean that during congestion, your “near-zero fee” transaction may simply not land at all unless you bid up. Users who skipped this detail during the 2024 meme coin cycles found out the hard way when their swaps reverted and their bots ate priority fees on failed attempts.
How this matters for presales
Most Solana-based presales pitch you the chain’s speed as if it were the project’s speed. It is not. A token contract deployed on Solana inherits Solana’s performance characteristics, but the actual user experience during a launch depends on:
- Whether the project’s claim infrastructure is centralised or smart-contract-based.
- Whether the launch is timed against periods of historical network stress (mint events, major listings).
- How the project handles the case where Solana’s mainnet degrades or pauses.
Almost no presale page addresses point three. If you are evaluating a Solana presale, our Solana presale calendar is one place to start, and you should pair it with the presale due diligence framework we use internally.
A practical reading method
When you see a TPS claim on a project’s site or in a thread, run it through this filter:
- Is the number sourced? If it links to a Solana Foundation doc or a public explorer, fine. If it is a screenshot from 2021, ignore it.
- Does the number include votes? If unstated, assume yes, and mentally divide by three to four to estimate user TPS.
- Is the project quoting theoretical or observed? If they are quoting 65,000, they are quoting theoretical.
- Does the project explain what happens during a network slowdown or outage? If not, that is a gap.
We covered some of the same critical-reading habits for other chains in our piece on evaluating Layer 2 throughput claims, and the methodology transfers reasonably well.
What we could not verify
We could not independently benchmark Solana’s mainnet. The TPS ranges quoted in this guide come from third-party dashboards (Solana Compass, Helius writeups) and the Foundation’s own documentation. Those numbers shift week to week, and meme coin or NFT cycles distort them upward in short bursts. Treat the 1,000-4,000 user-TPS figure as a 2024-2025 ballpark, not a fixed value.
Honest summary
Solana is a fast chain, but “fast” in the sense actual users experience is not the 65,000 TPS number printed on marketing slides. Real non-vote throughput sits roughly an order of magnitude below that, finality and liveness vary under stress, and the chain’s outage history is part of the cost of its speed. None of that makes Solana a bad place to build, but if a presale is leaning on TPS as its main selling point, it is using a number it probably does not understand to sell you something it probably has not engineered for the failure modes.