164 Million Bitcoin UTXOs Verified in 72 Seconds. The 9.39 GB File Is the Real Story
A Bitcoin developer just walked the entire unspent output set in 72 seconds, roughly 2.28 million coins per second. But the check arrives as a 9.39 GB artifact, with scripts and signatures left out and no outside review. Speed isn't the hard part of verification anymore. Distribution is.
One hundred sixty-four million unspent bitcoin outputs. Verified in 72 seconds.
That's about 2.28 million outputs a second. Faster than you can scroll a block explorer, and a real piece of engineering by any honest measure.
Now the part that gets buried. The check ships as a 9.39 GB file. Scripts and signatures are omitted. And no reviewer outside the project has gone through it yet.
So which is it? A breakthrough in Bitcoin verification, or a very fast pass through a very partial check?
Both, frankly. And the gap between those two readings is where the next year of this argument lives.
What Actually Got Checked
Start with what the UTXO set is. It's Bitcoin's working memory, the running list of every coin that exists and who can spend it. Every full node carries a copy. It's the thing you're actually trusting when you accept a balance as real.
Walking 164 million of those outputs used to be a grind measured in hours, sometimes days, depending on hardware and how you structure the pass. Doing it in 72 seconds is a different category of result. It means the state can be recomputed fast enough that you could plausibly run it as a checkpoint rather than a chore.
But read the fine print. Scripts and signatures are left out of the file. Those aren't trimming details. Signature checking is the expensive half of validation. Script execution is where the weird edge cases live, the ones that took years to harden and the ones that broke clients in the past. Strip both out and you've timed the cheap half of the pipeline.
There's also a second track here that people are conflating. This isn't browser proof verification. That's a separate effort, a separate set of tradeoffs, and a separate set of promises. The 72-second number belongs to the file-based check, not to anything you can run in a tab with a wallet extension.
And here's the one that should get more attention than it has. External review is outstanding. Nobody outside the team has kicked the tires.
The 9.39 GB Problem
Let me break this down with arithmetic, because the arithmetic is unforgiving.
9.39 GB divided by 164 million outputs is roughly 57 bytes per coin. That's not nothing. It's also not small enough to email, host casually, or expect a hobbyist node runner in a bandwidth-capped country to pull down on a whim.
Here's what matters: Bitcoin Core's chainstate, the actual live UTXO database on disk, sits somewhere around 12 GB these days. So this new artifact is roughly the size of the thing it's supposed to be an alternative to.
That's the tension in one line. If your lightweight check is 9.39 GB, you haven't made Bitcoin lighter. You've renamed the chainstate and added a step.
The numbers tell the story. Fast verification of a large state is only valuable in proportion to how many people can independently reproduce it. One machine, 72 seconds, no external review. That's a benchmark, not a proof. A proof nobody can rerun is a claim with better formatting.
From a risk perspective, the omission of scripts and signatures is the piece I'd flag hardest. Suppose the file check becomes the thing that exchanges, bridges, and wrapped-asset issuers lean on for state attestation. Now you've built a trust layer on a validation pass that explicitly skipped the part where Bitcoin's security actually lives. That's not a knock on the engineering. It's a warning about how the output gets used six months from now by people who never read the caveats.
Who wins here? Teams building light clients, state proofs, and cross-chain messaging. They've been starved for fast, reproducible state commitments, and 72 seconds changes their design space. Auditors and research shops get a new object to pick apart. That's healthy.
Who loses? The purists who've been arguing that verification should stay cheap enough for anyone to do it alone. And anyone who reads the headline and assumes the browser proof story got solved in the same breath. It didn't.
What Verifiable Actually Means
Ask yourself a blunt question. How much of Bitcoin's trust model survives if checking the state requires a 9.39 GB download and a reviewer with a graduate degree?
Some of it survives fine. Most people never verify anything themselves, and that was true before this week. They trust a node operator, an explorer, or an exchange. Pretending otherwise is nostalgia.
But the whole pitch of this asset class is that the number is checkable. Not checkable by a foundation with a server rack. Checkable by you, on hardware you already own, without permission.
So the takeaway isn't that 72 seconds is fake. It's that the interesting frontier moved. Speed is largely solved. Reproducibility isn't. Peer review isn't. Getting the artifact from 9.39 GB down to something a phone can digest, without quietly dropping the signature checks that make the whole thing mean anything, is the actual work.
Watch for two things. First, the external review. If independent researchers reproduce the result and the numbers hold, this goes from a demo to infrastructure. Second, watch whether the scripts and signatures come back into scope. If they do, and the timing stays in the same order of magnitude, that's a very different story than the one you're reading today.
Until then, keep the 72 seconds in your thesis. Just put the 9.39 GB right next to it, in the same font size.
Explore More
Key Terms Explained
An approval term meaning authentic, bold, or worthy of respect.
The first cryptocurrency, created in 2009 by the pseudonymous Satoshi Nakamoto.
A bundle of transactions that gets permanently added to the blockchain.
A website that lets you search and view everything happening on a blockchain, like transactions, wallet balances, and smart contracts.