A Bitcoin developer says he verified a snapshot of 164.5 million unspent outputs, a 9.39 GB dataset, in about 72 seconds on modest hardware. Published on Delving Bitcoin on November 13, 2026, the experiment does not change consensus rules, but it points to a serious route for cutting the time needed to bring a full node online. Here is what it measures, and where its limits lie.
๐ Key takeaways
- A 9.39 GB snapshot covering about 164.5 million UTXOs at block 972,500 was verified in 72 seconds.
- A STARK proof covers block headers from genesis and UTXO changes between blocks 935,001 and 972,500.
- Scripts and signatures are not checked, and no external security review has taken place yet.
- Bitcoin Core does not support this integration, and no public measurement of the time saved exists.
What the experiment actually measures
On November 13, 2026, the developer known as l-henri published on Delving Bitcoin an experiment verifying a snapshot of the blockchain’s transactional state. The file tested weighs 9.39 GB and contains roughly 164.5 million unspent transaction outputs, or UTXOs, as of block 972,500. A UTXO represents bitcoin that has been received but not yet spent. The full set of these outputs forms the database each node consults to determine which bitcoins remain available.
The wording deserves a note. The 164.5 million figure refers to UTXOs, not to 164.5 million bitcoins, even though several headlines and the Bitcointalk digest describe them as “unspent coins”. The main headline also rounds 164.5 million down to 164 million. This is a difference in vocabulary, not a contradiction in the reported result.
The test ran on an arm64 machine with about 0.5 GB of memory. The operation took roughly 72 seconds, most of it spent reading the 9.39 GB file. The number of blocks processed, the file size, the memory used and the runtime are consistent across the different reports of the experiment.

A STARK proof instead of replaying history
The system relies on STARK proofs (Scalable Transparent Argument of Knowledge). They let a computer verify that a given computation was carried out correctly, without re-running every step of that computation. In the experiment, the proof covers Bitcoin block headers from the start of the chain and checks the changes made to UTXOs between blocks 935,001 and 972,500.
The starting point is an existing Bitcoin Core snapshot, considered reliable, at block 935,000. Transactions prior to that point were therefore not replayed. Consistent sources distinguish two things: validating the header chain from genesis, and rebuilding the full set of transactions and the UTXO state, which was not performed.
| Parameter | Reported value |
|---|---|
| Snapshot size | 9.39 GB |
| UTXOs covered | about 164.5 million |
| Final snapshot block | 972,500 |
| Proof starting block | 935,000 (Bitcoin Core snapshot) |
| Proven change range | 935,001 to 972,500 |
| Verification time | about 72 seconds |
| Memory used | about 0.5 GB (arm64) |
| Compact proof | 1.6 MB, verified in about 0.5 s in a browser |
Two measurements should be kept apart. Reading and comparing the 9.39 GB file took about 72 seconds. Verifying the 1.6 MB compact proof, presented separately, took about half a second in a web browser. That second figure does not cover the comparison with the full file.
“The verification of the 9.39 GB file is separated from the verification of the proof in the browser, with scripts and signatures omitted and external review still to be done.”
BTCFans, article dated November 18, 2026
How a node starts today
Bitcoin nodes strengthen network security by independently verifying transactions and enforcing protocol rules. For a fully validating node, setup normally means downloading and processing the entire transaction history to rebuild the set of available coins. That step consumes time, bandwidth and computing resources before the node becomes operational.
Bitcoin Core already offers a shortcut called assumeUTXO. It lets a node start from a snapshot of the transactional state, then continues checking the history in the background to confirm that the snapshot correctly reflects the chain’s data.
l-henri’s proposal would accept snapshots accompanied by cryptographic proofs and skip this additional background history validation. That would depart from current behaviour, in which Bitcoin Core still checks historical records after loading a snapshot. The stated goal is to reduce setup time and the resources needed to run a new node.
Limits, caution and next steps
The system remains experimental and should be read as a research preview. It performs neither script checks nor signature checks, so it does not yet offer the same guarantees as a fully validating node. No external security review has been conducted, and Bitcoin Core does not currently support the proposed integration.
The next step is to determine whether the proof system can provide validation guarantees strong enough to replace historical checks without weakening the security assumptions nodes rely on. Three questions frame this work:
- Do the STARK proofs cover all the consensus rules required, including scripts and signatures?
- Can an independent review validate the code and the cryptographic assumptions?
- Can the real gain on a complete node setup, compared with assumeUTXO, be measured publicly?
No public test cited in the sources yet measures this saving in time or resources across the full setup process. The 72 seconds concern snapshot verification, not the installation of a node from start to finish.
Consistent coverage, no market reaction
CryptoSlate published its report on November 17, 2026 at 2 p.m. US Central Time. BobClawblaw’s Bitcointalk digest, posted November 17, confirms the November 13 date, the 72 seconds, the 9.39 GB file, the 164.5 million UTXOs, the arm64 machine, the 0.5 GB of memory, the 1.6 MB proof and the blocks involved. It also highlights the start from the reliable snapshot at block 935,000 and the experimental nature of the project.
A CryptoNews republication repeats the same figures without adding any independent result. The BTCFans page, dated November 18, 2026 at 4:00:40 a.m. Beijing summer time, credits CryptoSlate and notes that the experiment does not affect Bitcoin’s consensus rules. The sources report no direct market reaction, no adoption decision by Bitcoin Core and no deployment of the technology.
Conclusion: a technical step to watch
The demonstration shows that a STARK proof can attest quickly to a state of 164.5 million UTXOs, using modest memory. It does not yet show that a node could forgo historical validation without losing guarantees. The scenarios therefore remain open.
In a favourable scenario, external review, script and signature checks, and a public measurement of the full setup process could pave the way for future integration, subject to developer consensus. In a status quo scenario, assumeUTXO remains the reference and the experiment stays a research tool. In a rejection scenario, current security assumptions could prove too costly to replace. No protocol change has been decided at this stage.
Sources
- Delving Bitcoin, publication by l-henri, November 13, 2026
- CryptoSlate, report of November 17, 2026
- Bitcointalk, BobClawblaw digest of November 17, 2026
- CryptoNews, republication
- BTCFans, article of November 18, 2026
Cet article est publiรฉ ร titre informatif et รฉducatif. Il ne constitue en aucun cas un conseil en investissement. Faites vos propres recherches (DYOR) avant toute dรฉcision.


