Hadamard Get the app

Quantum Position Verification: How NIST Proved Where a Device Is

Quantum Position Verification: How NIST Proved Where a Device Is — Hadamard

Quantum position verification uses entangled photons and the speed-of-light limit to prove that a device is physically at the location it claims, and NIST researchers have now demonstrated it end to end. In a paper published in Nature on November 18, 2026, the team showed that verifiers can confirm a remote device’s position even if they do not trust that device’s quantum hardware. This is something no classical protocol can do.

Below we cover what the experiment did, how the protocol works step by step, why a loophole-free Bell test sits at its centre, and what still keeps it out of banks and power grids.

What NIST showed about quantum position verification

The paper, “Quantum position verification with remote untrusted devices” by Gautam Kavuri and colleagues (DOI 10.1038/s41586-026-11247-3), presents work by NIST physicist Krister Shalm and an international team of collaborators, including NIST physicist Scott Glancy, according to the NIST news release. The arXiv preprint, first posted in January 2026, lists 19 authors. They come from NIST, the University of Colorado Boulder, Oak Ridge National Laboratory, the University of New Orleans, the Joint Quantum Institute, Concordia University, the University of Maryland, Quside Technologies and ICFO/ICREA in Barcelona.

Shalm describes the result as “the first complete demonstration of such a protocol.” The pitch is a change in where trust comes from. In his words: “Our results show how digital security can instead be anchored to the physical world.”

Why classical location proofs can be faked

Most online location checks rest on data the user’s own device reports, or on network paths that someone farther away can relay or imitate. NIST frames this as a real cost. In its release, the agency says cybercrime cost U.S. citizens more than $16 billion in 2024. It adds that far fewer than 1% of cybercriminals are caught, partly because many operate from abroad.

The deeper problem is theoretical. The paper cites work by Chandran et al. showing that no classical protocol allows remote position verification. The intuition is that classical information can be copied freely. A group of colluding attackers placed around the claimed location could share and copy whatever challenges the verifiers send, and together produce answers that look like they came from the right spot.

Quantum mechanics changes this because quantum states cannot be copied. That is the role the no-cloning theorem plays in the NIST protocol.

How the protocol works, step by step

The experiment ran on NIST’s Boulder, Colorado campus. Two verifier stations stood 195 metres apart, linked by electrical cables and optical fibres routed through a research building. The prover, the party claiming a location, sat between them. The full paper gives the precise figures: the verifiers were 195.1 (±0.3) m apart and the prover was 92.8 (±0.1) m from verifier B.

NIST describes each round of the protocol like this:

  1. Entangled pair. One verifier generates an entangled photon pair, keeps one photon and sends the other to the prover.
  2. Random challenges. Both verifiers send random numbers, timed so they reach the claimed location at the same moment.
  3. Measurement. The prover uses those numbers to choose how to measure the photon’s polarisation, then sends the result back.
  4. Round-trip timing. Because nothing travels faster than light, an impostor farther away would answer too late. The timing pins down the position.
  5. Bell test. The verifiers check the correlations between their photon and the prover’s results with a loophole-free Bell test.

The timing has to be tight. According to the paper, the challenge bits reached the prover within 3 (±2) ns of each other. The signal cables carried signals at about 86% of light speed, but the security analysis assumes adversaries can use the full speed of light. That conservative assumption favours the attacker. Entanglement does not provide a faster-than-light channel either, a point we cover in our explainer on whether entanglement is faster than light.

Why a loophole-free Bell test makes it device-independent

A Bell test checks whether measurement correlations are stronger than any classical, pre-arranged strategy could produce. If you want the background, from Einstein’s objections to the experiments that settled them, see our EPR paradox and Bell test explainer.

In this protocol the Bell test does two jobs at once. It shows that the answers came from genuine entanglement measured at the right place and time. It also makes the scheme device-independent: the verifiers do not need to trust the remote quantum hardware. As NIST explains, any error or tampering causes the test to fail, and a pass also confirms the devices worked correctly. The same idea underpins device-independent quantum key distribution.

Why detector efficiency is critical

“Loophole-free” sets a high bar for hardware. If too many photons go missing, even an honest prover cannot pass the Bell test, because lost photons leave room for classical explanations. NIST used superconducting nanowire single-photon detectors, which caught nearly every photon, together with a fast, efficient entangled-photon source it developed about a decade earlier.

The paper reports detection efficiency above 81% at verifier A and at the prover. The protocol’s threshold is 2/3, so the experiment did not need a “fair-sampling” assumption, meaning it did not have to assume that the detected photons were representative of the missed ones.

The experiment in numbers

NIST says the procedure was repeated millions of times over a few minutes. The paper adds the details. Trials ran at about 250,000 per second, a rate set by the Pockels cells that switch measurement settings. A basic run lasted 2 minutes (about 30 million trials) and an extended run 4 minutes (about 60 million trials).

Date Duration Protocol Runs Passed
Sept 20, 2024 About 8 hours Basic 232 224 (96.5%)
Oct 7, 2024 About 7 hours Extended 103 102 (99.0%)
Total 335

The soundness parameter was δ = 2⁻⁶⁴. Soundness here bounds the chance that a cheating prover passes, so the value sets how strong each passed run is as evidence.

How much better than classical? Reading 2.47× and 4.53×

The paper’s abstract gives two comparisons. In one dimension, the localisation is 2.47(2) times smaller than the best classical protocol, and that classical protocol cannot work remotely at all. Compared with classical protocols that have the same latencies, the localisation volume is 4.53(5) times smaller.

Put simply, the quantum protocol fixes the prover to a tighter region than classical timing could, and it does this for a remote, untrusted device. The paper reports these as ratios, not as an absolute size in metres.

Limits, and when banks or power grids could use it

The entangled-attacker problem

Quantum position verification has a known theoretical ceiling. The paper cites Buhrman et al. (2011/2014), who showed that unconditional security against attackers holding unlimited entanglement is impossible. Every practical scheme therefore has to set a bound on what attackers are assumed to have.

The two versions in the experiment cover different threat models:

  • Basic protocol: secure against attackers with no prior entanglement, even if they have unlimited classical and quantum computing and communication during a trial.
  • Extended protocol: allows attackers some prior entanglement, with average robustness bounded by 8×10⁻⁶. The source produces about 2×10⁻³ per trial, so the attacker allowance is roughly 1/250 of the source’s output.

The authors say plainly that a “garden-hose” style attack using two entangled pairs at that level could break the version they ran. They describe the general protocol as loss-tolerant and secure against adversaries with substantial quantum resources. But full security against heavily entangled attackers in a deployed system is still an open problem.

Other practical constraints

The paper lists several further limits:

  • Localisation accuracy is currently limited by transmission latencies and by measurement processing time at the prover station.
  • Buried cables are generally unsuitable, because they do not follow line-of-sight paths.
  • The classical devices, the verifier labs and the timing must still be trusted.
  • Interference can cause denial of service, but it cannot break soundness.

Realistic timelines and use cases

NIST lists possible applications:

  • protecting healthcare records
  • stopping remote attacks on nuclear plants, power grids and water treatment
  • blocking unauthorised trades and authenticating stock transactions
  • limiting data access to people inside a secure building
  • stopping phishing by video or phone from unknown locations

Glancy is cautious about timing. He says real-world use may be more than a decade away, because verifying a bank customer would require high-fidelity quantum networks that share entangled photons between the customer and the bank.

Until then, the quantum threat to today’s systems is handled by a different tool: post-quantum cryptography, which runs on ordinary hardware. At Hadamard we build BB84, an Ethereum Layer 2 that replaces ECDSA signatures with hash-based one-time addresses. For where browsers, messengers and blockchains stand on that migration, see our Post-Quantum Migration Tracker.

FAQ

Can quantum position verification replace GPS or IP-based location checks?

Not in the near term, and it does a different job. It proves a claimed location to trusted verifiers, using light-speed timing and a Bell test on entangled photons. It needs dedicated verifier stations, line-of-sight links and quantum hardware. NIST expects real-world use to be possibly more than a decade away.

Can a group of attackers working together fake their location?

The protocol is designed to stop this. The speed-of-light limit means distant attackers answer too late, and the no-cloning theorem stops colluders from copying the quantum correlations. The version demonstrated tolerates only a small amount of pre-shared entanglement. Full security against heavily entangled attackers is still an open question.

Why does the protocol need a loophole-free Bell test and highly efficient detectors?

The Bell test makes the scheme device-independent: tampering or faults cause it to fail, so the remote hardware does not need to be trusted. That only works if few photons are lost. The protocol requires efficiency above 2/3, and the experiment measured above 81%.

When could banks or utilities actually use it?

Probably not soon. Glancy says practical use may be more than a decade away, because it depends on high-fidelity quantum networks that can distribute entangled photons between parties such as a customer and a bank.

Sources