The Problem Nobody Solved (Until Now)
In 2022, a group of mathematicians proposed a quantum analogue of one of the most useful tools in classical probability: the Wasserstein distance, sometimes called the earth mover's distance. The classical version, in plain terms, asks how much work it would take to reshape one pile of dirt into another, where work is measured as the amount of dirt times the distance it has to travel. Friedland and collaborators asked: can you do the same thing for quantum states? Their 2022 paper in Physical Review Letters, indexed as PRL 129, 110402, proposed a quantum cost matrix and a quantum coupling, then conjectured that for one particular quantum cost matrix — and for cost matrices in a small neighborhood of it — the resulting quantity would behave as a true mathematical distance, not merely a semidistance. In July 2026, that conjecture collapsed. A new comment paper, posted to arXiv as [arXiv:2607.07764] on 2026-07-08, exhibits an explicit family of triples of quantum states for which the triangle inequality fails. The triangle inequality is the property that makes something a distance: if state A is close to state B, and B is close to C, then A should be close to C. When this fails, the quantity in question stops being a true distance, full stop.
Why does this matter? Because distance measures between quantum states are the bedrock of quantum error correction. Every time a fault tolerant quantum computing system checks whether a logical qubit has drifted from its target state, it implicitly relies on a notion of distance. If that notion is mathematically broken, then any code or protocol built on top of it is built on sand.
The Core Finding
The new paper, authored by researchers whose identities are not disclosed in the publicly available metadata, does something disarmingly simple: it constructs counterexamples. Rather than proving a positive result, it shows the conjecture is false by exhibiting, in the authors' own words, an explicit family of triples of states for which the triangle inequality fails. This is a common technique in mathematics — to disprove a universal claim, you do not need a general argument; you just need one well-chosen counterexample. The author group here produced a family, meaning infinitely many, which is more decisive than a single example. The result is published as a Comment on the 2022 paper, following the standard practice in physics journals where a later paper formally responds to and disputes an earlier claim.
Think of it like this. Imagine someone tells you that, for any three cities, the direct distance from A to C is always less than or equal to the distance from A to B plus the distance from B to C. You say: here is a counterexample. You place A, B, and C on a curved surface and choose locations where the direct route goes the wrong way around relative to the two-step route. That is essentially what these authors did, but in the abstract space of quantum states, with the specific quantum Wasserstein distance proposed by Friedland et al. in 2022.
We disprove these conjectures by exhibiting an explicit family of triples of states for which the triangle inequality fails.
The State of the Field
Before 2022, classical Wasserstein distances had become foundational in machine learning, optimal transport theory, and statistics. Translating them to quantum mechanics is harder than it sounds. In the classical setting, you can move probability mass from one place to another. In quantum mechanics, the no-cloning theorem prevents you from copying states, and entanglement means that two quantum systems can be correlated in ways that have no classical analog. Friedland et al. proposed a clever workaround using quantum couplings — joint quantum states that mimic the role of couplings in classical optimal transport. They proved the resulting quantity always satisfies symmetry and one of the two distance axioms, but the triangle inequality held only in special cases. They conjectured it would hold for a particular natural choice of quantum cost matrix and for cost matrices close to that choice. The new comment shows this conjecture is wrong.
This is happening against a broader quantum computing landscape where distance measures are increasingly practical. Since 2024, multiple groups have demonstrated surface code error correction below the threshold needed for scalable fault tolerant quantum computing. Hardware vendors pursuing logical qubit prototypes have pushed fidelities upward. The mathematics underlying these advances, including how we measure the distance between a noisy state and a code space, is under active scrutiny. A 2026 disproof of a 2022 conjecture is small in scale, but it is the kind of correction that keeps the field's foundations honest.
From Lab to Reality
For scientists working in quantum information theory, the immediate consequence is methodological. Any future work that used the Friedland et al. quantum Wasserstein distance for the specific cost matrix in question, or for nearby cost matrices, must now find a different tool. Researchers in the optimal transport and quantum information communities who have built on the 2022 paper will need to revisit their results. The question now is not whether the conjectured property holds, but what weaker, still-useful properties do hold, and whether a modified definition of the quantum Wasserstein distance can rescue the triangle inequality.
For engineers, the impact is indirect but real. Quantum error correction systems in the laboratory rely on classical computers to process syndrome data and decide what corrections to apply. Some of these classical algorithms use distance measures as part of their decoders. The specific disproof in the comment paper does not invalidate the surface code, the leading candidate for fault tolerant quantum computing. But it does flag a class of theoretical tools that were being proposed as useful for analyzing quantum states, and tells the community they do not behave as previously believed. Quantum hardware roadmaps running from 2026 to 2030 are unlikely to be affected directly, since they are built on more mature distance measures and code distances.
For investors, the market impact is negligible in the short term. The quantum error correction market is forecast by multiple analysts to grow substantially through the 2030s as fault tolerant quantum computing matures, but a single mathematical disproof of a 2022 conjecture does not move the needle on capex decisions at major hardware vendors. It does, however, matter for any startup or research group that had been building tools around the Friedland et al. distance; their products and grant proposals need adjustment.
What Still Needs to Happen
Two technical challenges remain. First, the community needs a clear picture of when, exactly, the Friedland et al. quantum Wasserstein distance does satisfy the triangle inequality. The conjecture is false in general, but it might still hold for restricted classes of quantum states — for instance, pure states, separable states, or states with bounded entanglement entropy. Mapping out the boundary between works and does not work is the natural next step. Second, the field needs alternative quantum distance measures that do behave as proper distances, ideally ones that retain the operational interpretation of the Wasserstein distance as a transport cost. Research groups active in quantum optimal transport, including affiliates of the Perimeter Institute and several European quantum information centers, are likely to pursue this in the years ahead.
Both challenges are open-ended. There is no firm timeline, and no commercial milestone is tied directly to them. Realistically, expect incremental theoretical progress over 2026 to 2028, with any new distance measure taking years to propagate into the tools used by experimental groups. The path from a theoretical quantum Wasserstein distance to a working piece of fault tolerant quantum computing hardware is, charitably, a decade long.
Conclusion
In short: a 2022 conjecture about the quantum Wasserstein distance — a candidate metric for tracking how far quantum states drift during quantum error correction — has been disproved by an explicit infinite family of counterexamples that violate the triangle inequality, forcing researchers to reexamine the foundations of quantum state distance measures used in fault tolerant quantum computing research.
Frequently Asked Questions
What is the quantum Wasserstein distance?
The quantum Wasserstein distance is a proposed way to measure how different two quantum states are, modeled on the classical earth mover's distance from probability theory. It was introduced in a 2022 paper by Friedland et al. and is meant to capture the cost of transporting one quantum state into another using a quantum analogue of a transport plan.
How does the triangle inequality failure matter?
The triangle inequality is what makes a quantity a true mathematical distance. If the inequality fails, the quantity is not a distance in the strict sense, and tools that rely on distance properties, such as clustering, nearest-neighbor algorithms, or bounds on error accumulation, can give incorrect results.
How does this compare to other distance measures in quantum information?
Other well-established quantum distance measures, including the trace distance, the fidelity, and the Bures distance, do satisfy the triangle inequality and have been used for years in quantum error correction. The quantum Wasserstein distance was proposed as a more operationally meaningful alternative, but the new comment shows it does not always behave as a true distance.
When could this affect commercial quantum computing systems?
In the short term, almost never. Commercial quantum computers built on surface codes and other established error correction schemes rely on different mathematical tools. The impact is on the theoretical and algorithmic layer, which may take 5 to 10 years to propagate into deployed systems, if it ever does.
Which industries would benefit most from a working quantum Wasserstein distance?
Pharmaceutical and materials companies using quantum simulation could benefit from better tools for benchmarking quantum state preparation, while financial firms exploring quantum machine learning could use improved distances for portfolio state discrimination. Neither is directly served by the 2026 disproof, but both stand to gain from a future, corrected version.
What are the current limitations of this research?
The comment paper is a disproof, not a construction. It tells the community what does not work, but does not yet provide a replacement distance measure that is both operationally meaningful and mathematically well-behaved. Building such a replacement is the open challenge that the 2026 result creates.