⚛️ quantum physics

The NISQ Trap: Eight Years of Demonstrations the Hardware Was Built to Lose

This paper argues that the NISQ era has been a closed loop of failed "quantum advantage" claims because the specific circuit characteristics that current noisy hardware can execute are inherently the same ones that allow efficient classical simulation, proving that genuine quantum advantage remains contingent on the realization of fault-tolerant quantum computing.

Original authors: Amit Hagar

Published 2026-07-09

📖 6 min read🧠 Deep dive

CC BY 4.0

Original authors: Amit Hagar

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

The Big Picture: The "Eight-Year Loop"

Imagine a group of engineers building a new type of car engine (Quantum Computers) that is supposed to be faster than any car ever made. For eight years, they have been showing off "flagship" races where their new engines seem to win against old, standard engines.

However, this paper argues that every single one of those races was rigged by the rules of the track.

The author, Amit Hagar, claims that the engineers were forced to pick race tracks that were easy for the old engines to run on, just because the new engines were too fragile to handle the hard tracks. Every time they claimed a "win," a mathematician eventually came along and said, "Actually, we could have solved that same easy track with a pencil and paper just as fast."

The paper calls this the "Closed Loop."

The Hardware: The new quantum computers are noisy and fragile. They can only run on very simple, short, or highly structured tracks.
The Software: Because those tracks are simple, classical computers (the old engines) can actually simulate them perfectly well.
The Result: The quantum computer never actually gets to prove it's faster, because it's never allowed to run on a track that is hard enough to stump the classical computer.

The Core Problem: Two Sides of the Same Coin

The paper explains that there is a "two-faced" constraint:

Face A (The Hardware): Because the quantum chips are noisy, they can only handle simple circuits (like a short, straight road) or very specific patterns (like a perfectly symmetrical dance).
Face B (The Math): The exact same features that make a circuit easy for the noisy hardware to run (short length, symmetry) are the exact features that make it easy for a classical computer to solve.

The Analogy:
Imagine you are trying to prove that a new, shaky bicycle can go faster than a car.

To keep the bike from falling over (noise), you must ride it on a perfectly flat, straight, 100-meter track.
But on a flat, straight, 100-meter track, a car can easily drive alongside you and match your speed.
You claim, "Look! The bike is as fast as the car!"
The paper says: "No, you just picked a track where the car is already fast. You haven't proven the bike is special yet."

The "De-Quantization" Sequence

The paper lists a series of events from 2018 to 2026.

The Pattern: A company announces a "Quantum Advantage" (a win).
The Reaction: Within 18 months, a team of theorists publishes a paper showing how to simulate that exact experiment using a classical computer.
The Reason: The experiment was designed to work on the noisy hardware, which meant it had to use "short circuits" or "special patterns." Those are the exact things classical computers are good at.

The paper notes that out of more than 30 major announcements, every single one has been "de-quantized" (proven to be simulatable by classical computers), with one very specific exception that the paper treats with skepticism.

The One Exception: The "Quantum Echoes"

The paper mentions one experiment (from October 2025) that hasn't been fully beaten by classical computers yet. However, the author is very skeptical of it.

The Issue: The experiment measured a very complex signal, but to make the numbers look good, they had to apply a massive "correction factor" (like stretching a photo to make it look bigger).
The Catch: They only checked if their math was right on a tiny scale (40 qubits), but claimed the win on a huge scale (65 qubits) where they couldn't actually check the answer.
The Verdict: The paper suggests this might just be a physics experiment measuring the machine's own noise, dressed up as a computing victory.

What This Means for the Future

The paper makes two main points about what comes next:

1. The Only Way Out is "Fault Tolerance"
The "loop" can only be broken if we build a machine that can fix its own mistakes.

Current State (NISQ): The machine is noisy. It can only do simple things.
Future State (Fault Tolerant): The machine has "error correction." It can pump out the noise (like a refrigerator pumping out heat).
The Reality Check: The paper says we need to stop pretending the current noisy machines are the final product. The real "quantum advantage" will only happen when we build the fault-tolerant machines, which is what the original math from 1996 predicted we needed all along.

2. We Need to Stop Confusing "Engineering" with "Advantage"
The paper argues that the last eight years have been great for engineering (we built better chips, better wires, better controls). But we have been bad at science (we kept claiming we solved problems we didn't actually solve).

The Trap: Press releases have mixed up "We built a machine that works" with "We built a machine that is smarter than a supercomputer."
The Fix: We need to admit that the current machines are just diagnostic tools. They are good for testing physics, but they haven't actually beaten classical computers at anything yet.

The Conclusion: The "Perpetual Motion" Machine

The author ends with a sharp metaphor:

"Quantum advantage on noisy hardware is the perpetual motion machine of theoretical physics: each new prototype runs for a few months, attracts admiring write-ups, and then a theorist notices it is plugged into the wall."

Translation: The new quantum computers look like they are running on their own power, but they are actually just plugged into the power of classical computers (because classical computers can simulate them).

The paper warns that if we don't switch our focus to building truly "fault-tolerant" machines (the ones that can fix their own errors), we will just repeat this eight-year loop for another decade, spending money on demonstrations that classical computers can already do.

Technical Summary: "The NISQ Trap: Eight Years of Demonstrations the Hardware Was Built to Lose"

Problem Statement
The paper addresses a persistent pattern in the Noisy Intermediate-Scale Quantum (NISQ) era (2018–2026): flagship demonstrations of "quantum advantage" are consistently followed, within eighteen months, by either classical reproduction or a theoretical proof of simulability. The central problem is the identification of a structural constraint that forces NISQ hardware to operate exclusively in regions of circuit-space that are simultaneously the only regions accessible to current noisy devices and the regions most amenable to efficient classical simulation. The paper argues that the NISQ program, conceived as an interim strategy while fault tolerance remains out of reach, has inadvertently created a "closed loop" where hardware constraints and classical algorithmic capabilities converge, rendering genuine quantum advantage impossible within the current regime.

Methodology
The paper employs a retrospective synthesis of theoretical results published between 2024 and April 2026. Rather than presenting new experimental data, the author aggregates six distinct theoretical findings to map the boundaries of the NISQ regime:

Noise-Induced Depth Reduction: Utilizing results by Mele, Angrisani, et al. [3], the paper analyzes how local noise models cause quantum circuits to converge to logarithmic effective depth ( $O(\log n)$ ), erasing early computation layers.
Algebraic Compression: Drawing on work regarding fermionic dynamics [2] and linear optics [7, 8, 9], the paper examines how structured input states (e.g., tensor products of disjoint pairs) admit exact compression into determinants or Pfaffians.
Geometric and Markovian Constraints: Incorporating findings by Nelson, Rajakumar, Gullans [4] and others [5, 6], the paper considers how geometric locality and approximate Markovianity further restrict circuit complexity.
Comparative Analysis: The author cross-references these theoretical bounds against specific hardware demonstrations (e.g., IBM's Heisenberg chain experiments [10, 11], Google's Willow random-circuit sampling [13], and the "Quantum Echoes" experiment [14]) to demonstrate that the hardware's operational requirements (shallow depth, structured states) align perfectly with the conditions required for classical simulability.

Key Contributions

The "Closed Loop" Thesis: The paper formalizes the observation that the "two faces" of the NISQ constraint are identical: the hardware properties that allow a circuit to run with sufficient fidelity (low effective depth, strong algebraic structure, geometric locality) are the exact same properties that allow classical algorithms to compress and simulate the circuit efficiently.
Historical Diagnosis: The work re-characterizes the 2018–2026 NISQ era not as a period of progressive discovery, but as a "closed loop" where demonstrations were inevitably drawn from a region of circuit-space that classical methods were always destined to conquer.
Re-evaluation of "Advantage" Claims: The paper provides a critical framework for distinguishing between "competent diagnostic engineering" (running a circuit and matching a known answer) and "computational advantage" (solving a problem beyond classical reach). It argues that the press cycle has conflated these two, leading to funding for results that do not advance the field's ultimate goal.
Clarification of the "Exception": The paper scrutinizes the single demonstration (Google's "Quantum Echoes" [14]) that has not yet been closed by a theorem. It highlights that this experiment relies on error-model projections, global rescaling factors, and a scoring metric that removes variance, suggesting its claim to advantage rests on assumptions unavailable at scale (65 qubits).

Results

Systematic De-quantization: The paper documents that every major NISQ demonstration, with the noted exception of the "Quantum Echoes" experiment, has been either classically reproduced or proven simulable within 18 months of announcement.
The Identity of Constraints: The analysis confirms that the "structured/shallow" region of circuit-space is the only region NISQ hardware can access. Since this region is also the only region where classical methods (MPS, Pfaffian compression, tensor networks) are tractable, the perimeter of what NISQ can deliver is mathematically closed.
The Role of Noise: The results indicate that noise does not merely degrade performance; it fundamentally alters the computational class of the device, forcing it into regimes where classical simulation is efficient.
Future Hardware Projections: The paper analyzes upcoming Department of Energy (DOE) targets for 2028. It posits that without a breakthrough in active error correction (entropy pumping), future devices will fall into one of three categories: true fault-tolerant computers (which clear the NISQ loop), larger NISQ devices (still subject to the closed loop), or physics experiments (measuring their own dynamics) mislabeled as computational advantage.

Significance and Claims
The paper claims that the burden of proof has shifted entirely to the defenders of NISQ. After eight years and over thirty announcements, the empirical pattern suggests that "quantum advantage on noisy hardware" is a "perpetual motion machine" that cannot function because the hardware is "plugged into the wall" of classical simulability.

The significance of the work lies in its assertion that the exit from the NISQ loop is not found in better classical algorithms or slightly larger NISQ devices, but strictly in fault tolerance. The paper argues that the threshold theorems of 1996 correctly identified the necessary conditions (active error correction to pump entropy) for genuine advantage, and that the NISQ era was a historical retreat that has now reached its logical conclusion. The paper concludes that unless a demonstration can escape the current simulability theorems by operating in a regime requiring coherence budgets beyond current hardware capabilities, the NISQ program has exhausted its capacity to produce computational advantage.