## Does the Threshold Theorem Actually Guarantee Fault-Tolerant Quantum Computing is Possible?

According to Indiana University physicist Amit Hagar, the answer is: not without measuring four resource costs the original derivation left out. In a new analysis published as of August 2026, Hagar proposes that the feasibility of [fault-tolerant quantum computing](https://quantumintel.tech/glossary/fault-tolerant-quantum-computing) should be expressed as a single, measurable quantity — watts per decade of suppressed logical error — and argues that existing experimental data can already begin to test this. The paper identifies two specific measurements, which Hagar contends could be achievable this year on existing hardware, that would empirically settle a debate he characterizes as "a border dispute in which each side maps the territory in its own grid of coordinates."

The core claim is not that fault tolerance is impossible. It is that decades of theoretical confidence in the Threshold Theorem have rested on a resource accounting that quietly assigned zero cost to four critical physical processes: calibrating drifting devices, running decoders during correction cycles, maintaining [coherence time](https://quantumintel.tech/glossary/coherence-time), and flushing entropy with fresh ancillas. Hagar argues those omissions are not minor corrections — they are the crux of whether the theorem's promises translate to real machines.

---

## What Hagar's "Objective Probability" Framework Actually Claims

The philosophical machinery underpinning this argument dates to 1993 in Hagar's own account, with a formalized version co-developed with Giuseppe Sergioli in 2011. The central proposition is that the probability of a quantum state is not a statement of subjective belief or ignorance — it is a physical quantity: the energy-over-time required to create that state, relative to the resources available.

This is a direct challenge to the Bayesian or epistemic interpretations that dominate much of quantum foundations discourse, but the practical payoff Hagar is targeting is narrower: if probability is objective and resource-tied, then the [error threshold](https://quantumintel.tech/glossary/error-threshold) at the heart of the Threshold Theorem becomes a statement about whether target [logical qubit](https://quantumintel.tech/glossary/logical-qubit) states are realizable within a physical resource budget — not merely a statement about abstract noise rates falling below a mathematical bound.

The framing matters enormously for hardware developers. Under the conventional reading, a team that pushes physical error rates below threshold has, in principle, demonstrated a path to arbitrarily reliable computation. Under Hagar's reading, they have demonstrated only that one input to the resource equation is favorable. The other four costs — calibration drift, decoding overhead, coherence maintenance, ancilla entropy flushing — must also be measured and factored in.

---

## The Four Omitted Costs: Why They Matter Now

Each of the four costs Hagar identifies maps directly onto engineering challenges that contemporary hardware teams are actively contending with, even if they are not framing them in his terms.

**Calibration of drifting devices** is an acute problem for superconducting transmon systems, where qubit frequencies shift over timescales of hours. IBM Quantum and Google Quantum AI both run continuous recalibration infrastructure, but the energy and time overhead of that infrastructure is rarely counted against QEC cycle budgets.

**Decoding during correction cycles** is the bottleneck that companies like [Riverlane](https://quantumintel.tech/companies/riverlane) and [Quantum Machines](https://quantumintel.tech/companies/quantum-machines) are explicitly trying to solve with real-time classical decoder hardware. The classical compute cost of syndrome decoding at scale is non-trivial — and grows with code distance.

**Maintaining coherence** — keeping T1 and T2 long enough across the full correction cycle — is where trapped-ion and neutral-atom platforms often claim an advantage over superconducting qubits, though at the cost of slower gate operations.

**Flushing entropy with fresh ancillas** requires a steady supply of high-fidelity ancilla qubits reset to ground state. At scale, this is a throughput and thermodynamic constraint, not merely an engineering inconvenience.

Hagar's argument is that these costs, when properly aggregated, should produce a measurable energy signature. If the energy-per-logical-operation curve scales favorably as codes grow, the theorem's promise is real. If it scales unfavorably or plateaus, the theoretical guarantee dissolves into a resource wall.

---

## The Empirical Proposal: Two Measurements This Year

The most actionable — and most scrutinizable — element of Hagar's paper is the claim that two specific measurements on existing hardware could resolve the debate. The source text does not detail the exact experimental configurations of these two measurements, so their precise nature warrants scrutiny when the full paper is reviewed. What Hagar does assert is that the published record already contains initial data points, and that distinguishing between extrapolated and directly measured fault-tolerance data is critical.

This last point carries weight. Much of the optimism in current FTQC roadmaps — including public projections from major hardware vendors — relies on extrapolating small-code performance to large-code regimes. Hagar is explicitly calling for a shift away from that methodology toward directly measured resource curves.

For enterprise buyers and investors evaluating FTQC timelines, this distinction is not academic. Extrapolated roadmaps and measured roadmaps carry very different risk profiles.

---

## What This Means for the Industry

Hagar's framework does not invalidate any existing hardware approach. It proposes a new ruler. If the quantum computing industry adopts something like a "watts per decade of suppressed logical error" metric — even informally — it would create pressure for vendors to report resource overhead alongside gate fidelity and qubit count. That would be a meaningful shift in how hardware progress is benchmarked.

The NISQ-era metrics (quantum volume, CLOPS, two-qubit gate fidelity) were themselves contested additions to the benchmarking vocabulary before they became standard references. A resource-normalized fault-tolerance metric occupying a similar role in the FTQC era is not implausible — provided the underlying theory survives experimental contact.

The more immediate industry implication is for QEC software and decoder companies. If calibration, decoding, and ancilla overhead are formally recognized as components of the fault-tolerance budget, the value proposition of efficient decoder architectures and low-overhead QEC codes strengthens considerably.

---

## Skeptical Assessment

Several aspects of this analysis warrant caution before treating it as settled.

First, the source text summarizes a theoretical paper; the full derivation and the specific two proposed measurements are not available for direct evaluation here. The strength of the empirical claim depends entirely on whether those measurements are practically executable and whether their results are unambiguous — neither of which can be assessed from the summary alone.

Second, Hagar's objective probability interpretation of quantum mechanics has been in development since at least 2011 by his own account, and has not — to the extent the source reflects — been widely adopted in the quantum foundations community. A framework that has not yet reshaped foundations discourse faces a meaningful adoption barrier before it reshapes engineering benchmarks.

Third, the four omitted costs Hagar identifies are real and known. But the field's leading QEC teams are not unaware of them. The claim that these costs were "assigned no cost" in the original Threshold Theorem derivation is a statement about mathematical formalism, not about what practitioners ignore.

That said, the core proposal — express FTQC feasibility as a measurable resource rate, and test it empirically rather than by extrapolation — is a useful discipline regardless of whether the full philosophical framework is accepted.

---

## Key Takeaways

- Indiana University's Amit Hagar proposes that fault-tolerant quantum computing feasibility should be expressed as watts per decade of suppressed logical error, not merely as physical error rates below threshold.
- The Threshold Theorem's original derivation omitted four resource costs: device calibration, syndrome decoding, coherence maintenance, and ancilla entropy flushing — Hagar argues this omission has shaped decades of debate.
- Hagar claims two measurements on existing hardware, potentially achievable in 2026, could empirically test the theorem's core assumptions.
- The framework rests on an "objective probability" interpretation developed with Giuseppe Sergioli from 2011 onward, treating quantum state probability as a physical resource quantity.
- The practical implication for industry: vendor roadmaps built on extrapolated performance data carry different risk than those built on directly measured resource curves.
- QEC decoder companies and low-overhead code developers stand to benefit if resource overhead becomes a standard component of FTQC benchmarking.

---

## Frequently Asked Questions

**What is the Threshold Theorem in quantum computing?**
The Threshold Theorem states that if the physical error rate of quantum gates falls below a certain threshold value, error-correcting codes can suppress logical errors to arbitrarily low levels. It is the theoretical foundation for the belief that large-scale, reliable quantum computation is achievable. Hagar's paper challenges not the math of the theorem but its resource accounting.

**What does "watts per decade of suppressed logical error" mean?**
It is a proposed metric, introduced in Hagar's analysis, that expresses how much energy expenditure is required to suppress logical error rates by one order of magnitude. If this quantity grows prohibitively as code size increases, it would signal a physical resource wall for fault-tolerant quantum computing even when error rates are technically below threshold.

**What are the four resource costs the original Threshold Theorem omitted?**
According to Hagar, the original derivation assigned zero cost to: (1) calibrating drifting physical devices, (2) running syndrome decoders during error-correction cycles, (3) maintaining qubit coherence throughout those cycles, and (4) resetting ancilla qubits to flush entropy. He argues properly counting these costs changes the feasibility calculation.

**Why does the distinction between extrapolated and measured fault-tolerance data matter?**
Most current FTQC roadmaps project future performance by extrapolating results from small, few-qubit error-correction experiments to the large code distances needed for practical computation. Directly measured resource costs at those scales do not yet exist. Hagar argues this distinction is critical because extrapolation can conceal unfavorable resource scaling that only appears at larger sizes.

**Does this paper argue fault-tolerant quantum computing is impossible?**
No. Hagar explicitly frames the question as an empirical one — the viability of FTQC is "settled only when a functioning machine demonstrates the predicted resource curve." The paper proposes a framework for measuring feasibility, not a proof of infeasibility.