The Qudit Advantage: System-Level Joules-per-Solution Comparison of a Qudit Architecture Against 17 Conventional Qubit Quantum Computing Platforms

In the paper's words

The JPCUB Competitive Landscape v2.0 benchmarked 17 qubit-based quantum computing platforms across a single system-level energy-efficiency metric: joules per solution (J/sol). All 17 platforms use two-level quantum systems (qubits). This paper extends the JPCUB framework to qudit architectures — d-level quantum systems — and computes a joules-per-solution estimate for a qudit platform whose error-correction model is based on p-adic stabilizer codes on Bruhat–Tits trees with hierarchical ultrametric decoding and passive error resilience.

The abstract, verbatim, and its opening sentence is the claim. Every line below is quoted from the text.

Numbers it reports

1. The prior competitive landscape comparison reported the following values for the 17 qubit platforms, reproduced here for convenience: Platform Family JPCUB Range (J/sol) Key Determinant Superconducting $0.05$–$0.71$ Gate speed ($30$–$500$ ns) Neutral-atom $0.32$–$0.62$ Rydberg-gate speed with room-temperature operation Trapped-ion $8.5$–$16.3$ Gate time ($50$–$100$ $\mu$s) dominates despite room-temperature power Annealing $0.15$–$0.42$ Problem-embedding overhead Photonic $2.1$ Photon-generation efficiency The dominant factor across all families is gate speed — the time per operation determines $t_{\text{sol}}$ more strongly than the power consumption per gate.

   from 2.1 JPCUB Framework

2. 10 mK is: $$\frac{E_{\text{erase}}(300\text{ K})}{E_{\text{erase}}(10\text{ mK})} = \frac{300}{0.01} = 3 \times 10^4$$ However, this ratio does not directly translate to a JPCUB advantage — the Landauer bound is a per-erasure floor, and the dominant energy cost for qubit platforms is NOT the erasure energy but the active cooling overhead ($P_{\text{cool}}$).

   from 3.4 Cooling Advantage

3. For $d = 2$, this is 1 bit — the least efficient choice.

   from 4.3 Consilience — The 78-Year Silo

4. In the specific case $d = 2$, this yields 1 bit per symbol — the qubit is, in information-theoretic terms, the least information-dense carrier possible.

   from 1 Introduction

5. It models a qudit platform within the JPCUB framework and compares the resulting joules-per-solution estimate against the 17 qubit platforms from the prior competitive landscape comparison.

   from 1 Introduction

6. The qudit architecture may achieve a joules-per-solution advantage of at least one order of magnitude over the best 2026 qubit platform through three compounding factors: (1) dimensional encoding density ($\log_2 d$ bits per physical carrier vs.

   from 1.2 Core Claim

How it is done

1. This is not a quantum effect — it is the same $\log_2 d$ factor that appears in Shannon's source coding theorem for a $d$-ary alphabet.

   from 2.2 Qudit Encoding Density

2. For the JPCUB model, we set $P_{\text{decode}}^{\text{qudit}} \approx 0$ as a lower bound (optimistic estimate) — actual decoder implementation energy will increase the JPCUB value above the estimate reported here.

   from 3.3 Decoder Energy Advantage

3. In dimensionless Planck units, the Landauer bound for information erasure is: $$E_{\text{erase}} \geq T \ln 2$$ where $T$ is the dimensionless temperature (in units of the Planck temperature $T_P \approx 1.417 \times 10^{32}$ K).

   from 3.4 Cooling Advantage

4. The qudit advantage is parametrized by the dimension $d$.

   from 3.7 Sensitivity Analysis — The $d^*$ Crossover

What would falsify it

1. Disconfirmation: If a physical qudit platform with $d \geq 3$, subjected to independent adversarial validation, yields a measured joules-per-solution above $0.05$ J/sol (the 2026 superconducting-qubit floor), the claimed advantage is falsified.

   from 1.2 Core Claim

2. Falsification condition: If a physical qudit platform with $d = 3$, subjected to independent adversarial validation, reports $J_{\text{CUB}} \gt 0.05$ J/sol, the qudit advantage claim is falsified.

   from 3.7 Sensitivity Analysis — The $d^*$ Crossover

3. ID Prediction Test Window Disconfirmation Condition P1 A physical qudit platform with $d = 3$ achieves $J_{\text{CUB}} \lt 0.05$ J/sol 2027–2030 Measured $J_{\text{CUB}} \gt 0.05$ J/sol under adversarial validation P2 A physical qudit platform with $d = 7$ achieves $J_{\text{CUB}} \lt 10^{-4}$ J/sol 2028–2032 Measured $J_{\text{CUB}} \gt 10^{-4}$ J/sol P3 An external group independently computes a JPCUB for a qudit platform and the result is within $10\times$ of this paper's estimate 2027–2030 No external JPCUB for any qudit platform published by 2030 All three predictions are provisional; none has been validated.

   from 5.1 Pre-Registered Predictions

Where the claim stops

1. The model carries significant uncertainties.

   from 4.2 Constraints and Caveats

2. The 300 K operation assumed in this model has not been demonstrated for any qudit platform with the coherence times required for fault-tolerant computation.

   from 4.2 Constraints and Caveats

3. The model assumes a fault-tolerant regime (logical qubits, not physical qudits).

   from 4.2 Constraints and Caveats

Where to read

1. 1 Introduction570 words · 2 min
2. 2 Background540 words · 2 min
3. 3 JPCUB Model for the Qudit Architecture957 words · 4 min
4. 4 Discussion530 words · 2 min
5. 5 Calibration Register and Frontier Questions353 words · 2 min
6. 6 Conclusion218 words · 1 min
7. Declarations189 words · 1 min