Chalmers University models 100-qubit chips to predict their behaviour
Chalmers researchers introduced a GPU‑accelerated method to model 100‑qubit superconducting processors and predict their dynamics.

Modeling 100‑Qubit Processors with Continuous Unitary Transformations
Researchers at Chalmers University of Technology have demonstrated a new workflow for simulating large superconducting qubit arrays. By applying continuous unitary transformations (CUTs) they derive effective Hamiltonians for a 96‑transmon array with tunable couplers. The computational cost scales polynomially with qubit number, allowing a single‑GPU run to finish in under an hour for each design point.
From Effective Hamiltonians to Predictive Power
The CUT‑derived Hamiltonians capture key hardware metrics such as:
- Effective qubit‑qubit couplings
- State‑dependent ZZ interactions
- Spatial locality of eigenstates
Combining these models with semiclassical dissipative‑time‑evolution‑approximation (DTWA) sampling, the team predicted the quench dynamics of a hypothetical 100‑qubit, 180‑coupler processor. The simulated Hilbert‑space dimension (~10^29) is far beyond exact classical simulation. Validation against exact dynamics of smaller arrays showed agreement within sampling noise, confirming that the method reproduces relaxation, thermalization, and transport phenomena.
Practical Implications for Chip Design
Because each design iteration takes at most an hour on a single GPU, engineers can rapidly explore layout variations, coupling strengths, and detuning strategies. The approach highlights observables that remain classically tractable (local densities, low‑order correlators) while clarifying its limits—bit‑string statistics used in cross‑entropy benchmarks are not captured.
Outlook and Future Work
The current study assumes closed‑system dynamics and neglects decoherence, thermal excitations, and leaked coupler modes. Extending the framework to include these effects will further bridge the gap between simulation and experiment. Nonetheless, the work provides a valuable baseline for assessing which processor properties can be reliably predicted classically, guiding the development of next‑generation quantum hardware.
--- For the full technical details see the arXiv pre‑print linked below.
Hype check
PartlyThe paper presents a new modeling technique validated on smaller arrays, but it does not experimentally verify predictions for a full 100‑qubit device.
- Peer-reviewed paper, or press release only?Press release only — Only an arXiv preprint link is provided
- Error-corrected (logical) qubits, or physical qubits only?Logical / not a qubit claim — Claim is about modeling, not a qubit‑count result
- A new milestone, or a rebrand / restatement?New — Introduces a novel polynomial‑scaling CUT method for 100‑qubit prediction
- Independent confirmation?Not yet — No external replication or confirmation mentioned
- Does the headline claim match what the source shows?Matches — Headline accurately reflects the modeling of 100‑qubit chips
An editorial assessment of the claim as reported, not a judgment of the company or researchers. Checked against: arxiv.org. How we rate hype →
Source
Read original sourceWhy we picked this
Accepted from curated QuantumUnzipped source (relevance scoring disabled).