Quantum Litmus — The Weekly Reality Check for Quantum Computing
Independent, evidence-first analysis of what changed in quantum computing this week, why it matters, and what the evidence does not yet show.
Quantinuum puts error correction into the developer workflow, while Apollo remains a 2029 promise
Nexus can encode supported programs, test them in emulation and run supported workloads on Helios; the fully integrated fault-tolerant platform is still a roadmap.
This week’s top signal
Quantinuum is moving error-correction choices into the normal developer workflow, but the complete platform depends on future hardware.
In plain English
Nexus (Quantinuum’s software platform for building, testing and running quantum programs) can automatically add error correction when a program uses instructions and protection methods the platform supports.
Some programs may need to be rewritten, and the resulting calculation still has to fit on the machine. Developers can test a program in a software simulation before running it on Helios (Quantinuum’s current quantum computer, with 98 physical qubits and a demonstrated milestone of 48 logical qubits). Physical qubits are the individual hardware units. Logical qubits use groups of physical qubits to protect quantum information from errors.
Why this matters: Nexus takes over some of the specialist work of setting up error correction, so developers can spend more time on the problem they want to solve. In a separate October quantum memory test on Helios, Quantinuum showed that increasing the protection reduced errors. Its best reported error rate in that test was about five per million operations. That rate does not describe the performance of a full program using 48 logical qubits.
Much of the software capability arrived in September. The November announcement sets out how Nexus will evolve alongside Sol (the next quantum computer, targeting about 100 logical qubits in 2027) and Apollo (the computer planned for 2029, targeting hundreds of logical qubits and designed to run a broad range of programs with built-in error correction). The Sol and Apollo figures are targets.
Quantinuum reports that 215 organizations use Nexus and that job submissions in the 12 months through August 2026 were about ten times the number in the previous 12 months. That shows more activity on the platform. The announcement does not show a customer getting a useful result from an error-corrected calculation or say what a full run costs.
Our call is MONITOR. Nexus makes it easier to try error-corrected computing on Helios today. The next test is whether a customer can use it to solve a real problem repeatedly and show how accuracy, runtime and total cost compare with the best practical classical method.
Some programs may need to be rewritten, and the resulting calculation still has to fit on the machine. Developers can test a program in a software simulation before running it on Helios (Quantinuum’s current quantum computer, with 98 physical qubits and a demonstrated milestone of 48 logical qubits). Physical qubits are the individual hardware units. Logical qubits use groups of physical qubits to protect quantum information from errors.
Why this matters: Nexus takes over some of the specialist work of setting up error correction, so developers can spend more time on the problem they want to solve. In a separate October quantum memory test on Helios, Quantinuum showed that increasing the protection reduced errors. Its best reported error rate in that test was about five per million operations. That rate does not describe the performance of a full program using 48 logical qubits.
Much of the software capability arrived in September. The November announcement sets out how Nexus will evolve alongside Sol (the next quantum computer, targeting about 100 logical qubits in 2027) and Apollo (the computer planned for 2029, targeting hundreds of logical qubits and designed to run a broad range of programs with built-in error correction). The Sol and Apollo figures are targets.
Quantinuum reports that 215 organizations use Nexus and that job submissions in the 12 months through August 2026 were about ten times the number in the previous 12 months. That shows more activity on the platform. The announcement does not show a customer getting a useful result from an error-corrected calculation or say what a full run costs.
Our call is MONITOR. Nexus makes it easier to try error-corrected computing on Helios today. The next test is whether a customer can use it to solve a real problem repeatedly and show how accuracy, runtime and total cost compare with the best practical classical method.
Quantum Litmus assessmentMONITOR
Commercial Readiness Outlook — Industry
Early
Estimated broad enterprise window: 2030–2033
This week: Software & Algorithms ↑; Commercial Adoption ↑; Ecosystem Readiness ↑. Hardware Readiness →; Validated Advantage →; Economic Viability →. Readiness window 2030–2033 → unchanged.
Quantum Litmus provides independent editorial analysis, not investment advice.
Holdings disclosure: Holdings confirmed 5 October 2026: the editor personally holds shares in IONQ, IQMX, NVDA, PSQL, QBTS, QUBT, RGTI, and XNDU.TO. For this edition, the editor also discloses a personal holding in Quantinuum (QNT), the company covered in the lead story. Backloop Technology holds none. These holdings did not determine selection or the readiness judgment. About this report.