10 Things Driving Qubit Advancement Over the Next 18 Months
1 Error correction that improves as systems grow
Quantum error correction protects information by encoding it across multiple physical qubits. The critical milestone is demonstrating that increasing this protection consistently reduces logical errors.
Google’s Willow research established below-threshold error suppression in a quantum memory experiment. Over the next 18 months, watch for researchers extending that behaviour across larger systems and more demanding operations. Reliable computation requires protection throughout a calculation. [1]
2 Better two qubit gates
Two-qubit gates enable the interactions that quantum algorithms depend on. Their accuracy has a direct bearing on how much computation a processor can perform before errors overwhelm the result.
A 2026 paper describing Quantinuum’s 98-qubit Helios processor reported approximately 99.92% two-qubit gate fidelity. The next challenge is maintaining high accuracy as processors expand and more operations run in parallel. Consistent performance across a device will matter as much as its best individual gate. [2]
3 Error correction with a smaller hardware bill
Protecting quantum information consumes hardware. Reducing that overhead could change the size, complexity and cost of future machines.
IBM is pursuing quantum low-density parity-check codes, or qLDPC codes, as part of its fault-tolerance programme. The attraction is more efficient encoding; the engineering challenge includes delivering the connectivity and decoding those codes require. Watch for hardware demonstrations that test these requirements together. [3]
4 Neutral atoms combining scale with control
Neutral-atom research has demonstrated arrays containing 6,100 highly coherent atomic qubits. That is an impressive achievement in trapping and control, although it does not mean thousands of fault-tolerant computational qubits are available. [4]
The next stage is bringing large arrays together with reliable entangling operations, repeated error correction and logical computation. Progress across that combination would strengthen the case for neutral atoms as a scalable computing architecture.
5 Silicon qubits moving further into industrial manufacturing
Silicon spin qubits offer a potential route to using established semiconductor manufacturing infrastructure. Research has demonstrated industry-compatible spin-qubit unit cells with fidelities exceeding 99%. [5]
The question now is how consistently high-quality devices can be manufactured, connected and controlled. Over the next 18 months, fabrication yield and repeatability deserve close attention alongside laboratory performance. A manufacturing process becomes valuable when it produces working devices reliably.
6 Photonics bringing more components onto chips
Photonic quantum computing depends on generating, manipulating and detecting light with exceptionally low loss. Integrating those functions into manufacturable platforms could help reduce the complexity of assembling larger systems.
Research has demonstrated integrated silicon-photonic building blocks for these tasks, including chip-to-chip connections. Some reported performance figures are conditional on photon detection and exclude loss, so the whole-system picture remains essential. Watch for improvements that combine component quality with lower losses across the architecture. [6]
7 Cat qubits and other hardware efficient encodings
Another route to better quantum computers is to change how information is encoded in the hardware.
Bosonic approaches, including cat qubits, aim to make certain errors easier to suppress or correct. Researchers have demonstrated a logical memory combining cat qubits with an additional error-correcting code. The next tests are broader logical operations and evidence that the approach reduces the total resources required for reliable computation. [7]
8 Faster classical decoding
Quantum error correction relies on classical computing to interpret measurements and identify errors. That work must keep pace with the quantum processor.
IBM’s Relay-BP research targets decoding for qLDPC codes, illustrating how classical algorithms can influence the feasibility of quantum hardware. Over the next 18 months, decoding speed, accuracy and computing requirements should feature more prominently in assessments of processor progress. A strong error-correcting code needs a practical way to operate it. [8]
9 Modular systems and better connections
Scaling quantum computing also means connecting components while preserving performance. Packaging, wiring, cooling and interconnects all affect whether a promising processor can become part of a larger machine.
IBM’s announcement that it connected its first modular cryogenic systems in August 2026 highlights this infrastructure challenge. Watch for evidence that modular approaches can support expansion without introducing unacceptable noise, complexity or operating costs. [9]
10 Independent scrutiny of what progress actually delivers
Better benchmarking can help direct engineering effort towards the limitations that matter most.
DARPA’s Quantum Benchmarking Initiative is examining whether different approaches can achieve utility-scale operation by 2033, defined as computational value exceeding cost. Its evaluation spans the components, systems and algorithms needed to build a useful machine. My expectation is that this scrutiny will sharpen discussion of credible milestones and expose assumptions that headline qubit counts leave unanswered. [10]
What this means for enterprise leaders
Across these ten developments, the common question is whether improvements work together. More qubits, better gates and more efficient error correction each matter; their value ultimately depends on the performance of the complete system.
For the Quantum.Tech community, the next 18 months offer an opportunity to ask more precise questions about hardware progress: what has been demonstrated, what remains an engineering target, and which advances would make a relevant enterprise workload achievable?
Those answers will give us a much clearer measure of how far qubits have advanced.
Sources
[1] Quantum error correction below the surface code threshold
https://www.nature.com/articles/s41586-024-08449-y
[2] A 98-qubit trapped-ion quantum computer with all-to-all connectivity
https://www.nature.com/articles/s41586-026-10676-4
[3] IBM lays out clear path to fault-tolerant quantum computing
https://www.ibm.com/quantum/blog/large-scale-ftqc
[4] A tweezer array with 6,100 highly coherent atomic qubits
https://www.nature.com/articles/s41586-025-09641-4
[5] Industry-compatible silicon spin-qubit unit cells exceeding 99% fidelity
https://www.nature.com/articles/s41586-025-09531-9
[6] A manufacturable platform for photonic quantum computing
https://doi.org/10.1038/s41586-025-08820-7
[7] Hardware-efficient quantum error correction via concatenated bosonic qubits
https://www.nature.com/articles/s41586-025-08642-7
[8] Introducing Relay-BP
https://www.ibm.com/quantum/blog/relay-bp-error-correction-decoder
[9] IBM Connects Its First Modular Cryogenic Systems in Milestone Toward Fault-Tolerant Quantum Computing
https://newsroom.ibm.com/2026-08-19-ibm-connects-its-first-modular-cryogenic-systems-in-milestone-toward-fault-tolerant-quantum-computing
[10] DARPA Quantum Benchmarking Initiative and evaluation process
https://www.darpa.mil/research/programs/quantum-benchmarking-initiative