Quantum computers cannot simply grow their way to usefulness. The central obstacle to large-scale, fault-tolerant machines is not qubit count but a set of interconnected engineering problems spanning qubit quality, on-chip wiring, and the infrastructure that keeps everything cold and controlled. A review of scaling challenges published by researchers on the arXiv preprint server warns that significant unresolved issues in hardware, fabrication, and system architecture "could seriously undermine the arrival of utility-scale quantum computers for the foreseeable future." Understanding why requires looking at three distinct architectural layers — and at how a failure at any one of them constrains the others.
The Quantum Scaling Imperative: Beyond Qubit Count
Headline qubit numbers are the industry's favorite metric, but they measure the least meaningful quantity. As a survey of quantum computing scalability in the MDPI journal Quantum puts it, scaling "involves not only increasing the number of qubits but also enhancing their quality, managing error rates, ensuring efficient interconnectivity, and developing suitable software and control mechanisms." An IEEE-published review of hardware development makes the same point from the experimental side: today's quantum computers "can only manipulate a limited number of qubits for brief periods before succumbing to noise-induced disruptions," which "severely curtails their ability to execute complex computations." A machine with ten thousand noisy, poorly connected qubits is less useful than one with a few hundred stable ones. The real scaling problem is therefore architectural: every added qubit raises demands on fidelity, wiring, and cooling simultaneously.
Qubit-Level Challenges: Fidelity and Control
At the bottom of the stack sits the qubit itself. Quantum information is fragile; noise destroys it within short coherence windows, and every operation on a physical qubit carries some probability of error. The arXiv review argues that the road to scale could be paved by "adopting existing semiconductor technology to build much higher-quality qubits" — a telling choice of words, since it implies current qubits are not yet high-quality enough to scale. Better fabrication is the identified lever: the MDPI survey notes that improved fabrication techniques are being developed to enhance both coherence and scalability together. The reason is straightforward.
Error correction schemes need physical qubits whose error rates sit below a threshold; if individual qubits are too noisy, adding more of them adds more noise than the system can correct. Qubit quality is thus not a separate problem from scale — it is the precondition for it.
Chip-Level Challenges: The Interconnectivity Bottleneck
Even perfect qubits are useless if they cannot talk to each other and to their control systems. The MDPI survey identifies "ensuring efficient interconnectivity between qubits" as a significant scaling hurdle, alongside managing error rates and control mechanisms. The physical form of this problem is mundane and brutal: every qubit needs control and readout wiring, and as qubit counts grow, the wiring, signal routing, and miniaturization demands grow with them. The IEEE review calls the "relentless march towards miniaturization" a significant hurdle in its own right.
A chip that must route thousands of control lines into a cryogenic environment faces a congestion problem that has no analogue in classical processor design, because the signals must reach devices operating at millikelvin temperatures without injecting heat or noise. This is why interconnectivity, not qubit fabrication alone, is often described as the binding constraint on chip-scale growth.
Architectural Strategies for Interconnectivity: Monolithic vs. Distributed
Two architectural strategies dominate the response to the interconnectivity bottleneck. The first is monolithic integration: build one large, dense chip and solve the wiring and fabrication problems directly. The second is distributed scaling: build smaller, more manufacturable chips and network them into a larger logical system. The distributed approach trades the fabrication difficulty of a single massive chip for the engineering difficulty of high-fidelity links between modules.
IBM's announced plans illustrate the strategy in practice: the company says its Nighthawk r3 chip, designed to run 10,000 quantum operations on 520 usable qubits, carries 10 connectors for joining chips together, with two chips intended to form a single 1,040-qubit system in 2027. The modular design exists precisely because fabricating one 1,040-qubit chip would be harder than fabricating two 520-qubit chips and linking them. Neither strategy is universally superior; monolithic designs avoid link losses, while distributed designs reduce fabrication complexity and allow incremental growth.
System-Level Challenges: Infrastructure and Integration
The top layer of the stack is where quantum processors meet the classical world. The arXiv review frames the end goal as "distributed quantum computation within heterogeneous high-performance computing infrastructures" — meaning quantum processors will not stand alone but operate as components of larger classical systems. Reporting on IBM's platform plans makes the same point from the industry side: advancing quantum computers "requires continuous progress in how they interact with classical systems to mitigate errors and manage the increasing demands of complex workloads." This integration is not optional plumbing.
Error mitigation and eventually error correction require fast classical feedback, and the cryogenic infrastructure that keeps qubits near absolute zero must scale alongside everything else. The system layer is also where the arXiv review's warning lands hardest: it identifies "system engineering approaches" as one of the three pillars of a viable scaling path, alongside better qubits and distributed computation.
Quantum Hardware Scaling: A Three-Layered Model
The challenges above can be organized into a single diagnostic framework. Use it to categorize any scaling announcement you encounter: ask which layer the claimed advance actually addresses, and what new demands it creates for the other two.
| Layer | Core challenge | What the evidence shows | Knock-on constraint |
|---|---|---|---|
| Qubit-level (quality & control) | Coherence and error rates | Today's machines manipulate limited qubits briefly before noise-induced disruptions, per the IEEE review; the arXiv review says higher-quality qubits via semiconductor techniques could pave the road to scale. | Noisy qubits multiply the physical-qubit overhead needed for error correction. |
| Chip-level (interconnectivity & miniaturization) | Wiring and signal routing | The MDPI survey names efficient inter-qubit interconnectivity as a key hurdle; the IEEE review flags relentless miniaturization as a significant obstacle. | Every added qubit increases wiring and control-line congestion on the chip. |
| System-level (cryogenics & classical integration) | Infrastructure and integration | Progress requires continuous improvement in how quantum systems interact with classical systems to mitigate errors, per reporting on IBM's platform; the arXiv review calls for system engineering and distributed computation in HPC infrastructures. | Scaling any lower layer increases cooling and classical-feedback demands. |
The model's analytical value is in its coupling: a qubit-level improvement changes what chip-level wiring must support; a chip-level architecture such as multi-chip linking changes what the system layer must integrate. This is why simply increasing qubit count is not a complete solution — a claim every layer of the evidence independently supports.
Navigating the Path to Scalable Quantum Computing
To achieve utility-scale quantum computing, researchers and engineers must simultaneously address qubit quality, on-chip interconnectivity, and system-level integration challenges, rather than focusing solely on increasing qubit count. The three-layer model shows why: each layer sets the requirements for the others, and a breakthrough at one level without matching progress elsewhere yields no usable machine. The distributed, multi-chip approach — exemplified by IBM's plan to link two 520-qubit Nighthawk r3 chips into a 1,040-qubit system in 2027 — is best understood not as a qubit-count milestone but as a bet that modular architecture can sidestep fabrication limits while system-level integration catches up.
Progress in quantum computing scalability can be observed through advancements in qubit coherence times, reductions in physical qubit overheads for logical qubits, and successful demonstrations of multi-chip quantum system integration. Those three indicators, not headline qubit totals, are the milestones worth watching.





