As we saw in our previous blog, Schrödinger’s Cat: Paradox to Practical Quantum Computing, utilizing just fifty “good qubits” can — in the right setting — easily outsmart the most advanced supercomputers. So, the benefits of this technology are not in doubt. The real challenge is not jumping down the rabbit hole, but ensuring you come out at the other end, in other words effectively resolving the difficulties in successfully exploiting the qualities present in superposition and entanglement. Creating superposition is relatively straightforward. The real nightmare is keeping it there. Equally, maintaining high-quality entanglement brings its own set of headaches. In practice, these difficulties fall into the four broad categories of stability, scale, processing, and software, which we will now consider in turn.
Stability – The Constant Battle Against Decoherence
Stability is the constant battle against decoherence. Decoherence is what happens when a qubit in superposition collapses into a definite classical state (either 0 or 1). At that point the original quantum state is lost forever. This is why engineers dedicate so much time to developing quantum error correction. Since they cannot reverse decoherence after the fact, they work to prevent the logical information from ever collapsing in the first place.
In classical computing, problems caused by noise — temperature fluctuations, vibrations, or electromagnetic interference — are easily fixed. You simply make several copies of each bit and let the majority vote decide the correct value. In the quantum world this straightforward process is not physically possible. The no-cloning theorem states that you cannot make an identical copy of an unknown quantum state. Any attempt to measure the qubit to copy it would immediately collapse its superposition (the “spinning coin” stops spinning).
Continuing this analogy, quantum error correction works by placing that one important coin inside a large group of carefully entangled “helper” coins. If one coin in the group starts to wobble, or stops spinning, the group can detect the error and correct it — all without ever directly “looking” at the original important coin. Thus, the quantum information stays protected in superposition.
What sounds simple in theory, however, turns out to be one of the hardest engineering problems in the entire field of quantum mechanics. Even the best error-correction codes require between 100 and 1,000 physical qubits just to create one reliable “good qubit”. In a functional quantum computer, that quickly adds up to hundreds of thousands — or even millions — of additional physical qubits.
The headache does not stop there, however. These additional physical qubits must have error rates below about 0.1 % per gate, and some even lower. If the error rate is any higher, the error-correction code itself starts generating more errors than it fixes, and so the whole system fails. Securing such demanding tolerances is still a very long way off. Indeed, scientists are only now beginning to touch that threshold — and even these are limited to small-scale experiments.
Scale – Turning One Reliable Qubit into a Practical Machine
The next challenge will almost certainly be apparent to the reader at this point, the issue of scaling. It is not just that each “good qubit” requires redundancy in the order of three or four orders of magnitude, but that each physical qubit requires individual control wiring, cooling, precise microwave or laser pulses, constant error-checking measurements, and demands on power that can only be classed as a nightmare.
The solution, once you think about it, is obvious: build better physical qubits. Thus, current targets aim to reduce this overhead from around 1,000:1 down to somewhere in the region of 10–50:1 — a dramatic improvement. The technical approach required to achieve this goal centers on securing three key improvements: create purer substrates (high-resistivity silicon or sapphire), make better Josephson junctions (the super-thin “magic sandwich” that creates the space where superposition can occur), and reduced two-level defects (one of the biggest sources of noise and decoherence) through improved materials and atomic-level surface treatments.
Processing – Precision and Coordination at Quantum Scale
Processing is our third major hurdle. Even when you have enough stable qubits, they still have to be made to do something useful — and do it with an exceptionally high level of repeatability. The same quantum gate must produce essentially identical results millions to billions of times in a row with almost no variation. Every single-qubit gate, every two-qubit entanglement operation, and every measurement must achieve fidelities of 99.9% or better (some operations need 99.99%). Timing must be synchronized to within nanoseconds across the entire machine.
At the same time, the system must run continuous error-checking measurements and apply real-time corrections — all without introducing new errors. This creates an immense coordination challenge: sophisticated microwave and laser control systems, ultra-fast classical processors sitting right next to the qubits, and flawless synchronization across hundreds of thousands (or millions) of channels.
All of this points to an exceptional level of both complexity and physical size. This is no laptop. Today’s largest quantum systems already fill entire lab rooms with giant cryostats. A truly useful, fault-tolerant quantum computer will occupy dedicated data-center-scale facilities.
The Human Orchestra – It Takes a Village of Specialists
It also underscores the range — and the sheer number — of experts required for successful planning, construction, functioning and maintenance of a quantum computing system. This is not a job for a single polymath genius — no such person exists. Rather, it demands a large, tightly coordinated team of specialists from many different fields — physicists, cryogenic engineers, microwave and laser experts, materials scientists, classical computer architects, software developers, error-correction theorists, and more — working under strong technical leadership that can bridge the gaps between disciplines. It is for this reason that many of today’s leading quantum computing companies began life as university research groups or spin-offs.
Software – Entering an Entirely Different Intellectual Universe
Software is our fourth and final hurdle. In many ways it is different in nature from the previous three. Stability, scale, and processing are formidable physical and engineering problems. Software, by contrast, requires a genuine paradigm shift if we are to grasp the nature of what we are considering. Classical programmers are trained to think in clear, deterministic, step-by-step logic. Quantum programming requires thinking in probabilities, superposition, entanglement, and interference. You are not giving the computer a sequence of commands — you are designing quantum circuits that make the right answer emerge through constructive interference while wrong answers cancel out.
This requires very different mental abilities and attributes: strong mathematical intuition (especially linear algebra and probability), an unusually high tolerance for counter-intuitive results — often cited as the greatest mental challenge — and the ability to reason comfortably in high-dimensional abstract spaces. Many seasoned software engineers describe the transition not as moving to a different programming language, but as entering an entirely different kind of intellectual space.
As a result, true quantum software talent is exceptionally rare and in high demand. These specialists are at a significant intellectual premium: companies compete aggressively for them, compensation is high, and the talent pool remains very small.
The principal challenges facing the software itself are formidable. Finding efficient quantum algorithms that deliver a meaningful advantage for real-world problems is still extremely difficult. Even when good algorithms exist, translating high-level ideas into instructions that run efficiently on today’s noisy, limited-connectivity hardware is a major bottleneck. Developers must also build software that can extract useful results from imperfect, error-prone qubits, create seamless hybrid workflows that combine quantum and classical computation, and develop debugging methods that feel more like experimental physics than conventional programming. Finally, the software must work uniformly well across wildly different quantum hardware platforms — superconducting, trapped-ion, neutral-atom, photonic, and others — which all behave very differently.
For our readers, this last point may prove crucial. The software challenge is not just writing code — it is working in an environment where today’s quantum hardware platforms are locked in a fierce competition reminiscent of the early VHS versus Betamax days. While that contest was resolved quickly, this one is likely to last much longer. We will almost certainly see an extended period of coexistence and hybrid systems. Far from being a drawback, this diversity is a strength: it keeps the ecosystem open, encouraging new entrants, fresh ideas, and counter-intuitive innovations that could ultimately accelerate progress. Palomar’s Advanced Solutions Division exists to support exactly this environment. With more than four decades of leadership in semiconductor packaging, we have powered innovations from the very first personal computers and every generation of smartphones, through high-speed data center interconnects, the IoT, atomic clocks, drone sensors, and medical implants — and we are now delivering advanced solutions for quantum computing, encryption, and communication. From university research teams to the world’s largest corporations, we partner across the full spectrum of innovators. Ready to discuss how we can help with your project? Contact a member of our team here.
----
Dr. Anthony O'Sullivan
Palomar Technologies
Senior Director of Strategic Marketing