Quantum Computers Trapped in Perpetual Five-Year Promise
The Horizon That Refuses to Collapse
The question of when quantum computers will truly arrive can be approached from two directions. One is empirical: gather the data on how the machines are improving and look for a pattern. The other is philosophical: ask what a quantum computer actually is, and whether the answer is even stable. Christophe Jurczak, at investment firm Quantonation, took both routes. [1] Both led him to the same unsettling conclusion. His starting point was a conference that had nothing to do with quantum computing. Researchers there discussed a different emerging technology, and Jurczak realized he had first heard about it thirty years earlier. Back then, everyone said it was only five years away. Thirty years later, they were still saying the same thing. Deja vu. Then a worry set in: could quantum computers also be such a “perpetual five-year technology” — a PFYT? The label describes a field that keeps promising breakthroughs within half a decade, generation after generation, while the horizon retreats every time someone approaches it. The worry drove Jurczak to spend a year digging through philosophical literature. The paper that came out of that work delivers a blunt verdict: quantum computers are, right now, firmly in the PFYT category.
SENTENCE: Together with Brooke Abeles at ENS Paris-Saclay in France, he collected data about the many metrics that describe how well a quantum computer might work. [1]
CORRECTION: Together with Brooke Abeles at ENS Paris-Saclay in France, he collected data about the many metrics that describe how well a quantum computer might work. SENTENCE: Once the industry crosses a threshold for one metric, the five-year horizon does not collapse.
CORRECTION: Once the industry crosses a threshold for one metric, the five-year horizon does not collapse. SENTENCE: Take error rates: when researchers can guarantee that a single qubit makes an error less than once in a thousand times, the field does not declare victory.
CORRECTION: Take error rates: when researchers can guarantee that a single qubit makes an error less than once in a thousand times, the field does not declare victory. SENTENCE: Each milestone reveals problems that were invisible before the milestone was reached.
CORRECTION: Each milestone reveals problems that were invisible before the milestone was reached. SENTENCE: Building ENIAC was much more of an engineering challenge, he argues.
CORRECTION:
The second approach is philosophical. Jurczak found a thinker whose work seemed tailor-made for the situation: Gilbert Simondon, a philosopher of technology. Simondon’s theory of “technical individuation” says that a shifting assembly of elements becomes a coherent individual through a process that includes a whole ecosystem. That ecosystem consists of infrastructure, knowledge, and practices, maturing through iterative steps of trial and error. A collection of qubits and the devices that control and house them becomes a quantum computer only through interconnected developments in laser systems, in cryogenics, in the building of special chip foundries, in the growth of a specialized quantum workforce, and in basic research in quantum physics and computer science. The list goes on. The ecosystem is large and complex. The machine is not the only thing being built; the world around the machine is being built too.
When the Machine Defines Itself
The mechanism behind the resetting horizon is an identity problem. Forecasts of when quantum computers will arrive presuppose that we already know what a quantum computer is. Jurczak argues that this presupposition fails. “If the identity of a quantum computer is still being settled as the machine develops, that presupposition fails,” he writes. SENTENCE: The horizon will keep moving until it does not.
CORRECTION: That is an unusual situation for any industry, and it has consequences for planning. Companies draw road maps as if the destination were fixed. The destination is still moving.
Part of the difficulty lies in the diversity of approaches to building the machines. The basic components of quantum computers are called qubits, and they can be built in many different ways. Some qubits are made from extremely cold atoms. Some are made from tiny superconducting circuits, which carry electricity without resistance at very low temperatures. Some are made from particles of light. Each approach has its own champions, its own engineering challenges, and its own timeline. Then there is the question of what to measure. How many qubits does the machine have? How well connected are they? How reliably can they be used for different parts of a computation? How quickly can they be controlled? The list of relevant metrics is long. Perfecting one or even a few of these metrics typically does not guarantee that the others will become excellent too. A machine that excels on one measure can be ordinary on another.
This is why researchers and executives in the field speak with a certain mix of excitement and anxiety. They celebrate the fact that all the components of a truly powerful machine have been proven to work on their own. Then they try to keep at bay a worry about what will happen when those components are all put together. The contrast with classical computing is instructive. Researchers sometimes compare today’s quantum computers to ENIAC, the world’s first programmable general-purpose electronic computer, a room-sized digital behemoth. But Jurczak says that comparison is misleading. Building ENIAC was much more of an engineering challenge, he argues. SENTENCE: Quantum computing is still so close to scientific discovery that it cannot be easily wrangled, easily scaled, or easily accelerated by the usual tricks of the economy.
CORRECTION: Quantum computing is still so close to scientific discovery that it cannot be easily wrangled, easily scaled, or easily accelerated by the usual tricks of the economy. That is why the absence of a fully useful quantum computer is not a failure of prediction, and not a failure of the market. The delay reflects the groundbreaking, paradigm-shifting nature of the attempt to corral the denizens of the quantum world into doing computational work for us. “Don’t despair,” Jurczak says, putting a positive spin on it. “This is highly nonlinear, it takes time.”
What Comes After Fault Tolerance
None of this means the field is standing still. Over the past several years, quantum computers have made measurable progress. Several firms have presented road maps for the future of their devices, including goals for overall size and the performance of components, and then they met those goals. In 2026, the machines are much closer to being useful devices than they were just a handful of years ago. A chemist or a material scientist can access a quantum computer through the cloud right now and use it as a scientific tool. Quantum computers have not yet changed the world by aiding the discovery of a groundbreaking drug or a battery redesign. But the distance from scientific tool to world-changing device is narrowing. Jurczak himself recently ran an experiment remotely on a quantum computer made from extremely cold atoms. The machine was housed in a lab in Canada. He was on his couch in Texas. The experiment worked. Yet this did not settle the question of what quantum computing is, he says.
The industry currently has its sights on a goal called fault tolerance. A fault-tolerant quantum computer is one that is so good at catching its own errors that it can reliably run computations of unprecedented complexity. Many firms are projecting that they will reach this goal by the end of the decade. They have put the date on their road maps and are working backward from it. But will the arrival of fault tolerance answer the question of what a quantum computer is? Jurczak’s framework is not predictive, he says, but it does begin to unravel that question. Almost certainly, the horizon will shift again. Not all the machines will be fully individuated. The ecosystem will have matured, but it will not have finished maturing.
Jurczak is excited about the possibility of a genuinely heterogeneous quantum-computing system. No single qubit design will win, he predicts. Different designs will, through their own process of individuation, find applications they are well suited for. The question of what a quantum computer is will not be answered by just one device, but by many, in many different forms. This carries a lesson for the people who build and use these machines. “The people who will build and use these machines should have more of an open mind about what a quantum computer is as a technical individual and how it concretises and how it’s going to be used as well,” Jurczak says. Quantum computers will not remain perpetual five-year technologies forever. The constantly resetting horizon, he insists, represents possibility. Firms still need strategies for the next five years, Jurczak concedes. “We all want that and that’s needed, but think beyond.” The five-year plan is a tool, not a prophecy. The machine is still becoming itself. The horizon will keep moving until it does not.
Sources
1. Quantonation