Image courtesy of 123rf.com

Since its launch in May 2025, Europe’s VanEck Quantum Computing UCITS ETF (QNTM) has returned 13.57% value to shareholders, outperforming the 9.84% gain of the S&P 500 index in 2025. Starting with IonQ, the ETF not only holds all the quantum computing stocks but also the Big Tech companies supporting their own quantum initiatives, alongside large banks. This is not surprising as quantum-boosted algorithms stand to significantly improve risk modeling, portfolio optimization, personalized credit scoring, and cybersecurity. However, although not as exciting as AI hype, is quantum computing hype justified? In other words, what is the actual threshold for quantum computing to be practical and commercially viable?

The Biggest Problem for Quantum Computing

Quantum computing follows an inverse development path from classic computing. In the latter, the fundamentals were clear from the start – boolean logic, semiconductors, transistor physics and von Neumann architecture. This means that scalability was baked in from the start, with capacity and reliability just increasing over the years.

With quantum computing, the fundamentals themselves are not clear. That’s because its unit of computation – qubit – is inherently fragile, losing coherence in fractions of a second. To put it differently, reality itself is so rife with interference that it makes a usable quantum state vulnerable to environmental disturbances.

In practice, this means that quantum computing relies on quantum error correction (QEC). When investors hear that some new quantum system has crossed the 1,000-qubit threshold, such as IBM’s Condor, this should be taken with a grain of salt. The problem is that it requires a large number of physical qubits to form a single logical qubit, which serves as a usable unit of computation.

This range varies wildly, depending on the QEC technique used. Case in point: In mid-2024, Quantinuum announced a breakthrough on its H2 system, successfully entangling four logical qubits using quantum low-density parity-check (qLDPC) codes.

“The error correction step must be done so well that in the final calculations, you only see an error in less than one in a billion (or maybe even one in a trillion) tries. Correcting errors on a quantum computer is quite tricky, and most current error correcting schemes are quite expensive for quantum computers to run.”

Yifan Hong, Elijah Durso-Sabina, David Hayes and Andrew Lucas in their research paper

Although Quantinuum is not a publicly traded company, it is effectively owned by Honeywell International (HON). Likewise, IonQ uses Clifford Noise Reduction (CliNR) method with a ratio of three physical qubits per logical qubit. While this is considered cutting edge advancement, positioning IonQ as one of the top quantum exposures, it is still a large overhead for the purpose of scaling.

In essence, quantum computing relies first on reducing the physical error rate of the qubits, by employing a specific qubit system, and then on an error correction scheme to mitigate the physical system. Both are still in the Noisy Intermediate-Scale Quantum (NISQ) development stage, not suitable for business applications.

Quantum Approaches to Deal with Errors

Different quantum hardware setups produce different hardware noise, which algorithms (software) must then tackle. For investors, this means tracking a fundamentally sounder approach, as it would take less effort to correct errors. Moreover, for the code to work more effectively, the physical error rate would have to be under a certain threshold, typically around 0.1%.

Presently, quantum gate systems using trapped-ion qubits are more reliable. At the University of Oxford, in late 2024, one such system demonstrated control of trapped-ion qubits with microwave pulses, achieving an error rate of 1 per 6.7 million operations.

Although this 10⁻⁷-level qubit gate error is considered record-breaking quantum fidelity, it is still far off from the aforementioned goal of “less than one in a billion”. Both IBM and Google use a gate-based approach, delivering 99.5%-99.9% fidelity. Again, this percentage seems high but it is far from fault-tolerant.

Quantum annealing systems, employed by D-Wave Quantum (QBTS) in Advantage2, produce a higher physical error rate within 10⁻2 to 10-3 range, making them heavily reliant on error-correction. However, the upside is that quantum annealing has greater scaling potential due to a simpler architecture used for specific purposes – combinatorial optimization – alongside being more resilient to a specific type of noise.

On the other hand, gate-based quantum computers could be scaled as well, by reducing the need for extreme cryogenic cooling. For example, IonQ deployed an Extreme High Vacuum (XHV) approach using ion-trapped technology at room temperature, by cooling individual ions with lasers.

The Bottom Line

Investors should not expect to see a fault-tolerant quantum system until 2030, as set by IBM in June’s roadmap. Only after this milestone can true scaling efforts begin, kickstarting the real quantum race. In this interim period, it is already clear that IonQ, Quantinuum and D-Wave are making progress, in both physical hardware and software error correction.

These companies are likely to provide useful quantum computing under the existing umbrella of Big Tech’s cloud computing – Google, IBM, Microsoft – alongside Big Tech’s own advancements. Therefore, quantum computing today is less about immediate returns and more about strategic positioning.

In the end, the gap between quantum hype and reality may be wide, but the groundwork being laid in hardware fidelity and error correction is meaningful.

More By This Author: