Few markets as small as the quantum information sciences market generate as much lively discussion. Hyperion Research pegged the worldwide quantum market at $848 million for 2023 and expects it to reach ~$1.5 billion in 2026, according to its annual quantum computing (QC) market update presented at the Q2B Silicon Valley conference held in Santa Clara this week.
Bob Sorensen, Hyperion Research’s chief quantum analyst who presented the market update, told HPCwire, “I think that a positive, if not robust, market projection is justified. The QC ecosystem is becoming more sophisticated and granular with increased opportunities from QC processor suppliers, targeted classical control system vendors, QC systems integrators, software orchestration firms, and a growing base of sector-specific QC applications developers. All that adds up to a more finely-tuned QC solution well suited to the particular requirements for any potential QC end user, making quantum computing a more attractive compute option going forward.”
It is sobering that there are so many uncertainties remaining in QC writ large, ranging from figuring out what will be the quantum “transistor” (e.g. preferred qubit modality), to implementing needed error correction and scaling up system size, and ultimately building a library of quantum algorithms and applications to fulfill quantum computing’s tantalizing promise.
What’s not uncertain is the global race among quantum believers, including governments, companies, and academia all chasing the goal. For example, the U.S. is expected to reauthorize the National Quantum Initiative Act for a second five years sometime this month. Consider the major international organizations that assisted Hyperion in conducting its most recent QC market survey:
- Quantum Economic Development Consortium (QED-C)
- European Quantum Industry Consortium (QuIC)
- Quantum Industry Canada
- Japan Quantum STrategic industry Alliance for Revolution (Q-STAR)
- Australian Quantum Alliance (AQA)
- Korea Quantum Industry Association
Having missed out on the semiconductor revolution – the underpinning of the modern electronics industry – many regions (small and large) are jumping in so as not to miss the quantum revolution. For the moment, the quantum computing ecosystem retains its roughly bi-modal nature, with a few giants and very many smaller companies jostling for sway.
As shown below, the make-up of Hyperion survey is a broad reflection on the QC market. Twenty-four respondent companies had total (not just quantum) revenues of more than $10 billion and 39 had less than $15 million. Only two companies reported more than $50 million in quantum revenue. The long (irregular) tail of 66 companies with under $1 million is more broadly representative of the aspiring QC market.
A relative newcomer to the Hyperion outlook is a more bullish attitude towards deployment of on-premise quantum systems. Both IBM and D-Wave have deployed their systems at user facilities in the past, but no others. Just this year, both QuEra (neutral atom-based qubits) and IonQ (trapped ion qubits) have announced plans to offer on-premise systems, and HPCwire has talked with at least one quantum industry veteran who’s planning a quantum integrator business model to assist in deploying and integrating quantum systems into datacenters.
Sorensen said, “The positive future of QC installations on-premises is clear, at least to me. Despite many of the current advantages to QC access via cloud (pay as you go options, the ability to switch qubit modalities and vendors easily, and the relatively low capex requirements during the exploratory phase) there will be an increasing interest by QC end users firms that will have any number of reasons to use an on-prem QC, including the need to protect proprietary information, speed tightly integrated hybrid quantum/classical algorithms, ensure 24/7 access to a specific machine, and likely in cases where QC usage is high, secure a lower cost set-up than a cloud access alternative.
“In addition, many HPC sites are and will be looking to bolster in-house QC expertise and having a system on site offers more opportunity to do that versus a cloud-based option. That said, issues to be ironed out include buy versus lease, especially at a time when hardware advances are happening quickly, decisions about which quantum modality, architecture, and vendor to commit to, and the ability to effectively integrate an on-premises QC into an existing classical HPC ecosystem,” he said.
In keeping with past studies, the top targeted sectors remain steady, although the FS sector dropped from the top spot. Prospective QC end-user attitudes about demand drivers are interesting in that they reflect, for example, the growing recognition that the traditional HPC hardware paradigm is stuck. All netted out, QC user budget expectations are also up.
On balance, Sorensen’s view of QC prospects is positive.
“The QC sector currently is marked by a wide range of innovation with many questions about which quantum hardware and software will eventually reign supreme,” he said. “However, a sure sign of viable technology, especially one that could drastically redefine something as far reaching and entrenched as the classical IT sector, is that exploration is taking place across the academic, government, and the vast array of commercial entities.
“All this does is ensure that every considered quantum option will have its opportunity to shine, but only if it can prove its merits. There will be a range of companies that enter the market, with some departing, some being consolidated, or some pivoting to new opportunities. But as long as the overall scope of innovation stays on an upward trajectory, future prospects for the QC sector are good.”
A new consideration is the emergence of LLMs and concern regarding what impact it will have on efforts and funds flowing into the quantum ecosystem. At the moment, the quantum community doesn’t seem overly worried. It should also be noted that there are many efforts to harness LLMs as education tools for quantum computing as well as as coding aids to enable developers to write code for quantum computers without having to master quantum specific tools. Jay Gambetta, VP IBM Quantum, told HPCwire recently, “We [think] the full power of using quantum computing will be powered by generative AI to simplify the developer experience.”
As with all things quantum computing, a measure of caution is smart – Hyperion, for example, couched its outlook as estimates rather than firm forecasts. There are still a lot moving pieces in the gradually coalescing quantum landscape puzzle.