Quantum computing is very likely the next technology destined to ride a hype wave, and in fact it may be even more likely to do that than 5G or AI, because the whole concept is based on what a famous physicist said was “spooky”. Artificial intelligence is “ponderable” but quantum stuff is imponderable, making it a much better subject for hype. IBM, who is actually the company who’s worked the longest and hardest to realize AI’s actual business value, is now stepping up to try to save quantum computing from itself.
Most people know quantum computing only as a threat. The potential for it to break all of today’s encryption algorithms in seconds, exposing everything they’re supposed to protect, has been talked about for several years. Does this remind you of how AI’s threat to destroy humanity has been a hot topic? It should. OK, this is something that should be addressed, but first and foremost any promising technology has to realize its promise or it’s unlikely to get far enough to pose a threat to anyone. Quantum chips are the critical element of the future of the technology, because if you can’t make the cost manageable, next to nothing will justify its use, and to make a quantum chip successful you have to sell it to more than the hacking community. In any case, what quantum computers can break, the experts tell me it can render unbreakable. So we need to move from threat to promise, which is what IBM’s report aims to do.
The document was produced by IBM’s Institute for Business Value, looking at what the report calls “the innovators and risk takers—those who seized frontier opportunities and are now advancing out front.” These, IBM says, are concentrated in “aerospace, genomics, material sciences and financial services.” The primary focus within each of these verticals are problems that conventional computing technology has not addressed effectively, but future-proofing computing strategy and accelerating innovation were also cited by more than half the companies involved. Note that I’ve gotten quantum computing implementation comments from only 25 enterprises, and of that group only five were actually involved in trial/test applications, none yet in production.
One thing everyone actually involved in quantum computing agrees with is that it’s not the same as “traditional” computing. Quantum computing is about algorithms, about mathematics at a fundamental level. Linear algebra, a vector/matrix math, is arguably the underpinning of quantum computing, to the point where what doesn’t fit that model doesn’t optimally fit quantum as a use case, and arguably doesn’t even belong there. This is central to understanding how companies trying to identify quantum use cases, like those IBM cites, are working to accomplish their goals.
The key point to all the cited applications, IMHO, is that they’re “planning and analysis” more than “production and operations”. Right now, quantum computers aren’t practical purchases for most enterprises, so they buy time on systems for their applications. Material and chemical analysis and simulations used to assess techniques or systems seems to be the main missions. In these areas, quantum solutions are explored to enhance traditional computing models. Pangenomics, meaning population-scale analysis of genetic factors, is cited as an example of a field where there really are no models based on current technology, so they’re a proving ground for greenfield quantum solutions.
Vanguard, the financial/investment company, offers an example of quantum applications that are likely to be more readily adopted across verticals. Their work has focused on optimizing portfolio performance and modeling risk to assess downside potential. This is an approach that obviously many business processes could benefit from, by applying common methodologies to different statistical bases. Detecting illegal behavior like money laundering, one mission being explored, is a technique that could be applied to many different kinds of fraud/theft detection. Bond portfolio analysis techniques could be applied to optimizing inventory, shelf space, and even production relationship to sales patterns.
Simulation, though, seems to be the early-adopter brass ring for quantum computing. Almost any complex system can be optimized through simulation. In the design/deployment stage, this gets the “base” state working at its best, and this mission is suitable for quantum service users. As quantum chips improve, and the price per qbit comes down, more and more “operations” missions could evolve from the base design mission. Networks, utility grids, transportation systems, warehousing, and so forth could be optimized closer and closer to real-time with improved price/performance, and the techniques could all evolve from simulation applications used in that baseline planning mission.
Simulation can also guide R&D, which is where a lot of medical/health-care interest is focused. Screening for diseases and looking for a cure are both areas where real-world trials are essential, but simulation using quantum computing could identify the most likely fruitful approaches, reducing both time and cost to a final, proven, approach. The director of the quantum initiative for a university notes that getting a drug to market today is typically about 15 years, and so a ten or twenty percent reduction in that through quantum computing simulation and analysis could “change lives”.
Today, most quantum simulation is linked to traditional computing technology for real-world application. Simulation, then, essentially defines an approach that is implemented using familiar IT tools, including AI. One cited pioneer notes that quantum computing won’t replace AI but will “sit above it” in the sense of framing the application architecture in novel ways. Every quantum maven says that traditional computing in some form has to envelop the quantum-algorithm-and-linear-algebra applications, just to handle linkage with the real (and business) world.
There are obviously commonalities between quantum computing adoption and AI adoption. In both cases, my own enterprise contacts say that the primary barrier is internal skills. There is also a tendency for early applications to rely on a service rather than on self-hosting, and there’s an expectation that costs will have to decline considerably before large-scale deployment of actual infrastructure can be expected.
One interesting thing here, though, is that both enterprises I chat with and those cited by IBM in the report seem to believe in the notion of an “ecosystem” for quantum computing. Governments have been funding or heading initiatives on this; the IBM report cites one example, and IBM was awarded a billion dollars in USDoC funds to build a quantum computing chip foundry. IBM says it’s adding its own funds to this, and has signed more than $1.1 billion in contracts for quantum computing trials, tests, and even some actual production.
When will quantum computing see more actual applications? Right now, my enterprise contacts are even in trials at levels below statistical significance, and only three tell me that they have done any real production work. That doesn’t mean that no such work has been done; another ten indicated that they knew of companies who had actually leased time on quantum computer systems for real work. Healthcare and materials/chemicals seem to be the active areas today, but absent a determined broad-based survey effort, I don’t think we can get any reliable data on actual quantum applications. Enterprises think that there will be “some” production quantum-as-a-service adoption this year, and “significant” adoption by 2028. They think that self-hosting is likely to come along in the second half of 2028 or early 2029.
I think that while more people and companies will use AI than will use quantum computing, at least in the next ten or fifteen years, quantum computing may be more important to businesses, and indirectly to people, in the long run.
