Quantum computing (QC) is the utilization of the principles of quantum mechanics to perform calculations and solve problems. Using the unique phenomena of individual subatomic particles as compute elements, quantum computers have the potential to quickly solve problems that are impossible to calculate in time frames that are useful to humans.
Unlike so-called classical computers, which utilize transistors to process the smallest units of data, known as bits, quantum computers utilize quantum bits (qubits), which are subatomic particles, such as electrons and photons, to represent data. Unlike traditional bits, qubits can exist in multiple states, which is known as superposition. This means that instead of a qubit only representing a binary state of either zero or one, a qubit can be zero and one at the same time, which enables multiple calculations to be carried out at once.
Therefore, a quantum system is able to look at every potential solution to a problem simultaneously and not just the single “best” solution, but can also return thousands of close alternatives as well, all within a second. This presents significant advantages over a traditional computer, which must sequentially explore the potential solutions to a mathematical problem, taking significantly more time.
In essence, a simile highlighting the power of a quantum computer would involve a person being able to read and process every book in the library simultaneously, instead of having to read through each book one at a time, to find a piece of information that answers a single question.
The real power with quantum computers is the ability to increase compute power by adding additional qubits to the system, which provides an exponential speed increase every time a new qubit is added. In the example above, this would mean that each book could be read even faster.
However, quantum processors are not destined to replace the processors in personal computers or smartphones. A quantum computer only has a small overlap with the types of tasks and problems that can be solved by classical machines, because the technology is designed to address the types of problems that would not even be attempted on a classical machine, due to the immense processing power and time it would require.
[This article is from research firm Tractica’s report on quantum computing for enterprises. View full report details.]
Indeed, classical machines are still going to be better, faster and far more economical at solving most problems for the foreseeable future. Quantum computers are seen as best-suited to addressing optimization problems. These types of problems are considered extremely hard, because they are focused on the analysis and manipulation of multiple data points that must be considered and processed simultaneously.
A simple example of an optimization problem is the classic conundrum of organizing the seating chart at a round table for a dinner party. Each time a person is added, the number of possible arrangements is increased exponentially, because the result needs to account for each possible person occupying each possible seat. So, as additional people are added, it drastically increases the number of possible seating arrangements; for a table that seats 10 people, there are 3.6 million ways to arrange them.
Using a classical computer to process this type of problem would take an inordinate amount of processing power and time so as to not be practical, and optimization questions that are more complex, such as identifying all the possible molecular combinations needed to create a particular drug, or finding the best possible traffic routes within a city, would far outstrip the capabilities of a classical computer. The challenge lies in that every time a new variable is added, the entire model must be recalculated, thereby putting an immense strain on a classical machine.
The market for QC is being driven largely by the desire to solve these optimization problems that simply would take too long, or too much compute power, to be solved by classical computers. There are several factors driving this desire for more power and speed.
The influx of massive amounts of data. Both consumer-generated and business-to-business (B2B)-related data have exploded, in both volume and frequency, and today’s computers simply are unable to process this information in a reasonable amount of time. Moreover, enterprises are developing systems that are reliant on more so-called complex data, which require more powerful analytical tools and processing than is currently available.
Security concerns. There is a strong desire to devise new technology that can secure systems from hackers, many of which are expected to utilize QC technologies once it becomes available.
Increasing use of simulation and modeling. The use and demand to conduct digital simulations of real-world processes has increased significantly, largely as a way to reduce design, development and testing costs and time scales. Simulations are data- and processor-intensive and can greatly benefit from the exponential power of QC. These models and simulations are even being extended to modeling real-world processes, such as crowd prediction and control, fleet optimization, and logistics.
Despite the promise of quantum computers, there are significant barriers to overcome, both technological and non-technological in nature. While these are not insurmountable challenges, they will keep the market from generating significant revenue throughout much of the forecast period.
At present, there are several specific technological challenges that must be addressed over the coming years as QC matures. The first is that, as of the writing of this report, quantum computers have not been able to demonstrate quantum supremacy, which is defined as the ability to solve problems that classical computers practically cannot. This is a major barrier to adoption; quantum computers are difficult to build and control, and if they cannot provide an advantage over classical machines, enterprises will not adopt them.
Quantum computers are also prone to errors, as qubits used to calculate the result of a problem can be influenced by magnetic interference, noise from the surrounding qubits, or can simply fall out of a state of entanglement, which means the qubit can no longer process a calculation as a qubit. Correcting these errors is challenging, because current solutions require that a large number of qubits be grouped together to form so-called logical bits, to handle error correction, and that diverts power away from the actual qubits that need to do the work.
Because physical qubits are so unstable, they need the additional qubits to ensure error correction and fault tolerance. For semiconducting quantum machines, the number of physical qubits required to create a logical qubit can be as high as 3,000 to 1. Other researchers have posited that this ratio is far too low, instead citing estimates of 10,000 qubits to 1 logical qubit, or higher.
The ratio of physical qubits to logical qubits is extremely important in developing a quantum machine that can scale to address complex, real-world problems, and may constrain development until scientists are able to reduce this ratio or develop another form of error correction.
To achieve quantum supremacy, machines will likely need to be fitted with anywhere from 75 to 150 physical qubits, as the computational power grows exponentially with the addition of each individual qubit. While it is possible that researchers may develop a machine that is able to demonstrate quantum supremacy, developing a quantum machine that can actually address real-world, enterprise-grade problems will likely require qubit counts of more than 600 to address basic problems, and qubit counts in the thousands likely will be required to handle complex problems.
But it is not just technical challenges that are confronting the QC market. Relatively speaking, there are few quantum physicists who truly understand how to write quantum algorithms that will be useful to enterprises, and many of the top minds have already been recruited to work for the largest vendors and enterprises.
Further, quantum algorithms can be run on quantum simulators, but these simulators are only effective at mimicking quantum machines with about 50 qubits. Complex, real-world work (such as developing composite material simulations, chemistry simulations, or large-scale optimization tasks) simply cannot be simulated, and must be conducted on an actual quantum machine with hundreds or thousands of qubits (which does not exist). That is why it is extremely challenging for companies to devise future quantum algorithms that will actually yield meaningful results.
A key issue not often discussed is the specter of potential regulatory activity. While quantum computers are essentially a catalyst—specific algorithms must be written to solve specific problems—there is always the likelihood that privacy and data-protection advocates will want to ensure that specific policies and procedures are in place that extend to data that is used in quantum calculations or systems.
Further, because quantum computers will have the capability to ingest and analyze a wide variety of public and private data points very quickly, there is the distinct possibility that data that may not be tracked or used today could get swept up in a quantum-enabled system. Because of the exponential increase in speed and capacity to handle massive amounts of data, QC-enhanced systems are likely to attract additional scrutiny from privacy experts.
Quantum Computing Market Structure
The enterprise quantum computing market is largely focused on research and development (R&D) and identifying which use cases may be suitable for quantum computers. Vendors providing professional services, quantum algorithm development expertise and integration services are dominating the commercial activity in the market and are likely to continue to do so over the next several years.
As of mid-2018, there are quantum computers that are being used by enterprises, though largely these machines are focused on theoretical, rather than practical applications. The vendors developing quantum computers have a few primary goals, which include increasing the compute power of quantum machines by increasing the number of qubits in use; reducing noise (interference) and error rates to create a machine that can be used to address problems with more accuracy and reliability; developing an ecosystem of tools and platforms to allow enterprises to harness the power of quantum computers in the near term; and creating industry, use-case and process-specific quantum algorithms that can be patented and distributed, thereby creating a viable business model for QC applications.
Tractica projects that over the next several years, it is likely that quantum computers and classical computers will be used in tandem to solve problems. Much of the utilization of quantum computers in the enterprise will be as an accelerator to handle the “hard” problems, with a classical computer used on both the front end to program and control the quantum machine and on the back end to compile and integrate results with existing computing systems.
The QC market is expected to grow strongly over the next 8 years, with the bulk of revenue coming from the regions of North America, Europe and Asia Pacific. These regions are home to the vendors that are developing the hardware, software and services required to fill out the QC ecosystem, as well as the large enterprises that are trying to find suitable use cases where QC will be beneficial. Underlying these two segments of the market are the academic and other research institutes, which are sources of both expertise and future workers to the commercial companies operating in the market.
Tractica projects that total QC market revenue will reach $2.2 billion annually by 2025, up from $39.2 million in 2017. North America will lead the world with $718.3 million in revenue by 2025 but will be followed closely by Europe ($695.8 million) and Asia Pacific ($650.9 million).
Not surprisingly, Latin America and the Middle East & Africa will each post 2025 revenue figures below $115 million in 2025. However, each region is projected to grow strongly throughout the forecast period, posting 2017 to 2025 compound annual growth rates (CAGRs) of more than 62%.
From an industry perspective, Tractica projects that a relatively small number of industries will be actively spending and generating revenue from QC within the forecast period. Generally speaking, the industries that will be investing in quantum technology will feature problems that dovetail nicely with the types of solutions that quantum computers can provide, largely focused around optimization, simulation, design and chemistry); and will have use cases that generate significant revenue streams to support the initial and running costs of using quantum technology.
Tractica projects that there will be 15 industry groups that will generate at least $80 million in cumulative revenue related to QC between 2017 and 2025. Five of these industries—life sciences ($718.3 million); aerospace ($647.9 million); oil, gas and mining ($627.3 million); agriculture ($617.6 million); and automotive ($616.5 million)—will dominate the forecast period.
At present, there is very little actual product that can be purchased, and even by the end of the forecast period, it is highly unlikely that most Fortune 500 companies will be lining up to purchase a quantum computer. As such, much of the revenue in the market will be generated by services, including general consulting services that are provided by vendors, professional services firms and third-party industry consultants. Cloud-based access to quantum computers, which Tractica believes will be the dominant way that enterprises will use quantum machines in the near term, will also contribute to services revenue.
The market for QC is clearly on the upswing. However, many of the actual applications and use cases mentioned in this report are likely to be deployed in a test phase; there is simply too much noise and instability inherent in the quantum computers of today (and during the next five years) to rely on them for actual work.
However, the industries and use cases mentioned in this report, along with revenue forecasts, should provide directional insight as to where enterprises believe there may be benefits down the road. While QC may be in its infancy now, a decade will elapse very quickly, and no enterprise or vendor will want to be caught without having a clear map to guide them.
View details about the full Tractica report.