No investor focused on the Diversified Industrials sector would expect to come near quantum physics. Could that be about to change? At last month’s Electrical Product Group conference, a leading industrial company surprised its audience by announcing that it is ‘incubating quantum computing’, adding: ‘We do expect revenue […] from this effort late this year or next year [and] when it takes off, it's really going to change the vector of who we are and bring tremendous value to customers to solve some very unique problems that can't be solved with computing that's available today.’
This ought to be explored. Unfortunately, reading ‘Quantum Physics for Dummies’ and not understanding a word was not a particularly gratifying experience. After some further research, I hope to be able to summarize some basic concepts for the uninitiated.
Quantum physics deals with the subatomic world where the classical laws of physics do not automatically apply, something which left more than one scientist, including Albert Einstein, utterly bewildered. According to classical physics, particles are particles and waves are waves. Particles and waves are simply two distinct states. In quantum physics, waves can act as particles and vice versa, which is called the wave-particle duality. This happens to open a completely new world with many applications, including computing.
Traditional computers are essentially smart calculators which can store and process data and work on their own by executing programs. A critical component is the transistor where information is stored as bits. There can be tens of billions of transistors packed on a microprocessor nowadays. Over the years, they have become as small as they can be. The next miniaturization stages would lead to the subatomic world and its strange laws.
In conventional computers, long sequences of bits can be used to store information ready to be processed, with each bit being a 0 or a 1 depending upon whether the transistor is switched on or off. In quantum physics, the bits become quantum bits or ‘qubits’ which, unlike binary bits, can be in multiple states, i.e. a zero, a one, both zero and one, at the same time. Each qubit can thus contain a lot more information than a bit, which allows for various solutions to be tested in parallel. Consequently, quantum computing can be millions of times faster than traditional computing for certain problems (simple example available here).
The technology is nascent, and its technical challenges are immense, starting with the manufacturing of qubits and their storage, followed by the need to control them precisely under cryogenic temperature. Furthermore, quantum computers remain inherently unstable and prone to errors. Finally writing quantum algorithms is said to be a complex discipline. Quantum computers are already a real thing, though, as evidenced for example by the brand-new IBM Q System One. It will just take time before their potential is exploitable from a commercial point of view.
Quantum Machine Learnings
Quantum computers are demonstrably faster when a complex problem can be described easily and its solution can be checked as easily. This would apply to modelling chemical reaction and, closer to home, data analytics and machine learning which is currently suffering from a lack of computing power. Indeed, quantum computers are said to be particularly efficient at solving optimization problems, i.e. when the best possible outcome must be found given a particular set of circumstances. Quantum machine learning is attracting the attention of high-profile institutions in the public and private sectors, including M.I.T.
An industrial company may well have found a golden nugget in quantum computing. More generally, I would expect quantum computing to stay quite distant from the day-to-day vocabulary in Diversified Industrials for now. Many things remain to be developed around ‘digital solutions’, including when it comes to commercially sensible business models, before any industry participant gets distracted by the ambition to solve larger-scale optimization problems with ‘quantum solutions.’
That being said, some experts have expressed surprise at the fast pace of development in quantum computing over the last year. They suggest that the evolution of the underlying technology be monitored carefully. A close contact with the leading tech companies and institutions in the space may well be already warranted.