Quantum Computing: 1 million qubits

Nothing has changed the world more than our ability to compute. A large fraction of the global population now carries up to a billion transistors in their pockets, and universal internet access is rapidly following. Every business is soon to be an information business, as I’m reminded while running in the countryside, where even farming is being transformed. Livestock are being equipped with sensors to monitor their health remotely, and a farmer’s decisions on what crop to plant when, as well as how much to irrigate and fertilise, are increasingly based not just on weather forecasts but also on complex computational models making market and economic predictions.

We are learning that computers can help us with almost everything, even in places we might not expect. Perhaps even the creation of fertiliser itself could be improved by computing. A significant percentage of the world’s energy is used to make these nitrogen-based products, yet we know a better solution is possible since microbes in the soil do the same job effortlessly.

More energy-efficient production of fertiliser is just one of a multitude of critical problems that could be solved if we could create an accurate simulation of how molecules behave. Similar problems exist across climate change, healthcare and energy, where computing could change our lives profoundly. An example that touches all these areas is designing an economically viable process to capture carbon from the atmosphere, which could have the side benefit of generating fuel for cars, heating and electricity in the form of methanol.

The fact that we can even begin to tackle these problems with computing is thanks to our science-fiction like ability to make silicon computer chips. We make billions of these every year, each one containing up to several billion transistors — the building blocks of a computer. We’ve managed to shrink these transistors over a period of decades to such an extent that we can now measure their size in terms of individual silicon atoms. The latest transistors are just tens of atoms across; thousands of them could fit across the width of a human hair.

The problem with conventional computers

Many of the problems that seem so suited to computing are still incredibly hard, and our conventional computers are not only currently inadequate but will forever remain so. In the past, we could rely on regular increases in computing power. The number of transistors within computer chips has impressively doubled every two years, a trend known as Moore’s Law. But just when we’d like to rely on it most, Moore’s Law has come to an end. In 2015, Tom Conte, president of the Institute of Electrical and Electronics Engineers’ Computer Society, stated prophetically that “Moore’s Law is reaching its limits: the doubling of transistors per unit area is slowing down . . . and is projected to end at seven nanometres circa 2020.”

These big problems that we’d really like computing to solve are so difficult that, even if Moore’s Law did continue, conventional computers could never solve them. When these problems increase in complexity, the time required to solve them doubles or even worse. Simulating a molecule is an example of such a problem: while it is possible to simulate small molecules in seconds on a laptop, the time needed to run the simulation explodes for larger, more complex molecules. Using conventional computers, exact simulation of molecules with just a few hundred atoms could take longer than the age of the universe.

A quantum solution

The good news is that a solution is at hand — a drastically different approach to computing that is profound both in terms of the fundamental laws of physics it exploits, and the transformations it will bring about in our lives, society and economy.

At their most basic level, conventional computers represent each “bit” of information — the logical zero or one — in the on-off state of a transistor. But by exercising careful control over some of the smallest constituents of our universe, quantum computers instead work with “qubits”. A standard bit can only exist in the zero or the one state, whereas a qubit can adopt a uniquely quantum superposition of the two logical states.

Any carefully controlled system obeying the laws of quantum mechanics can be used to form a qubit; popular choices are trapped ions, superconducting circuits and single particles of light, known as photons. However you choose to do it, changing how you compute at this fundamental level opens up new approaches to problems.

I first saw the revolutionary potential of quantum computing as a student, when I came across a science magazine article on the subject. By good fortune, I was engrossed in an undergraduate quantum mechanics course at the time. Quantum mechanics is the theory of physics that describes nature at its most fundamental level. My classmates and I were learning that a single particle could be in a strange “superposition” of more than one place at the same time, and that two particles could be inextricably and powerfully linked, or “entangled”.

Energised with purpose, I ended up describing this incredible idea and lending the magazine to almost everyone in my quantum mechanics class. I never got it back but I finished my degree, went on to take a PhD in quantum computing and have been working towards the goal of building a quantum computer ever since. Today, I lead a research centre of more than 100 people, focused on creating this technology.

The quantum advantage

Many problems evading conventional computers are well suited to a quantum computer — molecular simulation being a prime example. A large fraction of today’s supercomputing power is used to perform molecular and materials simulations. But these simulations are limited to small systems and imperfect approximations. Although precisely simulating the quantum mechanical behaviour of molecules is insurmountable for a conventional computer, a quantum computer is perfectly suited to represent these kinds of intrinsically “quantum” problems.

A quantum computer with several hundred logical qubits would be able to tackle the problems of modelling nitrogen fixation and designing catalysts for extracting carbon from the air. A quantum computer with about 1,000 logical qubits may help us design a room-temperature superconductor, capable of transmitting electrical power with negligible loss (we currently lose about 10 per cent of electrical power just in the transmission from power plant to consumer).

We may ultimately use this simulation power to design new pharmaceuticals, clean energy devices and polymer membranes for fuel cells. In fact, we could apply quantum computing to the design of any material for any purpose — from transport and construction to sensors and prosthetics — since those materials are ultimately made of molecules and atoms, and understanding their properties and interactions is a quantum mechanical problem. This is one of the most compelling features of quantum computing: it’s a technology that expands the way we can think, and the extent of the possible solutions we can investigate.

But the benefit of quantum computers is not limited to molecular applications. So-called quantum algorithms allow us to come up with powerful approaches to seemingly “unquantum” problems. For example, quantum algorithms can search databases faster, perform pattern matching (important in genomics and genetic engineering, for example), and even perform computer graphics operations more efficiently.

These algorithms are hard to come up with, because they require us to think in a quantum way, but as quantum technologies become more ubiquitous and we become more proficient at thinking like this, we can expect more and more to emerge. There are even quantum algorithms that can perform key elements of machine-learning tasks, which are vital for big data business analytics, and in growing areas of artificial intelligence such as self-driving cars.

When will we have one?

Given the promise of quantum computing, when will this world-changing technology emerge? Prototype systems with up to a dozen qubits have already been developed in industrial and university research labs around the world. Many people are excited by this, because when we reach around several dozen qubits, these machines should start producing results that simply cannot be replicated by a conventional supercomputer. It would be an indisputable demonstration of the otherwise unachievable power of a quantum system, dubbed “quantum supremacy” by Professor John Preskill at CalTech.

That’s a thrilling prospect and I believe we will get there within the next few years. Unfortunately, we don’t yet know how to use such a relatively simple quantum computer to calculate anything useful. To tackle useful problems, we will need many more qubits.

The catch is the need to correct the inevitable errors experienced by qubits. Because quantum computers are prone to errors in ways that conventional computers are not, it is currently understood that to create a quantum computer with 100 logical qubits, we would need a system with about a million actual qubits. By adding this redundancy, we are able to distil a perfect “logical” qubit from many imperfect ones.

That is a big price to pay, and I believe that the only way we will make quantum computers at this scale any time soon is if we exploit the incredible silicon manufacturing capabilities currently used to make computer chips. As remarkable as it may sound, the technology on your desk may be just the thing we need to build a quantum computer.

The path forward

So here is my call to action: let’s not fret over the end of Moore’s Law. Far from leaving us in the lurch, it may be the stepping stone to an even bigger computing revolution. Let’s take the crowning achievement of our last computing revolution — the silicon chip manufacturing capability we have today — and redirect it to build a quantum computer. Given the potential impact on climate change, energy generation and healthcare, our future might just depend on our ability to quantum compute.

If nothing else, it might finally mean that I can take a run through the countryside in peace.