The state of quantum

Quantum chips are sparking into life. But what is this hardware going to be used for?

Forward: features are independent pieces written for Mewburn Ellis discussing and celebrating the best of innovation and exploration from the scientific and entrepreneurial worlds.

There are two camps in quantum computing. There’s the cautious let’s-not-oversell-it types, who are mindful of the Gartner hype cycle and its Trough of Disillusionment. And there are those who simply can’t help but radiate their excitement at the dawn of the quantum era.

If there’s a spokesman for the latter camp, it’s whurley. Known officially as William Hurley, and to everyone else by his coding username, whurley is one of the best-known faces in the industry. He’s the founder of Strangeworks, a quantum computing ecosystem, as well as the $125m venture capital fund Ecliptic Capital, based in Austin, Texas.

William Hurley (also known by his coding name "whurley"), founder of Strangeworks and Ecliptic Capital

‘The quantum revolution is often compared to the industrial revolution, and for good reason,’ says whurley. ‘Just as the industrial revolution dramatically changed the way we live and work, quantum computing aims to do the same on an even grander scale, potentially revolutionising every field that relies on computation, from medicine and finance to material science and pharmaceuticals.’

He’s eager to spell out just how big this revolution is going to be. ‘My interest in the future of quantum computing isn’t about how it will help us to develop new drugs and treatments for diseases, it’s about eliminating disease and debilitating conditions. It’s not about finding new chemistry for batteries; it’s about finding completely new forms of transportation we’ve yet to dream of. This is the power of quantum computing and the ridiculously exciting potential it could bring. To transform the world for the betterment of humanity and the planet.’

It’s hard not to feel a buzz at this vision. New forms of transportation? Transform the world? But quantum computers barely exist yet. We are only getting glimpses of how they perform. How can he assess the potential?

‘In all honesty, I started off saying “We write software for computers that don’t exist”,’ admits whurley. ‘However, that was 2018, and by 2020 there was already enough progress on the machines that I couldn’t say that anymore. Today we have Fortune 500 companies across multiple industries using the Strangeworks quantum computing ecosystem to drive internal innovation using quantum computers.’

Strangeworks now works with a long list of hardware partners. These include D-Wave; IonQ (maker of the Aria quantum computer which operates with 20 algorithmic qubits, which the company claims makes it the most powerful in the industry); and Xanadu. In June 2022 the Xanadu Borealis completed a statistical task in 36 microseconds, which it estimated would take Fugaku, the most powerful supercomputer at the time, 9,000 years to finish. It was a tantalising display of ‘quantum supremacy’, where a quantum computer beats a supercomputer at a given task.

The predictions of evangelists such as whurley are not mere speculation. They are increasingly grounded in real-world progress made by the quantum industry. 

What to do with unfathomable power

The emergence of quantum hardware highlights the question: What are quantum computers going to be used for? What exactly does one do with almost limitless compute power?

‘Financial markets will be greatly affected by a general-purpose quantum computer in ways that far exceed portfolio optimisation or risk management,’ says whurley. ‘Portfolio diversification, rebalancing portfolio investments, and streamlined trading settlement processes also come instantly to mind.’

But for him, the impact lies beyond merely doing standard stuff faster.

‘I believe that 2023 will be like 1963. In 1959, Jack Kilby invented the integrated circuit. Before this innovation you purchased transistors one at a time. By 1963, his creation was in bloom and kicking off the march towards all of the modern computers we have today. So I’ll ask your readers this… In 1963, how many people were thinking about artificial intelligence? Autonomous vehicles? Drones? Facebook? TikTok? My point being that 2023 will produce the most advancements in quantum computing in a single year. But still, this will be nothing more than a starting gun. The race? To use our first step away from the architectures John von Neumann created to completely reinvent the world.’

We are already seeing glimpses of this new way of thinking. QC Ware is a software consultancy based in Palo Alto, California, and, like Strangeworks, is focused on helping companies and developers harness the unique properties of quantum computing. It offers Forge, a platform that provides access to algorithms, developed by QC Ware, which run on hardware supplied by a variety of providers. Current clients include Goldman Sachs, Roche, and BMW.

‘I think we still haven’t found the killer application,’ says Yianni Gamvros, SVP of sales and marketing at QC Ware. Nevertheless, he’s able to reel off a variety of real world uses. ‘Chemical structures is a promising area,’ he says. ‘When we want to look at how a potential drug will react with a protein in a given situation, we currently use classical computers and get a fuzzy picture of that interaction. With quantum computing we will get a high-definition picture. We will also be able to run hundreds or thousands of simulations, which currently takes years.’

Yianni Gamvros, SVP of sales and marketing at QC Ware

Investment banks are also a natural fit. ‘Financial institutions want to price risky assets, such as derivatives, in a fraction of the current time. Today, when they want to run a simulation of market conditions, they need to run computers overnight. Classical computers take that long. The promise of quantum computers is to run simulations throughout the day and get results as soon as market conditions change.’

The pharmaceutical industry, too, offers a variety of examples. QC Ware worked with Roche on medical image recognition. The project deployed machine learning applications for biomedical image analysis and ran them on quantum machine learning and neural networks. The goal was to identify disease in retinal images faster and more accurately than before. Tests were run on IBM machines. The results illustrate the power and the challenge of quantum computing; whilst some results showed promise, noise introduced by the hardware limited the accuracy. The paper concluded: ‘It is quite clear that the larger circuits we used are pushing the boundaries of what the current quantum machines can perform reliably and one would need better and more robust hardware to increase confidence on how quantum machine learning techniques can help in image classification tasks.’

Finding the signal in the noise

The Roche study highlights a serious aspect of quantum chips – namely, the noise problem. Quantum machines are fragile. The quantum vehicle in question, be it an electron, caesium atom, or ytterbium atom, are highly susceptible to disruption from various sources; radiation from wi-fi or mobile phones can interfere, as can the Earth’s magnetic field, or in-chip factors. For that reason, we are in the NISQ era of quantum machines – the Noisy Intermediate-Scale Quantum era. It’s not a flattering term. The chips are not yet able to implement continuous error correction. The chips run, but the results are noisy.

The problem is so profound, simply adding more qubits to a chip may be counter-productive, owing to excess noise cancelling out the gains in speed. Slower, more noise-resistant chips, may offer superior results.

Software is also a factor in this debate. By writing the software in an optimised way, it’s possible to run algorithms on slower, more stable quantum chips and get better results. This way, software can bring the era of quantum supremacy forward by years.

‘You can actually run the same code on a quantum computer as on a classical computer,’ explains QC Ware’s Gamvros. ‘But you’ll get the same results and, as a result, no improvements. If you rewrite it to take advantage of the special capabilities of quantum computers, that’s when interesting things happen. That’s what we do as a company.’

As the software is optimised, so the performance soars. ‘When you see a press release by a hardware manufacturer, it’s normally announcing they’ve upped the number of qubits in their chip, from say 50 to 100 qubits. They’ll also talk about the qubit quality and how much error their machine introduces into the calculation. Well, when we make an announcement on the software side, it can be that we can run a calculation that used to require, say, 10,000 qubits and now requires 1,000. So the trends are moving in two directions: chips and software. When they meet, that’s when you’ll get applications that work really well.’

Remarkably, the shortcomings of current quantum machines aren’t holding back research into software. It is possible to run quantum applications without the hardware being ready.

‘Our field has produced several simulators that work on classical computers, which can simulate how a quantum computer would behave up to a certain number of qubits. We can run tests and see what is going to scale. The manufacturers build simulators, as do the software companies. Us too. It’s a low-cost sandbox environment to make sure you understand the theory and for you to test algorithms on.’

Almost there

The timeline for the quantum era is clearer than ever. For example, Rigetti, a quantum chipmaker, expects its 1,000 qubit model to be ready in late 2025 , and a 4,000 qubit system around 2027, a little later than initially forecast. Professor John Morton, co-founder of Quantum Motion, told Mewburn Ellis he’s expecting production-ready chips with multiple 100-qubit modules running in parallel on the same chip, on a ‘five-year horizon’. IBM recently published an updated timeline to its 4,000-qubit Kookaburra quantum chip with parallel processing (see box out).

For whurley, we are already one foot in the quantum era. ‘Quantum computers are becoming more and more prevalent today,’ he says. ‘At Strangeworks we already have dozens of early machines available. However, timelines for when quantum computers will become an everyday thing are still unclear. The way we like to look at it is that pessimists think it’s 10 years away, optimists think it’s 3 years away, realists are preparing today.’

And which companies need to prepare for the quantum revolution? ‘Every. Single. One,’ says whurley. ‘It’s a starting gun in a race to a future that’s yet to be written. Where will we go from here? Think about how much our lives have changed in the 63 years since the invention of the integrated circuit. Now think of 2023 as the breakout year for quantum computing and close your eyes and imagine how much our lives will change in the next 63 years.’

The full potential of quantum computers

Andrew Fearnside, Senior Associate and Patent Attorney, Mewburn Ellis comments:

 The World Economic Forum recently reported that private and public investment in quantum technologies is estimated to have reached $35.5 billion in total by 2022. Much of the initial investments in quantum computing were made by governments who realised that leadership in this technology has now become a matter of geopolitical importance. Since then, around two-thirds of all funding in quantum technologies since 2018 has reportedly been in the form of private investment. This pivot from public towards private investment is necessary for sustained growth of a healthy commercial landscape for quantum computing and quantum technologies more generally. Today’s quantum computers operate on the scale of hundreds of qubits, and the aim of the industry at large is to reach the million-qubit scale. Then, we can expect to see the full potential of quantum computers realised. Private and public research is making tremendous progress in this direction, so there is good reason to be excited about the state of quantum.

Written by Charles Orton-Jones