QC Ware: Quantum computing for the masses
At first glance, Matt Johnson’s journey to CEO of QC Ware--the Palo Alto company developing accessible, hardware-agnostic quantum computing software--immediately stands out from the mathematicians, physicists and computer scientists you’d expect to find running a quantum computing company. With an impressive list of Fortune 100 customers, investment from Citigroup, Airbus and Goldman Sachs, and partnerships with NASA and the US National Science Foundation, the company has already established itself as a leader in this emerging field.
Matt is, in his words “unconventional”. After serving five years as an active duty officer in the US Air Force and a further long stint in private equity, Matt wanted to return to his roots. “I was looking to do something meaningful. I wanted to work on building something new, and I wanted to tap into my long-standing interest in aerospace technology." From 2011 to 2013, Matt spent time looking at upwards of 150 information technology and aerospace startups and projects, in search of something that appeared to be on what he calls the ‘tipping point’ between being a science project and a buildable product.
“I’m not a scientist, nor a physicist,” he admits. “So what we’ve done is build a very strong team of quantum information scientists, physicists and mathematicians. We’ve managed to build a critical mass of real leaders in the research field and it’s their deep insights, their thought leadership, their track records and their networks which are allowing us to build out from that base."
Matt is quick to deflect attention from himself and onto the team, which he is obviously very proud of. “We’ve tried very hard to create an environment in our company which offers our quantum computing team [the opportunity] to do very challenging quantum computing research and development, that specifically addresses this problem: how can we use early-generation quantum computers --the ones available now and over the next few years--to solve real-world problems? Our team is dedicated to answering the fundamental question for the quantum computing industry at this stage, which is: what are quantum computers good for?”
In exploring answers to this question, Johnson and the team at QC Ware have initially prioritized three classes of problems: simulation, optimization and machine learning. Simulation, Johnson predicts, will “hit first” - the first breakthrough application for quantum computing, modeling chemical reactions and molecular dynamics in pursuit of new drugs, chemicals and super-materials. The focus on optimization is really a focus on dragging the problems of the present into the quantum age: resolving computing bottlenecks for computationally complex problems in verticals from financial services to logistics to manufacturing.
“If you look at the largest 500 companies in the world, most of them will have a wide variety of optimization problems across various business processes,” Johnson says. “They’re currently solving these problems using ‘classical’ systems, but the vision is that quantum computing will be able to shorten the time-to-solution for a number of their ‘hard’ problems.”
Outside of its government partnerships, QC Ware’s roster of clients currently includes a number of those Fortune 500 companies, but its medium-term goal is to expand its customer base to classically trained data scientists, optimization, chemistry experts and more.
Making quantum computing easily accessible for as many users as possible is what drives QC Ware and its product strategy. Accessibility, both in terms of hardware access to different hardware platforms, and in terms of abstracting away complexity so that more classical experts can use quantum computing algorithms.
“Our target user, the target customer we go after, is everyone who has a GitHub account: every programmer, every data scientist, dev, quant, whatever,” he says. “There are about 25 or 30 million of these people. These people are what we’d call ‘classically trained’ - they’re trained to write programs that run on CPUs and GPUs, classical or traditional processors. So what we want to offer to those parties is a way to stay in their high-level language - Python, Java, C++ or whatever - but to run their job on any of the quantum computers that are available. We take care of that in our black box. The user doesn’t need to know that, but the user has the ability to target or direct his or her job onto any of those hardware platforms - that’s what we mean by accessibility.
“This is the thing that all these big companies want; they say, ‘I’ve got a computing bottleneck and if I could resolve this bottleneck and solve this problem more quickly it would be valuable to me. So we look at these problems and say, ‘OK, can I express this problem in a way that would run on a quantum computer?’ That’s how we’ve been successful in collaborating with these customers.”
Johnson is also personally involved in the development of the wider quantum computing community, as a member on the governing board of the Quantum Economic Development Consortium, a partnership between the US government, industry and academia to steer America’s quantum computing research in the most promising new directions over the coming years.
“Our job through the QED-C is to advise the government on where explicitly the $1.2 billion specified under the National Quantum Initiative legislation should get spent over the next 4-5 years. It’s just an advisory consortium, but it’s meant to take input from various quantum technology companies as to which particular research topics really need to be invested in now. That’s a nice thing for QC Ware to be involved in, because it helps us contribute to the community.
“The other thing we do that I think is really noteworthy is that we organize an annual conference called ‘Q2B’, a three-day industry symposium for quantum computing. The goal is to accelerate application discovery and application development. The conference gets big enterprises that have hard problems together with the technology heavyweights and researchers who are building quantum computing hardware and algorithms in the same room to answer the same question that drives QC Ware: ‘What are these machines really good for?’”
Over the next one-to-two years, Johnson and QC Ware will continue to develop the algorithms and software to address that question and its customers’ specific demands, while also building on the upcoming beta release of its software platform. But extending that timeline to 3-5 years, Johnson hopes to achieve the “true objective”: a quantum performance breakthrough in at least one real-world business setting, that can then be spread across industries and clients as a foundation for future problem-solving.
“Somewhere in 3-5 years what we hope will happen is to achieve a true speed up for one or more practical problems,” Johnson says. “That would be the objective in that timeframe: to be able to identify at least one real enterprise problem where you’re introducing quantum processing into the computing workflow. We want to have our cloud service set up so that the functions and features of that service are very useful to a wide range of users. So, for optimization we would want an engineer at Raytheon to use the same function call that a fund manager at PIMCO might use. Two, three, four years from now, we want our cloud service to proliferate across these users in different industries.”
Beyond QC Ware’s five-year plan, Johnson - like so many quantum tech CEOs - has an eye on machine learning and artificial intelligence. But despite his atypical background, he’s a realist: cognizant of the wilder promises being made by quantum’s more zealous promoters and direct about where AI falls on the company’s list of priorities.
“The reason we are sensitive to not hyping machine learning is that there is limited technical evidence to support the view that there are large machine learning problems that can be accelerated by quantum computing. We as QC Ware, with some of the world’s leading experts in the field, and others still need to do a lot more research on Quantum Machine Learning to support that claim. There also needs to be more hardware development, as current-generation QPU’s can not compete with the large clusters of FPGA’s and GPU’s that provide computing power for machine learning problems today. So if anyone asserts right now that there’s a wide body of information around how to use quantum computing to gain speed-ups in AI or ML, that would be a misstatement.
“Machine learning and artificial intelligence are tantalizing targets for people in the quantum computing community. We are pulling on some interesting threads and there is a growing number of people in the research community working in this domain. We are all very incentivized, because artificial intelligence and machine learning are seen as so important for the future of computing and business.
“It’s very sexy,” he says. “It’s very appealing and very important. It is less tangible, however, than optimization and chemistry - that’s for sure.”
For more information email Amit Das directly on firstname.lastname@example.org
Visit QC Ware's website: https://qcware.com/