Five years after IBM made its first five-qubit quantum processor available for users to access over the cloud, the company is now showing off the first quantum computer that it has physically built outside of its New York-based data centers.
All the way across the Atlantic, scientists from Germany's Fraunhofer Institute have just unveiled the IBM Quantum System One – the country's first superconducting quantum computer that Big Blue was contracted to build especially for the organization.
The device, which contains one of IBM's 27-qubit Falcon processors, came online a few weeks ago and has already been made available to Fraunhofer's scientists and some of the institute's partners. German academics and organizations outside of Fraunhofer will, from now on, be welcome to arrange monthly contracts to use the computer too for research, education and training purposes.
Fraunhofer's partnership with IBM was signed last year, marking the start of a global expansion for Big Blue's quantum hardware. The company released the Quantum System One in 2019, pitching it as the world's first commercial quantum computer; but until now, users have only accessed the device over the cloud, by connecting to IBM's Quantum Computation Center located in Poughkeepsie, New York.
SEE: Building the bionic brain (free PDF) (TechRepublic)
Physically bringing the hardware to a new location for the first time was never going to be easy – and the global COVID-19 pandemic only added some extra hurdles. Typically, explains Bob Sutor, chief quantum exponent at IBM, the company would've shipped some key parts and a team of in-house specialists to Germany to assemble the quantum computer, but the pandemic meant that this time, everything had to be done remotely.
IBM's engineers had to rely on NASA-inspired methods of remote assembly. "How do you train people that are thousands of miles away, when you can't just run up to them and say: 'Do this'?" Sutor tells ZDNet. "We had to train local teams remotely and work with them remotely to assemble everything and get this machine running. We developed new techniques to actually put these systems around the world without travelling there. And it worked."
To train German engineers from the local IBM development lab, Sutor's team put together a virtual course in quantum assembly. From installing the computer's refrigeration system to manipulating the Falcon processor, no detail was left out and the device successfully launched in line with the original schedule.
For Fraunhofer, this means that the institute and its partners will now have access to a leading-edge quantum computer built exclusively for German organizations, instead of relying on cloud access to US-based systems.
Since the partnership was announced, the institute has been busy investigating potential applications of quantum computing and designing quantum algorithms that might show an advantage over computations carried out with classical computing.
This is because quantum computing is nascent, and despite the huge potential that researchers are anticipating, much of the technology's promise is still theoretical. Existing quantum processors like IBM's Falcon come with too few qubits and too high an error-rate to resolve large-scale problems that are relevant to businesses. The research effort, therefore, consists of spotting the use-cases that might be suited to the technology once the hardware is ready.
"For users, they need to get in now, they need to understand what quantum computers are, what they're useful for and what are viable approaches using quantum computers that will get them an advantage over using classical computing," says Sutor.
At Fraunhofer, researchers have been looking at a variety of applications ranging from portfolio optimization in finance to logistics planning for manufacturers, through error correction protocols that could improve critical infrastructure and molecular simulation to push chemistry and materials discovery.
Working in partnership with the German Aerospace Center, for example, the institute has been conducting research to find out if quantum algorithms could simulate electro-chemical processes within energy storage system – which, in turn, could help design batteries and fuel cells with better performance and more energy density.
For Annkatrin Sommer, research coordinator at Fraunhofer, the choice of IBM as a quantum partner was a no-brainer. "We really wanted to go for cutting-edge technology where you have the ability to start developing algorithms as fast as possible," she tells ZDNet.
IBM's offer in quantum computing has some significant strengths. Since the release of its first cloud-based quantum processor, the company now has made over 20 Quantum System One machines available, which are accessed by more than 145 organizations around the world. Two billion quantum circuits are established daily with the cloud processors, and IBM is on track to break a trillion circuits before the end of the summer.
The Falcon processors used in the Quantum System One are 27 qubits, but the company is working in parallel on a chip called Hummingbird, which has 65 qubits. Big Blue recently published a quantum hardware roadmap in which it pledged to achieve over 1,000 qubits by 2023 – enough to start seeing the early results of quantum computing. Ultimately, IBM is aiming to produce a million-qubit quantum system.
"If I were to throw out a toy system and say: 'Here you go, play, I don't know if it'll ever get better' – no one would care," says Sutor. "People need confidence that the machines and the software and apps on them will reasonably quickly be able to do work better than just classical computers."
For an institute like Fraunhofer, the rapid scaling of quantum technologies that IBM is promising is appealing. And the German organization is not alone in placing its bets on Big Blue. This year will also see an IBM Quantum System One installed in Japan as part of a partnership with the University of Tokyo; and back in the US, the Cleveland Clinic has just placed a $500 million order for IBM to build quantum hardware on-premises.
But despite IBM's credentials, Fraunhofer's research team is also keen to stress that it is too early to tell which approach – or approaches – to quantum computing will show results first. The industry is expanding fast, and with new companies jumping on the quantum bandwagon every so often, it is hard to differentiate between hype and reality.
This is why, in addition to investing in IBM's superconducting qubits, Fraunhofer is also investigating the use of different approaches like ion traps or diamond.
"Currently, it's not clear which technology will be the best," says Sommer, "and we will probably have different technologies working in parallel for different use cases. It makes sense to start projects with different approaches and after some time, measure how far you got and if you reached your goals. Then, you decide with which technology you should proceed."
It remains that Germany's shiny new Quantum System One puts the country in a favorable position to compete in what is increasingly shaping up to become a global race to lead in quantum computing.
The German government has already launched a €2 billion ($2.4 billion) funding program for the promotion of quantum technologies in the country, which comes in addition to the European Commission's €1 billion ($1.20 billion) quantum flagship.
Meanwhile, in the US, a $1.2 billion budget was allocated to the National Quantum Initiative Act in 2018. And China, for its part, has made no secret of its ambition to become a leading quantum superpower.
The UK government has also invested a total £1 billion ($1.37 billion) in a National Quantum Technologies Programme. In the next few years, the country is hoping to follow Germany's lead and launch its very first commercial quantum computer, which will be built by California-based company Rigetti Computing.