IBM Q isn’t vaporware. It’s a project years-in-the-making that could help quantum computation reach its massive potential. The future of quantum computers may arrive sooner than you think.
When news arrived of IBM’s move to offer the first commercially available universal quantum computer last week, it was characterized as a “handoff” from IBM Research to IBM Systems. According to the company’s CTO and vice president of quantum computing, technical strategy, and systems, Scott Crowder, that’s not entirely the case.
“It’s not quite a ‘handoff,’ it’s really a partnership,” explained Crowder. “This is definitely a transition point from it being pure science, pure research, to also being engineering, and development, and commercialization.”
“But this is the ‘40s,” he continued, drawing a comparison to a defining era in the development of classical computers. “This is not the ‘60s or ‘70s, this is the 1940s. We’re still in the very early days; we’re still driving a lot of the basic fundamentals, at the same time that we’re trying to make the systems accessible to a wider group of people.”
People often wonder how quantum computation will impact their day-to-day life. By offering the first commercially available universal quantum computer, IBM hopes to find out.
Bringing quantum computers into the real world
IBM’s press release about its upcoming plans for quantum hardware was packed with potential applications for the technology, and animated quotes from company personnel and experts in the field. By contrast, it was light on details of when its quantum computers will be released. The company’s announcement pledged that universal quantum computers will be delivered to partners “over the next few years.” It also stated that aims to construct systems with around 50 qubits “in the next few years.”
The fact that only a vague timeframe has been made public might lead some to believe that IBM Q is an attempt to secure a spot at the quantum table without setting any plans in stone. However, Crowder maintains that there’s a much firm schedule behind closed doors.
“Yes, we have a rigid definition of ‘the next few years’ internally,” he laughed. “We announced that our intention is to offer systems to a select group of industry partners this year. We’ve said that we’d upgrade them to roughly 50-qubit size within the next few years.”
IBM Q is not a project conceived for good press coverage alone. In fact, Crowder specifically referred to the company’s intention to make quantum systems “real,” rather than just being “PowerPoint presentations and vaporware.”
While the specifics of the timeframe are being kept under wraps, IBM has high hopes that IBM Q can help foster the next evolution of quantum computing. And while the company isn’t ready to share a detailed schedule with the public, it has no such reservations about delineating how its quantum hardware differs from other systems on the market.
What’s in a Name?
Unless you’ve been following the development of quantum computing very closely, you may have been confused when IBM announced the first commercially available universal quantum computer last week, as D-Wave has been supplying quantum hardware to enterprise users for several years.
However, D-Wave has been criticized for the way it presents its hardware. The company promotes its products based on their qubit count — its most recent system has access to 2,000 qubits, which is far more than IBM’s aspiration of producing a 50-qubit system in the next few years. However, the two strands of hardware can’t be compared directly, as D-Wave builds quantum annealers, which can only tackle certain problems. IBM is working on universal quantum computers which, like a classical computer, can tackle many types of problems.
“I think for the folks who are deeply into this area, they understand the differences,” said Crowder. “But to the average technologist, even, it’s really confusing.”
“What we’re offering is not a fixed-function system, it’s a universal system,” he added. “But more importantly than that, the power of quantum computing that has everybody really excited is this concept that you can explore an exponential space with it.”
Crowder is referring to the concept of complete quantum entanglement, which allows qubits to occupy a wider swathe of states than the binary on/off of traditional bits. It would only take around 40 or 50 qubits in this state to create a quantum computer that can’t be simulated on a classical computer. At present, there’s some debate as to whether D-Wave’s hardware can offer these advantages over traditional systems, even at much higher qubit counts.
Even in its current form, IBM’s quantum hardware is a step beyond what’s previously been offered commercially. But to fulfil its lofty ambitions for the project, the company isn’t just looking for customers; it’s looking for partners.
Like ‘Early Access,’ but for quantum computers
IBM expects to roll out its first “early access” quantum systems to select industry partners by the end of 2017. However, this won’t be a traditional transaction in terms of purchasing hardware from a manufacturer — the company isn’t just paying lip service when it refers to early adopters as “partners.”
“For the first couple that we do, we definitely want to have partners that we will be working with very closely,” said Crowder. “We hope to find partners that are driving the ecosystem.”
IBM’s partners will benefit from this set-up, as the hardware itself will be maintained and periodically upgraded by the company. A quantum computer has strict requirements in terms of space, temperature, and other practicalities. Keeping the system under IBM’s watchful eye will make it easier for partners to concentrate on putting its capabilities to good use.
IBM already has a proven track record of delivering quantum computation via the cloud, thanks to the IBM Experience. “It’s not like you do five qubits in a lab, you kick the thing a couple of times and it works, you get your results, then you kick it again two months later to get it to work again,” said Crowder. “No, we demonstrated that we can have a real quantum computer running continuously since May .”
On the other end of the bargain, IBM hopes that these partnerships will be able to inform the future of the IBM Q initiative. There are still big questions about how and why companies will choose to implement quantum hardware, collaborative work with early adopters might offer up some answers.
This should make it clear that IBM Q is a long-term project. Partners getting in on the ground floor are signing up to explore uncharted territory alongside the company — which could help them snag first place in line when the business benefits of a quantum system become clear. IBM is thinking three moves ahead. However, the company isn’t ready to drop its classical computing endeavors just yet.
A real quantum leap
“Absolutely,” Crowder replied when we asked him whether IBM expected to see quantum and classical computers co-exist going forward. “Absolutely, for much longer than my lifetime.”
“What classical computers do well is store and process lots of data, and they do that much more efficiently than quantum computing does today, and for as far as I can see in the future,” he explained. “Theoretically, universal quantum computers can do any function that a classical computer does, but in my opinion, they’re not as good at processing lots of data. What quantum computers are good at is exploring large problem spaces.”
IBM isn’t looking at quantum computers as the next generation of classical computers. It’s an entirely new category of hardware that has its own strengths, its own weaknesses, and the potential for some very powerful applications.
Of course, it’s still immature technology. Researchers all over the world have made massive advances toward a working large-scale universal quantum computer in recent years, but less headway has been made in terms of hashing out potential applications. It’s something of a vicious cycle.
“With a roughly 20-25 qubit system, you can still simulate that quantum system on your laptop,” said Crowder. “When you start getting to 40s, and around 50, you can’t even really explore that entire possibility space on the world’s largest supercomputer. It’s an interesting discontinuity.
“You actually need to have access to a quantum system to really understand how you build algorithms and use cases for a quantum system,” he added. “You can’t simulate it on a classical system anymore.”
Setting the benchmark
At present, we’re so early in the process of building universal quantum computers that there isn’t even a standardized method of comparing one system to another. IBM hopes to solve this problem via a metric dubbed quantum volume, which considers the number of qubits, the quality of quantum operations, qubit connectivity, coherence time, parallelism, and more.
“It’s all of those things together that tell you how powerful your quantum computer is, in terms of what’s the size of the space it can explore,” said Crowder. “We’re very open to other people weighing in and perfecting that benchmark, but we’re going to need some kind of benchmark like that over time, so that people can easily get their head around the value of this quantum computer versus another quantum computer; this generation versus the next generation.”
The fact that there’s no such benchmark in place demonstrates just how much work there is to be done to transform quantum computation from promising theoretical work to a workable commercial product.
“Like I said, we’re in the ‘40s,” said Crowder. “We don’t have an industry benchmark yet to classify how powerful a quantum system is, but that’s the kind of thing that I think you’ll see rapidly developing over the next couple of years, as IBM and others keep pushing the threshold on these systems.”
IBM is laying down infrastructure that could prove to be very valuable in years to come. Between its quantum value metric, its relationships with IBM Q partners, and the IBM Experience, the company is already deep into the process of making quantum computing a key component of its business interests.
“We do believe that we’ve gotten to the point where we see a path to larger systems in the near future,” said Crowder. “I know the announcement doesn’t have specific timelines in it, but we do see that path to larger systems. That gets it to the point where it gets interesting.”
IBM has already invested a large amount of time and resources into quantum computing; IBM Q represents the next step. The company has ambitions of spearheading quantum computing as a commercial product. If it can find the right partners at this crucial stage, IBM Q could play a pivotal role in taking the technology out of research labs and carving out its real-world applications.