On EQT Ventures’ investment in Seeqc, the digital quantum computing company
When my colleague Ted and I first met Seeqc’s co-Founder/co-CEO, John Levy, earlier this year, he started the conversation in a rather unconventional way. Dispatching quickly with the typical niceties that occur between founder and investor in a first meeting, John instead dove right in: “so tell me, where are you guys on quantum?”
With the growth of quantum computing startups in the last few years, our team has looked at a number of approaches, including ion trap, photonic, and superconducting. However, having not made an investment in the space yet and with this being our first meeting with John, we must have said something deflective and obviously not all that memorable. John leaned forward in his chair and narrowed his gaze. “Well then, allow me to tell you where Seeqc is. Let’s dig in…”
Over the next hour and then the following months, one thing became abundantly clear. John and team were on a mission to transform superconducting quantum computing and there wasn’t a minute to lose.
For more than a decade, John and his co-founder/co-CEO Oleg Mukhanov led Hypres, the world’s leading developer of superconducting electronics. Here, they applied a proprietary logical framework to address some of the US government’s most challenging problems. During this time, the once-thought impossibilities of quantum mechanics started to become a reality. The world’s largest technology companies began committing extraordinary resources to achieve quantum computing supremacy and a market soon followed.
In 2017, the world’s top quantum experts made a seminal declaration that quantum computing’s commercialization was only five years away and would soon dominate the largest industries on the planet. The only problem was that arguably the world’s best quantum technology was still analog, heavily inundated by noise and plagued by the challenges of scalability.
Computing had come full circle. Quantum computing needed to be scalable in order to achieve commercialization and unlock the next frontier, and making it digital had become a strong contender to do so.
Like many of the great technological breakthroughs of the last century, John, Oleg, and their colleagues recognized that their work could be more broadly applied as it contained a potential answer to a foundational hurdle of commercialization. After an arduous planning process and the backing of some of the world’s top seed investors in this space, Seeqc recruited quantum physicist Matt Hutchings as co-founder and spun out from Hypres just over a year ago. Today, Seeqc is on the cusp of transforming not only superconducting quantum computing but redefining the realm of the possible within quantum technology.
At EQT Ventures, part of our founding was built on the principle that everything that can be digital, will be digital. Most often, we find great founders and companies that are on a mission to transform an industry from analog to digital. It’s rare to have the opportunity to invest in a team and a technology that has the potential to go from digital, to, well, the next thing, for the biggest industries in the world.
Until now, advances in quantum computing have been premised on a prevailing maxim — the more qubits, the better. The world’s largest quantum players have centered their entire operations to win an arms race in employing an ever-increasing amount of qubits. This is understandable. Qubits are the core engine of quantum computers and the fundamental differentiation from a classical computer.
They are also incredibly powerful, doubling a quantum computer’s power with each additional qubit added to the system. This fundamental difference unlocks the power to make laughably impossible tasks, like the factorization of a 2048-bit number, rather simplistic for a quantum computer. Of course, if you’re a regular Medium reader, you already knew that.
Like the transistor in eras past (and microprocessors made up by transistors), qubits have been the first building block by which the foundation of the quantum computing industry has been constructed. However, we do not believe that the future of quantum computing will be defined by the qubit alone. Not being quantum physicists, we recognize that this is a relatively audacious statement and will surely elicit plenty of fan mail (for the sake of good order, please send all of it to me.) But we’ve been convinced that an outsider’s perspective is the right one to see what the industry has been missing. Let us explain why.
There are two fundamental points that are essential to understand here about all of this qubit magic. First, the qubits themselves are just computational vehicles that need to be run in parallel with a classical computer, which is responsible for controlling qubits, and providing and receiving the data that the qubits are providing. Second, most designs for qubits only work in a cryogenic state. This means that the classical computers are run at room temperature and have to transmit data (most frequently and effectively by electricity) to the qubits, passing through a complex configuration of wires, processors, RF transmitters, Digital-to-Analog converters, and microwave generators all connecting to refrigerated dilution chambers
In order to process more data, you need more qubits. You can see where this is going. More qubits means more wires, more heat, more space, and larger cryogenic chambers. Not to mention, once something breaks or needs to be adjusted within this complex system, the entire computer would need to be taken apart, adjusted, then reassembled. So goes the saying, quantum computing is a physicist’s dream and an engineer’s nightmare (fun fact: bugs are called bugs because physical bugs came into early room-scale computers).
So: the more qubits, the more gear, cables, conversions, magnetic fields, the more error or, in quantum speak, noise (technically quantum decoherence). All of this activity drastically distorts the qubit and the signal that the operators are able to glean from it. So: more qubits = more noise = harder to filter out the signal = not useful. VC math, nice.
Therein lies the industry’s conundrum: a 500-qubit machine is more powerful in raw computing power than all of the world’s computing power combined. To be the first to build such a machine would be an extraordinary milestone. But, without integrating any way to mitigate or correct the errors coming from this incredible technology, the most powerful computer in mankind’s history would have no practical use. More than 99% of the signal provided by this computer’s incredibly powerful qubits would consist of noise.
However, the qubit-centric mentality has been more pragmatic than the conundrum would imply. It’s simply a matter of how damn hard this technology is to design, build, and implement. First things first. Consequently, the obvious limitations of qubits in scalability and application to real-life problem sets have largely been put off and filed under the category of “we’ll deal with that when we get there”.
Well, here’s the catch: Seeqc is already there.
Early experiments with superconducting qubits were carried out using “wet” dilution refrigerators. These required expensive liquid helium to operate, but more critically the cooling power was low, not more than about 50 μW. This meant that superconducting qubit experiments were limited to only a few RF control lines controlling a limited number of qubits.
Today, a typical commercial dilution refrigerator is “dry” (requires no liquid helium), with a cooling power of 1000 μW and can be bought with 200 factory installed RF lines. Such a system should be able to cool a quantum processor with about 400 qubits.
There is no technical reason for why one could not build even bigger dilution refrigerators and this is likely to happen. This should make it possible to scale up to circuits with perhaps thousands of qubits. However, in order to go beyond this other solutions will need to be found.
One such approach for this is using “conventional” electronics (or electronics originally intended for use in satellites) inside the refrigerator. CMOS (complementary metal–oxide–semiconductor), commonly used in electronic devices has shown to be the only technology allowing both cryogenic operation and the integration of the billions of transistors required to operate a very large number of qubits. However, CMOS has to operate at high power resulting in heat and thus noise.
Seeqc offers another approach called RSFQ (Rapid Single Flux Quantum), which has the potential to solve not just the noise problem, but also making Quantum Computers modular, digital, scalable, cost effective and thus usable. The foundational work was developed by Seeqc co-founder/Co-CEO Oleg Mukhanov while a PhD student at Moscow State University in the 80s, before quantum computers could be realized.
Once he and John (then the Chairman of Hypres) realized the commercial potential of what they had built, they recruited Matt, who was designing IBM’s first qubits, and spun out the technology to start Seeqc, bringing with them an entire superconductive chip foundry, a significant portfolio of issued patents and a team of experienced superconductive electronics engineers and SFQ circuit / system designers.
In just over a year, Seeqc has finalized a digital system-on-a-chip (SFQuClass Digital Management Quantum), addressing noise and performance issues such as latency, energy efficiency, system complexity, calibration, as well as cost. The architecture could be seen as analogous to the evolution of classical computers from tubes to transistors to microprocessors. The chip requires orders of magnitude lower levels of energy, and replaces racks of expensive, high energy, analogue microwave circuitry with proprietary digital chips that are co-located with qubit chips as multi-chip modules in the same cryocooled environment.
Most importantly, these chips are coupled directly with qubit chips and communicate without cables. This represents a clear, and cost-efficient, pathway towards addressing existing challenges with analog, microwave-controlled architectures and will help scale the current generations of superconducting quantum computers beyond the NISQ era.
And that’s just the beginning of the story. Seeqc has thought much farther ahead than just hardware-based error correction as the DQM system-on-a-chip architecture enables the company to co-design specific quantum hardware that match the unique requirements of quantum algorithms and applications. They’ve designed an entire platform for not just what quantum computing needs today, but for where it is ultimately going.
With that, we’re extremely excited to announce the company’s $22.4M Series A investment, led by EQT Ventures with investments from FAM AB, M Ventures, BlueYard Capital, New Lab Ventures, and the Partnership Fund for New York City. In this round, Seeqc is also backed by M Ventures, the strategic corporate venture arm of Merck KGaA in order to begin co-design work with Merck in drug discovery.
Ted, myself, and the rest of the EQT Ventures team could not be more thrilled to join Seeqc on this mission and be a part of shaping what’s next!
