One Million Qubit Quantum Computing? How SEEQC plans to scale quantum computers

As exciting as it can be to hear about progress 127 qubits quantum computers or promises of 1000 qubits by 2023, the reality is that quantum computing is not yet a mainstream computing technology that produces revolutionary real-world results with massive, obvious, and immediate impact.

There’s still a lot of promise Y potential Y future in quantum computing.

That’s something the New York-based quantum computing company SEEQC is looking to change. And she faces an uphill battle.

The company is starting by building unique, purpose-built quantum computers in the same way that the first classical computers were built to do just one thing, like calculating missile trajectories in the early days of rocketry. One project, QPharma, is the company’s venture with Merck and other partners to build a quantum computer for pharmaceutical research. One key technique the company uses is essentially fusing classical computers and quantum computers inside the dilution coolers that quantum computers need to reach their optimal operating temperatures (within a minimum distance from absolute zero). The goal: to reduce the instruction round trip and digital-to-analog conversions required by the external quantum computer’s control mechanisms, and to perform the reading, control, and even possible error correction essentially inside the computer. quantum.

That’s using chips with impressive speeds, 30 to 40 gigahertz, that run at very low temperatures to prevent destabilization of qubits.

Perhaps the most challenging work SEEQC is doing is looking for ways to scale quantum computers by multiple orders of magnitude.

“What we’re really focused on is figuring out a strategy on how to scale a quantum computer,” CEO John Levy recently told me at the TechFirst Podcast. “As much as IBM gets to a thousand or thousands of qubits… the question is how do we get to ten thousand, a hundred thousand, a million qubits so that we can scale quantum computers to the complexity of problems that big. companies worry.

“And that’s what we’re doing”.

This starts with using the SEEQC foundry to design and build chips in near real time (think iterations every week) that the company can mix and match for various needs, such as running a quantum variational autosolver or an approximate quantum optimization algorithm. . It helps that the company is a spin-out of hyperactivewhich develops and markets superconducting integrated circuits.

But for Levy, what is critical is not just building the quantum components of a quantum computer, the qubits themselves. What is just as important is integrating them efficiently into a computing environment that makes sense for how we actually solve problems and use information.

Listen to the interview behind this post:

“If you look at superconducting computers… we’re still dealing with analog signals that have to be converted to digital and back to analog, from cold to room temperature to cold. There is a lot of latency. It’s very difficult to use,” says Levy. “But we are used to working with computers, real computers that are all based on chips. So our chips are based on something called Single Flux Quantum, which is different than CMOS (much faster, much quieter, much less power) and allows us to build functional chips for a quantum computer that does things like reading and control, possibly error correction. , co-processing, etc. all in the same dilution cooler, all at chip scale.”

It is an approach similar to that used by computer scientists and physicists at the University of Chicago recently demonstrated – and won the best paper award at the 2021 IEEE International Conference on Quantum Computing and Engineering.

The researchers found that the result of this approach is better scalability and lower error rates of quantum operations.

“We know that in order to scale large, complex systems, we have to shrink them down to a series of chips…it just has to happen,” says Levy.

And what it has the potential to do, according to SEEQC, is take quantum from the computing equivalent of the Werner Braun V2 rocket to something much more SpaceX-like: scalability, repeatability, and lower cost.

Much lower cost.

Up to 97% less cost.

“Right now, the cost of building a qubit… with readout, control, tunable couplers… is around $10,000 per qubit,” says Levy. “Not long ago, it was around $40,000. So there has been a real improvement there. But imagine for a moment that you could do all of that with, I don’t know, five or six chips, something like that, and where you could multiplex a lot of qubits. Well all of a sudden your cost goes from the range of, let’s call it $10,000 to, you know, not a lot of dollars to build a set of these kinds of chips that you can build on six-inch and eight-inch wafers, where you get a lot of these records.

Upon further investigation, Levy told me that “not many dollars” means “hundreds” of dollars per qubit.

That’s a significant game changer.

While there is talk of Shenzhen SpinQ Technology launching a very basic quantum computer for around $5,000, that’s essentially a two-qubit toy for high schools that isn’t even close to typical classical computing speeds. “Real” quantum computers for actual use in industrial or university settings cost millions of dollars. D-Wave, for example, sold a particular quantum computer model for $15 million in 2017, but is now focusing on making quantum cloud computing accessibleso it activates quantum processing just like it activates a new node in Amazon Web Services.

Making quantum computing affordable is a key step in making it useful.

While SEEQC is making specific machines today, the ultimate goal, Levy says, is general-purpose quantum computers. And that requires scale, which fits in quite well with a foundry-based approach to making quantum computers.

“We know that to scale large complex systems, we have to reduce them to a series of chips,” says Levy. “It just has to happen.”

It will work?

Only time will tell. The company is well connected and has multiple grants for its operations in the US, UK and Italy totaling several million dollars. You have a well-connected matrix with useful technology and you’re in an industry that’s growing rapidly in 2021 and you’re it is expected to continue to grow in 2022 with applications in AI, natural language processing, and cryptography.

But it also faces significant competition from the likes of Honeywell, IBM, Google and Intel. And cutting-edge technology tends to make the companies that invent it bleed.

Levy thinks they will, but he doesn’t think the SEEQC approach is the only one that will work.

“Within the next three to five years … you will see quantum computers built on this, on this and other scalable platforms,” ​​he says. “I think you’ll see one based on photonics, and you’ll see us, you’ll see one on superconducting electronics… a superconducting qubit version.”

subscribe to first technology; get one transcript of our conversation.

Leave a Reply

Your email address will not be published.