QC Design: Shaping the Future of Quantum Design Automation
“If we want to address truly transformative applications with quantum computers, we need to make them fault-tolerant.”
However, the path to fault tolerance is far from straightforward, demanding precision at every step—from optimizing individual qubits to engineering their overall connectivity. It requires solving a lot of complex engineering challenges. Yet, fiddling with qubits in the lab can be both time-consuming and expensive.
What if you could simulate fault-tolerant quantum computers, much like the semiconductor industry is using electronic design automation software?
QC Design was founded in 2021 by Ish Dhand and Martin Plenio to develop design software and architectures for fault-tolerant quantum computers. Their quantum design automation software allows quantum hardware manufacturers to build the ultimate computing machines allowed by the laws of physics.
It raised a pre-seed round from Vsquared Ventures, Quantonation, and Salvia in 2022 and recently secured €4M in funding through the EIC Accelerator program by the European Innovation Council.
Learn more about the future of quantum design automation from our interview with the co-founder and CEO, Ish Dhand:
Why Did You Start QC Design?
It was a journey from when I finished my postdoc with my now-cofounder Martin Plenio in 2018. Around that time, we felt that quantum computing had matured enough to bring useful, real-world applications within reach. After working for some time at Xanadu, a photonic quantum computing startup in Canada, learning a lot about fabricating and designing chips, and getting excited about startups and taking innovations from zero to one, I got back to Germany to found QC Design.
My co-founder and I felt that quantum computing would be truly transformational for humanity, so we started QC Design to help our customers, quantum hardware companies, to get to quantum fault tolerance.
Today, even with the best quantum computers, one in a thousand operations fails, which is simply not good enough. To address truly transformative applications, such as developing new drugs or battery materials, you’ll need to bring down error rates at least to one in a million—several orders of magnitude improvement. And the only reasonable way to get there currently is to use lots of error-prone physical qubits to encode one clean, logical qubit.
How Does Quantum Design Automation Work?
Engineers use AutoCAD software to design buildings before anyone starts to build them. It’s a simulation tool that runs on a classical computer and helps to put ideas and designs into software and check whether they would work—whether the building would be stable and the materials have the right properties and strength to support the building.
We’re building AutoCAD for quantum computing companies, helping their teams to understand the noise and resulting device performance of any quantum hardware.
One important question to achieving fault tolerance is how many error-prone, physical qubits do you need to build one clean, logical qubit. Depending on the type of noise in your quantum computing hardware, the requirements are very different, and it can make a difference of several orders of magnitude in the number of physical qubits needed to encode a logical qubit.
For example, if you had only one type of noise for your physical qubits, you’d need fewer, maybe only a hundred physical qubits for one logical qubit. In practice, however, there are more like 20 different types of noise, and you need about 10,000 physical qubits to implement one logical qubit, which is an enormous amount given that most quantum computers currently have dozens, maybe hundreds of physical qubits.
Also, if your noise level and, thus, your error rate is too high, you won’t get any logical qubits irrespective of how many physical qubits you have: if adding a physical qubit introduces more errors than it removes, you’re only making the problem worse by adding qubits. You need to get to a point where adding qubits for error correction effectively removes errors.
We help quantum hardware teams understand whether they can build logical qubits at all and, if so, how they can improve. With our software, they can answer such questions and explore new architectures for their quantum machines.
There are four aspects you should keep in mind when developing the architecture for a fault-tolerant quantum computer:
- Qubit connectivity: How are the physical qubits connected, e.g., is it a square or hexagonal lattice? This influences the choice of your error correction code.
- Decoding problem: Which classical algorithm is used to find errors? As you can’t measure all the qubits, you need a sophisticated sequence of measurements to find out where an error has occurred.
- Qubit definition: How should you connect hardware components to build physical qubits? For example, in photonics, there are many ways to use photons as a qubit, and some are more efficient than others.
- Qubit control: How will you control, feed in, and output information? For example, for superconducting qubits, you may use microwave pulses, but you need to figure out how long, how intense, and how focused they should be.
All four have a huge impact on quantum computer’s architecture and how to make them tolerant of many different faults.
How Are Engineers Using Your Software?
Our users are quantum engineers, designers, and architects who use our software not only to produce the initial design of a quantum computer but also when making any changes to its design during development: you make a design, run tests to validate whether it works, and then update your design accordingly.
Like in traditional chip design, we expect it to be the same in quantum computing, and people will iteratively work on improving their designs. We may also provide architectures and IP in the future, but for now, we see ourselves as a software tool.
Customers can access our design tool without having to reveal any parts of their IP. It runs on-premise; they can add their own control sequences and study the expected performance of their quantum hardware. We see our job as making their lives easier when they’re designing their quantum computing hardware.
We have implemented just about every quantum correction code proposed in the literature in our software, which gives designers a major productivity boost. Previously, if you wanted to try a new, say, LDPC error correction code, it took a long time to implement it yourself manually. With our software, you can select the code from a dropdown menu, and you’re ready to go. You can start with everything you know from arXiv papers, and we have already taken away the pain of implementing it.
Finally, we can also help with designing quantum sensors or components for quantum communications and building the quantum internet—all of those also need fault tolerance.
How Do You See NISQ Versus FTQC?
We’re very focused on applications, and if customers get value from noisy intermediate-scale quantum (NISQ) algorithms already today, that’s awesome. Many people at quantum companies work on simulators to benchmark quantum algorithms and explore pathways to get to quantum utility as early as possible.
That said, the applications that will totally transform humanity, like developing new drugs, new fertilizers, or new ways to capture carbon, will require fault-tolerant quantum computers (FTQC). And that’s what we’re focused on: helping other quantum companies to get to fault tolerance and unlock those truly transformative applications.
How Did You Evaluate Your Startup Idea?
As we sit in the middle of the stack, and own neither the quantum hardware nor access to end users, you may think we’re in a tricky spot. But we look at this from a different lens: as long as we build something truly useful for our customers, something that can accelerate their hardware developments by weeks, they’ll be happy to use us. Build something useful, and good things will happen.
When you look at the semiconductor industry, electronic design automation (EDA) software has a 1-2% share of the entire industry and has produced giants like Cadence and Synopsis. It’s the software that’s necessary for everything the semiconductor industry builds.
Quantum computing will unlock truly transformative applications and many 100Bs in market value, so if we do something useful, we’ll capture a decent share of that. Also, we can work with any quantum hardware company—we’re not limited in how we choose our partners. We’re sure that when the quantum wave comes, we’ll ride that wave.
What Advice Would You Give Fellow Deep Tech Founders?
I learned throughout my journey as a founder that a big part of the job is to take punches and get back up. Things go wrong all the time; you need to fix bugs, make sure you hit your performance goals, hire the smartest and kindest people, and work with the right people in university, VC, and the industry.
It took me some time to realize that taking punches is the job. It’s the norm, not the exception. And once you start to see it as part of your job as a founder, it’s like any other skill you can train and get better at. The moment I started to see it as a skill and part of my job, it helped me not to get bogged down but come back with vigor the next day, and it made a difference in how quickly we could move as a startup.