DCVC DTOR 2024: With quantum computing, we invest in the picks and shovels

The stubborn reality standing in the way of practical quantum computing is that qubits are exquisitely sensitive to quantum noise (random fluc­tu­a­tions arising from the funda­men­tally inde­ter­mi­nate state of matter) that can cause them to decohere. Whatever physical system is being used to embody qubits in a state of coherent super­po­si­tion and keep multiple qubits entangled, it will require elaborate and expensive controls. Even then, high error rates are a fact of life for quantum computing hardware builders.

At DCVC, our approach to quantum computing has been to invest in companies selling picks and shovels to businesses toiling to make quantum systems, rather than building the quantum computers themselves. (The exception is DCVC portfolio company Rigetti Computing, a pioneer in quantum computing that went public in 2021.) For example, we’re ardent backers of Q‑CTRL, a Sydney-based startup founded and led by quantum control researcher Michael Biercuk. The company initially specialized in quantum error suppression software, which pinpoints the sources of error in a quantum calculation and helps operators reduce that error by optimizing the control pulses used to manipulate qubits. IBM was so impressed by Q‑CTRL’s work that it decided to include the software in its Quantum Pay-As-You-Go Plan, which offers cloud-based access to its quantum processors. ​“There’s a growing recognition in the industry that in order to get any kind of useful results from quantum compu­ta­tions, Q‑CTRL’s error suppression technology will be indis­pens­able,” says DCVC general partner James Hardiman.

Lately, Q‑CTRL has also been surging ahead in another area: quantum sensing. Essentially, quantum sensors take the noise problem in quantum computing and turn it on its head. ​“The thinking is, if these things are so sensitive, let’s just make sensors out of them, so that the bug becomes a feature,” Hardiman explains.

Q‑CTRL starts with validated quantum sensor designs and adds its proprietary AI and quantum control algorithms to suppress noise. In this way it’s developing prototype quantum-enabled inertial sensors that could provide position, navigation, and timing information in situations where GPS data is unavailable (say, due to a military conflict or a geomagnetic storm). Its ultra-sensitive gravimeters and magne­tome­ters are also helping to create more detailed geophysical maps, which provide yet another backup form of navigation, as well as data about potential ore deposits or water tables.

We still believe in the promise of quantum computing. In fact, we see it as an absolute requirement for the future, since there are classes of problems, such as training future high-dimensional AI models with trillions or quadrillions of parameters, that conven­tional computers simply can’t solve given a reasonable amount of time or energy. We hope that fundamental efficiency improve­ments in conven­tional computing algorithms, together with the emergence of specialist AI chip archi­tec­tures such as those being developed by DCVC-backed Mythic, can make the path to practical quantum computing smoother. And in the meantime, we continue to invest in adjacent, enabling tech­nolo­gies like quantum error control and quantum sensing that help innovators put the potential of quantum technology to practical use in the much nearer term.