It is clear computing is in for a revolution if a company can have unicorn status with its main product still a few years away. In years past this mismatch in funding to goods would seem absurd but quantum computing is setting new trends, especially among VCs lately.
If we had been placing bets about which quantum computing startup would be the first to raise a billion dollars three years ago, D-Wave and Rigetti might have topped the list. The former has raised $256 million since its inception in 1999 and since its inception in 2013, Rigetti has rounded up just shy of $200 million.
So why is it that a quantum upstart few have heard of has managed to capture $665 million to date?
It might have something to do with manufacturing scalability and reliability, as we described in a chat with GlobalFoundries and the startup, PsiQuantum, in May. It might also be because the high sights they’ve set—building a fault-tolerant, million-qubit quantum system—seem to have grounding in manufacturable, scalable reality with GF as a partner that has already made its own investments in superconducting materials, photonics, and quantum system making.
PsiQuantum announced $450 million in Series D this morning, with investors including Microsoft’s VC firm, M12, and entrants including Ballie Gifford, along with Blackbird Ventures and Temasek with BlackRock leading. The next step for companies in understood segments is often going public but for a quantum computing company, especially one that has not yet brought products to the broader market, that might be a bit of a stretch.
It’s not common to see Series E or F rounds before a company goes public or gets snagged by one of the few companies able to make such an acquisition. But in a conversation with company leaders today, it became clear they’re playing the long game. For now, at least. Microsoft, after all, has its own quantum ambitions and the company’s VC arm, M12, has been tracing PsiQuantum’s evolution since its Series C with strong affirmations from M12’s managing director, Samir Kumar today.
“We have a very good alignment in terms of the philosophy with Microsoft, they’ve been doing this a long time with a big team and they’re very serious,” PsiQuantum CEO and co-founder, Jeremy O’Brien tells The Next Platform. “In particular they’ve been rare in that they understood from the beginning the importance of error correction and fault tolerance.
PsiQuantum and Microsoft have a shared perspective on the need for a good number of logical qubits enabled by fault tolerance and error correction on 1 million-plus physical qubits, when it comes to building a truly useful quantum computer,” says M12’s Kumar.
PsiQuantum is exclusively focused on building a fault tolerant quantum computer supported by a scalable and proven manufacturing process. The company has developed a unique technology in which single photons (particles of light) are manipulated using photonic circuits which are patterned onto a silicon chip using standard semiconductor manufacturing processes. PsiQuantum is manufacturing quantum photonic chips, as well as the cryogenic electronic chips to control the qubits, using the advanced semiconductor tools in the production line of PsiQuantum’s manufacturing partner, GlobalFoundries.
PsiQuantum is raising the entire range of capital it will need to move from concept to reality with the Q1 quantum computer, which recently entered manufacturing at GlobalFoundries. The two companies collaborated to introduce superconducting photon detectors into the 300mm manufacturing process to work toward scalable production of quantum devices. While PsiQuantum does not yet have a functional quantum computer, it sees a path to a million-qubit, fault tolerant system coming in the next few years. In other words, this investment will have to stretch to build the teams and further technologies over the long haul.
As Pete Shadbolt, PsiQuantum CSO tells us, the startup has a “handful” of paying customers even without a functioning quantum system. These customers are looking at algorithms in chemistry and finance. “We’ll spend the next few years continuing to develop the technology to the point where it’s useful. That’s expensive. We have 150 people, most in Silicon Valley, and we do novel semiconductor process development, for instance, we have been putting superconducting photon detectors into large R&D like GlobalFoundries. That’s expensive. We’re building a big machine, we’ll spend all this money building the quantum computer.”
Shadbolt adds that this is an expensive undertaking all around to date, with a repeat of some of the manufacturing investments coming again soon. We have to wonder if that means another process node but the company did not clarify. “We’ve put tools into the production line at GF over the last few years and enabled new manufacturing steps in the process we use to build our chips. We’ve put the superconducting photonic detectors in the 300mm manufacturing process at GF. That is pretty expensive, we’ve been through a large number of wafers doing that and it’s a complicated manufacturing process. That’s an example of something we’ll do again in the near future, and that isn’t cheap.”
The quantum devices are based on a blend of traditional packaging, control, and other technologies well-understood at GlobalFoundries with the addition of novel photonics and cryogenic features. After all, once built and proven at scale (manufacturing and in production), any number of cloud builders who have yet to provide their own quantum devices for remote users would be keen to have a stable manufacturable quantum technology to offer. And in the cloud, no one can see your IP.
Speaking of which, the company has a sizable list of patents ranging from novel materials to photonic operations. “We’ve been patenting our work since the founding of the company, patenting across entire technical spectrum, all the way to basic semiconductor processing to materials development up to quantum algorithms and everything in between—packaging, devices, photonics, CMOS electronics, cryogenic systems, error correcting codes,” Shadbolt says.
There is a reason we emphasized the Microsoft angle so heavily here. We are making the bold prediction that one of the most noteworthy quantum acquisitions of the 2020s will be Microsoft buying PsiQuantum. While other cloud builders like Amazon are becoming “marketplaces” for quantum services from multiple device makers, Microsoft has made its own investments in building its own machines. With the GlobalFoundries partnership and assumed manufacturability and usability, the first cloud to provide this kind of service could own the future of quantum since placing these systems in ever supercomputing center, at least as they’re built today, will not make sense, especially given near-term limited application reach.
No matter what happens with the company, CEO and co-founder O’Brien believes in a world with million-qubit systems. He tells us this isn’t something that is incrementally better, or even an order of magnitude better than conventional computing—it’s world-changing.
“This is the only path to a million-qubit error correcting system which is extraordinary value. Our plan is to become a truly great company, not to be acquired. We uniquely have a path to the most profoundly world-changing technology that humans have uncovered to date. We are categorically not building an incrementally better chip delivered none months later, this is the foundation of a future economy and a transformation in how humans go about doing things. It’s not unprecedented but it is akin to major revolutions in the past.”