IonQ has shattered the glass ceiling that previously separated pure-play quantum computer companies from public listed offerings.
There were rumors last year that IonQ was considering an IPO. But rather than raising capital with a conventional IPO, IonQ chose to go public by using a merger with dMY Technology Group III, a SPAC (special-purpose acquisition company). The deal will create a combined entity estimated to be worth about $2 billion.
A few days after the announcement, Patrick Moorhead, president of Moor Insights & Strategy, and I had an interesting conversation with Peter Chapman, CEO and President of IonQ, and Niccolo de Masi, CEO of dMY. There are hundreds of articles explaining details about the transaction and its structure. Our focus here is primarily on IonQ’s cash injection and IonQ’s plans for the money.
IonQ was founded in 2015 by Dr. Christopher Monroe and Dr. Jungsang Kim. They did it with $2 million in seed funding from New Enterprise Associates. At the time of its founding, Professors Kim and Monroe had previously spent more than two decades combined researching quantum physics. And as a result of their academic work, they were confident they could build a programmable quantum computer using tiny atomic particles called ions.
Between 2015 and 2018, IonQ raised an additional $20 million from Google Ventures, Amazon Web Services, and NEA. In 2019, under the guidance of a new CEO and president, Peter Chapman, IonQ raised another $55 million from notable investors that included Samsung, Lockheed Martin, Airbus Ventures, Bosch, Hewlett Packard, and Mubadala. That year, Chapman also opened up cloud access to IonQ by announcing partnerships with Microsoft Azure and Amazon Web Services’ Amazon Braket.
Equity and cash injection for IonQ
Architects of IonQ’s historic deal were Peter Chapman, CEO of IonQ, and dMY SPAC executives Chairman Harry You and CEO Niccolo de Masi.
Chapman told Moorhead and me that the process initially began when an IonQ investor suggested that he consider using a SPAC as an option to go public. The investor also offered to introduce Chapman to de Masi. After meeting with de Masi, Chapman became interested in the SPAC concept and presented it to the IonQ board. Later, after de Masi had raised $300 million in a previous public offering, Chapman and the board gave de Masi a thumbs up to explore a possible merger.
IonQ will receive $650 million cash and 64% of the new equity. The $650 million comes from two components, $300 million raised by dMY III and $350 million from new PIPE strategic investors Hyundai Motor Company and Kia Corporation, MSD Partners, Bill Gates’ Breakthrough Energy Ventures, Marc Benioff’s TIME Ventures, Silver Lake, and Fidelity. Before the merger is fully closed, it remains for the SEC to review and approve the deal.
There appears to be no lack of confidence for IonQ in the investment community.
“This was the easiest PIPE to raise of our career,” de Masi said. “Harry [dMY Chairman You] and I have done five SPACs between us. This PIPE was the most oversubscribed so far. We weren’t able to give everyone what they wanted. Because demand was high, some companies placed orders for a lot more than we were able to provide to them.”
IonQ’s initial investors must be patting themselves on the back. They will be getting a healthy return in the form of new IONQ equity.
Peter Chapman said, “I’ve had to raise money for the company many times. Some people I talked to said it was too early for them to invest in quantum. But if you look at the investors in every previous round, we were the lead investment in their portfolios. And with this deal, we’ve now made our existing investors hundreds of millions of dollars to-date.”
Many technologies are used for quantum computing qubits such as superconducting, photonics, silicon-based, spin qubits, trapped-ions, and several others. IonQ’s quantum technology of choice is trapped-ion qubits created from an isotope of a rare-earth metal called ytterbium.
Ion qubits are created by precision lasers that remove an outer electron from an atom of ytterbium to form a positively charged ion. Next, lasers are used like tweezers to move ions into position in the trap. Once in position, oscillating voltage fields hold the ions in place.
IBM and Google use fabricated superconducting qubits. Honeywell Quantum Solutions uses trapped-ion ytterbium qubits in a different architecture called QCCD, rather than linear traps used by IonQ.
Trapped ion qubits have many advantages, including long coherence times, high fidelities, and fully interconnected qubits. Additionally, ion qubits are a product of nature. Natural ion qubits are perfect and identical to each other. IonQ’s most recently announced quantum computer, currently in development, uses 32 fully-connected ion qubits. That is the largest number of qubits being used in a trapped-ion quantum computer.
How the SPAC sees IonQ and its market
Niccolo de Masi is a physicist by education. He told us there were several reasons why he believed this was the right year for IonQ to go public.
First, cloud computing is essential. It’s the common denominator for connecting everything that’s a service. Growth and dependence on the cloud amplify the importance of IonQ’s partnerships with Amazon and Microsoft. To de Masi, the formula is simple. IonQ’s research team can focus on building leading-edge quantum computers. At the same time, the Amazon and Microsoft cloud services help IonQ monetize its products.
“IonQ’s cloud revenues for the last several months remind me a lot of Apple’s app store revenues in 2008 and 2009,” De Masi said. “It’s lumpy, but sometimes it’s huge and similar to apps. People will continue writing new APIs for IonQ’s quantum computers.”
De Masi went on to comment about quantum market growth. He believes IonQ revenues will begin to take off in a very smooth and impressive way over the next two or three years. He pointed out there are 85 venture-backed companies in the quantum field, most of which are focused on 2030. But IonQ has already made many impressive breakthroughs.
Peter Chapman and Niccolo de Masi believe that quantum computers will eventually solve problems that are impossible for today’s classical computers to solve. As an example, today we use supercomputers for various optimization problems, such as financial Monte Carlo simulations and energy applications. In a few years, quantum computers will allow for faster and more robust simulations and, on a much larger scale, provide more in-depth insight, higher efficiency, and better forecasting.
Almost all experts believe within a decade, large fault-tolerant quantum computers will also be able to accurately simulate large and complex molecules. That capability will likely lead to the creation of new pharmaceuticals and new materials. These applications and many others will eventually create major quantum computing revenue opportunities. Ultimately, de Masi expects IonQ’s market cap to rival that of companies like NVIDIA.
While revenue is essential, Chapman emphasized that over the next several years, his primary focus will be on building a quantum computer that can surpass supercomputers’ capabilities. By doing that, he is also focusing on future revenue growth.
With the new cash injection, Chapman plans to execute a plan for IonQ that evolves over three phases.
Chapman said IonQ is currently in phase one. Developing a very high gate fidelity is vital if quantum computers will be able to solve complex problems. Ion traps already have a low error rate compared to other technologies. Still, it is essential to eventually
drive two-qubit gate fidelity to 99.999% if IonQ is to outperform classical computers. To achieve this kind of gate fidelity, error correction is needed.
In phase two, Chapman plans on achieving a broad quantum advantage for select applications. As an example, IonQ has performed a lot of research and experimentation using machine learning for signature verification with excellent initial results. “Based on our current success,” Chapman said, “we think that machine learning will be the first application area with the widespread adoption of quantum computing.”
Quantum computing error correction is a key element in phase II. IonQ researchers must focus their efforts on refining and implementing error correction. Classical computers rarely make computation errors, perhaps one in a trillion or so calculations. On the other hand, today’s quantum computers make an error every few hundred calculations. Various components and environmental factors create qubit errors. A method is needed to correct these errors while the computation is in progress. The lack of a viable means to correct quantum errors is one reason we cannot build large quantum computers today. IonQ has already developed a working prototype for error correction that uses 16 error-correcting qubits for every logical qubit. Additionally, Chapman recently added Dave Bacon to his staff. Bacon is a physicist who is well-known for his contribution to the Bacon-Shor error correction code and for previously leading Google’s quantum software division.
Chapman calls the third phase of his plan the “holy grail of quantum.” This phase defines the ultimate stage when multi-million qubit fault-tolerant quantum computers become available. It is a stage where quantum computers will be able to solve problems beyond the capability of any supercomputer.
At this stage, the plan calls for IonQ quantum computers to range from tiny desk models to small rack-mounted, room temperature devices to large multi-million qubit quantum computers. All models will be able to be interconnected with fiber optics and remotely operated as a massively parallel device.
Phases I, II, and III hardware evolution
Over Chapman’s three phases, IonQ’s ion trap and vacuum chamber will undergo significant downsizing. In 2016 a trapped-ion quantum computer required a large vacuum chamber. Today, the new 32 qubit system’s quantum processing unit (QPU), including vacuum chamber, is about the size of a card deck.
By 2023, IonQ researchers expect to put the QPU optics, the ion trap, and the vacuum chamber on a single chip. Optical interconnects will allow the chips to be networked together with optical cable. IonQ has already developed working prototypes. Chapman expects a relatively low-cost rack-mounted room temperature system to be available in 2023.
Qubit counts Phases I, II, & III
Peter Chapman and Niccolo de Masi both believe it will require about 72 trapped-ion qubits and a 99.999% two-qubit gate fidelity to outperform a supercomputer. Chapman defines IonQ’s technical roadmap in terms of algorithmic qubits. Algorithmic qubits use just variable amounts of error correction. For exceptionally large fault-tolerant quantum computers, more error correction will be needed requiring multiple physical qubits for each algorithmic qubit. In the near term, reduced error correction is possible, reducing the total number of physical qubits required.
Phase I will increase the number of algorithmic qubits from 22 to 29.
Phase II will have a high-water mark of 64 algorithmic qubits. There is a big difference in the computational ability of 64 logical qubits compared to 29. Many research breakthroughs should occur during this period. For each additional logical qubit, a quantum computer’s computational power is doubled.
In Phase III, IonQ’s trapped-ion quantum computers are planned to have 256 qubits. Optical interconnects will make it possible to dramatically increase QPU power by connecting modules together and increasing qubit count 4x from 2025 to 2026.
In May 2020, I spoke with Peter Chapman after he had been at IonQ for a year. After the interview, Moorhead and I reviewed IonQ’s accomplishments. We were both
impressed with the progress Chapman had made over the 12 months he had been there. Also, we both came to the same conclusion, so I closed the previous 2020 Forbes article with these comments:
“From an analyst’s perspective, IonQ has strong business and technical leadership. The company is doing all the right things at the right time, including infrastructure and process improvements, expansion, and preparation for future opportunities in quantum computing. Peter Chapman is a savvy business leader who understands what is essential to customers and is not afraid of tackling bold challenges.”
It appears that our 2020 assessment of Peter Chapman and IonQ has been validated. Being the first pure-play quantum computing company to go public as a listed stock is definitely a bold move.
- Most experts agree that quantum computing is still in an experimental stage. We believe the $650 million injection into IonQ will benefit the entire quantum ecosystem and shorten the time required to achieve quantum advantage.
- Chapman has it right. IonQ’s priority must be to improve the native gate errors and find a path that allows a larger number of physical qubits to use error correction at scale. If that fails, it will either delay or jeopardize the entire plan. Error correction is critical for the entire ecosystem.
- There are significant differences between assessing a computer’s performance with quantum volume and counting algorithmic qubits. While both are dependent upon gate fidelity, algorithmic qubits are arrived at by calculation and provides a general expectation of a quantum computer’s capabilities. Quantum volume, on the other hand, is determined by running circuits whose results depend not only on fidelity but also on many other factors. We think everyone agrees that a standard quantum computing benchmark is needed. Perhaps even different benchmarks for different domains.
- The door to IPOs has been opened, so this deal is likely to encourage other quantum IPO’s or SPACs. Let’s hope that innovative research being done today isn’t replaced by an IPO-inspired focus on commercialization and pre-IPO prepping of products. We saw that happen when competitive pressures forced redirecting a large portion of AT&T Bell Labs’ financial resources to developing commercial products. That was done at the expense of pure research.
Note: Moor Insights & Strategy writers and editors may have contributed to this article.