The arrival of ChatGPT moved up the timetable for the artificial intelligence (AI) technology industry. As Generative AI, once considered a future that would come someday, quickly seeped into everyday tasks such as search, customer service, development, design, and content creation, corporations rapidly reshuffled their investment priorities.
What this experience left in the market is the expectation of "what is the next inflection point." Quantum computing is a field often mentioned as a candidate.
On the 12th of last month at Korea University in Seongbuk District, Seoul, Mike Piech, vice president of business development at the U.S. quantum computing company Rigetti Computing, said, "The error rate, the key performance indicator for quantum computers, will improve gradually over time like an asymptote," adding, "There is no magical moment." An asymptote refers to a curve that does not reach a target value all at once but gradually approaches it. Vice President Piech majored in computer science at the Massachusetts Institute of Technology (MIT) and received an MBA from INSEAD in France. After serving as a vice president at Red Hat, Piech joined Rigetti in 2024.
Rigetti, founded in 2013, is a superconducting-based quantum computing corporation that pursues a full-stack strategy, building in-house everything from quantum processing units (QPUs) to quantum computer systems and software (cloud). In September last year, it won a three-year, $5.8 million (about 8.4 billion won) contract related to superconducting quantum networking from the Air Force Research Laboratory (AFRL).
While today's computers process information in 0s and 1s, quantum computers use a unit called "qubit." Using the properties of quantum mechanics, they have drawn attention for their potential to compute in a way completely different from conventional computers.
There are two main axes most often cited when describing quantum computer performance. The first is the number of Qubits, and the second is the error rate. As Qubits increase, there is more room to tackle larger problems, but if errors are frequent, the results are hard to trust. Rigetti uses "fidelity," defined as 1 minus the error rate, as its benchmark. For example, if the fidelity is 99.5%, the error rate is 0.5%.
Piech explained that it is hard to look at the number of Qubits and the error rate separately. To make computation valid when errors have not been completely eliminated, you need to use multiple Qubits together to correct errors, which can greatly increase the additional Qubits required in the process. In the end, increasing the number of Qubits and reducing errors are intertwined, making it difficult for quantum computers to reach a usable stage by focusing on only one side.
Piech added, "Quantum computing is still in its early stages, and it is an industry in which every layer must mature at the same time rather than just one issue," and "Not only the quantum processor chip (QPU) but also the surrounding electronic equipment and the operating system and user-environment software must advance together for customers to approach useful computing where they can weigh returns on investment."
Separate from this reality, investor interest in the quantum field has heated up over the past year. Piech said, "Investor interest does not move to exactly the same rhythm as the concrete milestones of technology development," and "Rigetti is consciously avoiding hype and is choosing to set aggressive yet realistic goals and then achieve them."
Rigetti has currently achieved 84 qubits at 99% fidelity and 36 qubits at 99.5% fidelity, and plans to offer a near-term target of 108 qubits at 99.5% fidelity. Around the end of this year, it presented a target of at least 150 qubits at 99.7% fidelity, and around the end of 2027, at least 1,000 qubits at 99.8% fidelity.
It also plans to expand cooperation with Korea, the United Kingdom, and Israel. Piech visited Korea to sponsor the "Quantum AI Competition (Hackathon)" organized by Norma, a Korean partner. The competition was held to promote participants' learning and knowledge sharing and to raise awareness of quantum computing.
Piech added, "Over the past year, Korea has shown a clear trend of trying to preempt the quantum industry through government investment," and "We want to broaden our touchpoints with the Korean ecosystem. The goal is to install at least one, and if possible several, quantum systems on-site in Korea to grow the ecosystem."