By 2027, PsiQuantum anticipates that photons will deliver a million-qubit supercomputer.
Alpha System
Although there are many ambitious goals in the field of quantum computing, California-based startup PsiQuantum has set one of the most ambitious ones to date: by 2027, it plans to provide a fault-tolerant quantum computer with about a million qubits. If this goal is accomplished, quantum computing may bring about the long-anticipated revolution in disciplines like chemistry and materials science.
The business, which was founded in 2016 by four researchers from British universities, has raised $1.7 billion to construct a silicon photonics-based optical quantum computer. At a new facility in Milpitas, California, PsiQuantum is now putting together the “Alpha System,” its first completely functional prototype. The business is speculating that by expanding on established networking and photonics technology, it may achieve critical scale more quickly than rivals that use “matter-based” strategies like superconducting qubits, trapped ions, or neutral atoms.
You can also read Orange QS Celebrates 5 Years of Quantum Innovation in Delft
Planning for Scale: The Fault-Tolerance Imperative
Focusing solely on developing a full-scale, fault-tolerant quantum computer from the beginning is a fundamental component of PsiQuantum’s approach. There is now increasing agreement that complete fault tolerance is necessary for actual utility, whereas early industrial hopes focused on smaller, “noisy intermediate-scale quantum” (NISQ) computers performing meaningful work without error correction.
The decision to seek optical techniques was motivated by this original assumption, according to Pete Shadbolt, cofounder and chief scientific officer. Four major issues must be resolved in order to reach the millions of qubits required for error correction: cooling, control, communication, and manufacturability. Photonics may make these issues easier to handle than competing technologies.
A New Approach to Cryogenics and Connectivity
Because matter-based qubits frequently need to be chilled to almost absolute zero, they are extremely vulnerable to radiation and temperature changes. Since photons, or light particles, are naturally resistant to radiation and heat, they can theoretically function as qubits at room temperature.
Since the design of PsiQuantum’s hardware depends on superconducting photon detectors that function between 2 and 4 kelvins, cryogenic temperatures are still necessary in practice. But it’s far simpler to reach these temperatures. PsiQuantum has created cryogenic cabinets the size of a server rack that can accommodate roughly 250 chips, in contrast to the dilution freezers used by superconducting qubits, which often house one or two chips. A cryoplant built by engineering behemoth Linde will cool three of these massive cabinets at the Milpitas facility.
Packing control electronics near the qubits is made possible by photons’ resistance to heat and radiation, which helps streamline system design. Furthermore, the ability of qubits to be transferred over regular telecom fiber significantly simplifies communication. Recently, the business showed that it could send qubits with 99.7 percent fidelity over 250 meters of fiber.
Leveraging the Semiconductor Industry for Mass Production
Since the majority of existing systems are custom devices, manufacturing is one of the main barriers to large-scale quantum computing. In order to solve this, PsiQuantum developed a commercial fabrication technique for its chips by utilizing established silicon-photonics technology.
Shadbolt highlights that the company was established with the realization that it was necessary to take advantage of the trillions of dollars that had been invested in the semiconductor industry over the previous fifty years in order to build millions of high-maturity devices. Global Foundries is now producing thousands of PsiQuantum’s chips commercially in a semiconductor fabrication facility in Malta, New York, even though the chips contain innovative components like superconducting photon detectors and ultrafast optical switches.
You can also read Federal Reserve Board Warns Of Quantum Attack On Bitcoin
The Technical Hurdles of “Flighty Photons”
Notwithstanding these benefits, the photonic method has many drawbacks. Simon Devitt of the Center for Quantum Software and Information at the University of Technology, Sydney, says photon generation is nondeterministic due to the linear optics system; therefore, immediate fault tolerance is needed.
About 25 to 50 percent of the time, gate operations in the PsiQuantum system fail. While the business uses “multiplexing” to conduct numerous photon-generation attempts and choose the ones that work, this only addresses a portion of the issue. Error correction is required to address the remaining gate failures. According to Devitt, this implies that a sizable amount of the error-correction budget is spent just correcting gate failures, leaving little room for other causes of errors.
After gate failures, optical loss is mentioned as the second most common cause of mistakes. Waveguides, photon detectors, and optical switches are the three main components whose performance largely determines this loss. Even though data indicates that the waveguides and detectors seem ready, the company’s switches continue to have excessive losses.
Shadbolt is still certain, though, that the problem has less to do with a basic restriction of material science and more to do with purity and production. He maintains that there are no significant obstacles and that thousands of little, gradual advancements in the chips’ geometry, design, and manufacturing accuracy will be the key to success.
Testing the System, Not Seeking Supremacy
The first real test of the business’s overall design is the Alpha System, which is now being built in Milpitas. Mercedes Gimeno-Segovia, vice president of system architecture, makes it clear that quantum algorithms will not be used in these preliminary tests. PsiQuantum feels that NISQ machines perform too differently from fault-tolerant ones to offer meaningful information, in contrast to other businesses that employed NISQ prototypes to demonstrate quantum supremacy on “toy problems.”
Rather, the Alpha System is intended to determine whether the behavior of the system is consistent with the expectations of the company’s theoretical models, which is essential for developing large-scale systems in the future.
Shadbolt anticipates having the Alpha System cooled by the end of the year, allowing for the start of trials in early 2026, if PsiQuantum continues on its current trajectory. Although he warns that the financial difficulty of raising the required money beyond the substantial cash already secured may prove to be the main obstacle, analyst Paul Smith-Goodson thinks the company has a realistic possibility of achieving its ambitious technical ambitions. PsiQuantum specifically plans to shift its attention away from security and cryptography and use its future utility-scale machine to address significant global issues like medicine development, materials research, and climate change.
You can also read Adaptive Quantum Circuits Conference 2025 (AQC25) At Boston
