PsiQuantum wants to build an impossible quantum computer: one that calculates with light particles

On the south side of Chicago, on the grounds of a former steel plant, a 6,000-square-meter building is rising that could house the world's first fault-tolerant, useful-scale quantum computer. Behind it is PsiQuantum, a Palo Alto startup that has decided to go for something that almost no one else tries: build the machine using light particles.

Quantum computing has been promising for decades to solve problems that would take current computers millions of years, from the design of drugs to the simulation of materials or batteries. The obstacle is that existing prototypes are still too small and error-prone to be useful. PsiQuantum wants to skip that intermediate phase and directly build a large-scale machine.

While IBM, Google and Amazon build their qubits with superconducting circuits, and other companies use trapped ions or atoms, PsiQuantum has opted for photons. It is a path that it shares practically only with the Canadian Xanadu among the major players in the sector.

The advantage, according to Terry Rudolph, one of the four physicists who founded the company in 2016, is that photons maintain their quantum state for a long time. The drawback is that they barely interact with each other, which is essential to calculate. The solution came in 2001, when a team from Los Alamos and the University of Queensland showed that these interactions could be simulated by sending light through a network of mirrors and detectors. That discovery is the basis of all PsiQuantum technology.

Unlike its competitors, which gradually scale up larger and larger systems, PsiQuantum has focused its efforts on developing the technology and manufacturing capacity needed for a system of around one million physical qubits in one fell swoop. Its CEO, Victor Peng, admits that the company needs to have almost the entire process running before building the first machine, which lengthens the initial times but, he defends, makes it easier to replicate on a larger scale later.

To achieve this, it manufactures its own chips together with GlobalFoundries, engraved with channels that direct light instead of cables, and incorporates barium titanate, a ceramic compound that allows photons to be switched with great precision. Another advantage of the optical approach: Only the detectors that measure photons need to be cooled near absolute zero, while other technologies require cooling the entire system.

Illinois Governor JB Pritzker has promoted the Illinois Quantum and Microelectronics Park with $500 million in state funds to turn the area into a quantum hub comparable to what the Stanford Research Park was for Silicon Valley. Pritzker does not want to repeat the mistake of the 1990s, when talent trained at Illinois universities—the creators of the Mosaic browser, among others—ended up leaving for California. The park already has six other tenants committed, including IBM, and will host a DARPA test program.

PsiQuantum has raised more than $1 billion in financing and has received $100 million from the US Government through the CHIPS Act, in addition to building a second plant in Brisbane, Australia. The company has moderated its promises: it no longer claims to have a functional quantum computer in 2027, but rather "operational" facilities, with cooling systems ready to install the hardware.

DARPA, which has been evaluating its technology since 2023, is increasingly optimistic about the sector: its current head, Micah Stoutimore, has even stated that someone could build a quantum computer of useful scale around 2033. If PsiQuantum meets its deadlines, Chicago could be the place where that machine is born.