Xanadu Quantum Technologies has launched a project with Mitsubishi Chemical to develop quantum algorithms for extreme ultraviolet (EUV) lithography.

The two companies said they hope to establish the first concrete use case of quantum computing to advance new semiconductor chip fabrication technologies.

EUV lithography is a wafer-patterning technique necessary for the manufacturing of the most advanced chips.

Under the terms of the partnership, researchers from Mitsubishi Chemical's Materials Design Laboratory will provide information about the molecular structures, compositions, and reactivity of EUV photoresist materials, while a team from Xanadu will provide quantum algorithms expertise to support the design of simulation algorithms that “model light-matter interactions and secondary electron effects.”

"Continued progress in chip miniaturization hinges on breakthroughs in EUV lithography and the design of superior photoresist materials,” said Torin Stetina, senior quantum scientist at Xanadu. “Precisely modeling how these materials interact with EUV light remains a formidable challenge. Using quantum computers to simulate these interactions represents an exciting frontier in tackling this problem, offering a path to uncover material properties for future semiconductor generations."

Founded in 2016, Canada-based Xanadu has made a number of announcements in 2025, following a number of successful fundraises to support its efforts to build and commercialize photonic-based, fault-tolerant quantum computers.

In February 2025, Xanadu built a prototype of what it said was the world’s first universal photonic quantum computer. Dubbed Aurora, the 12-qubit machine consists of four photonically interconnected modular and independent server racks, containing 35 photonic chips and 13km of fiber optics.

The system operates at room temperature and is fully automated, which Xanadu says makes it capable of running “for hours without any human intervention.”

Last month, Xanadu opened a CA$10 million (US$7.3m) advanced photonic packaging facility in Toronto, Canada. Described as the only end-to-end, ultra-low-loss photonic packaging facility of its kind in the country, the site will be open to external customers across both academia and industry.

Australian science agency uses quantum machine learning to model electrical resistance of chip materials

Elsewhere this week, engineers at Australia’s national science agency, CSIRO, said they have demonstrated the world’s first use of quantum machine learning to fabricate semiconductors.

In a new research paper published in the journal Advanced Science, CSIRO detailed how it modeled Ohmic contact resistance – the electrical resistance at the interface between a metal and a semiconductor material in a device – by testing its quantum machine learning model on data relating to 159 experimental samples of GaN HEMT (gallium nitride high electron mobility transistor) semiconductors.

The team then narrowed its set of parameters for the experiment from 37 to five, in order to ensure it could be compatible with the “very limited capabilities” of currently available quantum computers, explained Professor Muhammad Usman, head of quantum systems research at CSIRO’s Data61 team, to CSIRO-owned science news outlet Cosmos.

A novel Quantum Kernel-Aligned Regressor (QKAR) architecture was then developed by the team, which included a quantum feature map to translate classical data into quantum states in the form of five qubits.

CSIRO said its study is the first to show that the chip fabrication process can be improved via the application of quantum machine learning to real experimental data and that, because only five qubits are needed, the method is immediately applicable to current quantum architectures.

“This classical machine learning technique takes that outcome that the quantum method has extracted, and then it’s trained to guidance back to the fabrication. It can tell us what the important parameters in the fabrication process are, which play the critical role, and what needs to be changed or tuned to optimise fabrication,” Usman told Cosmos, adding that the QKAR model can be used for other materials besides the initial proof-of-concept test on GaN.