Sydney – October 24, 2026 -- Diraq, a University of New South Wales (UNSW) spin-out that builds quantum computers from modified standard transistors, expects chips with thousands of qubits as early as 2028 and universal fault-tolerant machines in the 2030s, according to founder and CEO Andrew Dzurak.
Diraq adds $15M Series A-2 extension to $20M initial round
Diraq raised a $20M Series A in 2022, led by Allectus Capital. This spring it announced a Series A-2 extension led by Quantonation, which brought in an additional $15M. The company spun out of UNSW in May 2022 and opened a company lab on the university campus a few weeks ago, keeping it close to ongoing research.
Modified MOSFET transistors serve as 20-30 nm spin qubits
Diraq's qubits are standard MOSFET transistors, the device type Dzurak says is used in 99% of integrated circuits today. The company reduced the electrons in each transistor to a single electron or a few electrons, confined in nanoscale quantum dots within the silicon substrate. Information is encoded in electron spin, and neighboring dots interact to perform two-qubit logic gates.
Metal gate electrodes between the dots switch the interaction between qubits on and off. Diraq says it has developed a complete physical architecture that allows for quantum error correction. It also says it benefits from the most advanced semiconductor process nodes and smaller feature sizes.
Diraq argues 20-30 nm footprint avoids superconducting scaling limits
Dzurak puts superconducting qubits at tens to hundreds of microns each, against 20-30 nm for Diraq's silicon qubits. In his assessment, packing and cooling a million superconducting qubits is a major obstacle. Scaling beyond a few thousand while maintaining entanglement across the chip and moving large volumes of data between qubits would be an enormous engineering challenge.
He adds that operating such a machine would require thousands of dilution refrigerators at massive cost. Diraq says its chips and cooling infrastructure are much smaller and cheaper, allowing for thousands of quantum computers in a data center alongside classical compute.
Fault tolerance requires millions of qubits for computation and error correction
Dzurak says a powerful quantum computer needs many qubits for the computation itself and many more for quantum error correction. He puts the required scale at around millions of qubits. Diraq's stated goal is the first fault-tolerant quantum computer, built on CMOS foundry manufacturing.
Chemistry is the first target application from 2028
Diraq expects its chips to reach thousands of qubits as early as 2028 and to target first applications, such as chemistry. Dzurak names molecular design as a domain where combining classical high-performance computing with quantum computing can deliver early value. Once full error correction is in place, he expects the machines to address a wider range of problems, with universal fault-tolerant systems expected in the 2030s.