Australian quantum startup Diraq has moved its technology out of the lab and into a working commercial data centre, a step its founder believes could eventually help rein in the soaring electricity needs of artificial intelligence.
The company has installed a small quantum computer built on a silicon chip inside Equinix's Sydney facility. Diraq describes the installation as the first field test of its kind anywhere in the world.
Founder Andrew Dzurak argued, in an interview with AFP syndicated by KMVU FOX 26 Medford, that quantum machines will be ready to take on part of the data centre workload within five years.
Why AI's Power Problem Matters
The largest data centres used to train and run AI models now draw power on a gigawatt scale. Meeting that demand often means building new power plants or placing extra strain on existing grids, and local communities have increasingly pushed back.
Dzurak's case is that much of this consumption comes from a category of problems that conventional hardware handles inefficiently. In his view, many of the calculations next-generation AI systems are attempting would be far better handled by quantum processors.
Inside the Sydney Field Test
The machine now running in Sydney uses eight qubits on a silicon chip. Qubits are the basic units of quantum information and rely on quantum mechanical effects to store more data and perform certain calculations faster than classical bits.
Diraq's system is self-contained and consumes less than 20 kilowatts while operating next to a conventional computer at room temperature. The company plans to scale the installation gradually to 150,000 qubits by 2029.
Its longer-term goal is a two-million-qubit silicon chip by 2031. Dzurak said such a system would use roughly a couple of hundred kilowatts of power.
Beyond AI: Drugs, Batteries and Chemicals
Dzurak also pointed to applications outside AI, including fast, low-energy modelling of molecules for pharmaceuticals, lighter batteries and the chemical industry.
Diraq estimates that applying quantum computing to drug development could raise trial success rates by 10 percent and cut the cost of bringing a new medicine to market by a third, provided the computing runs for under $100 million a year. Cost remains the deciding factor. "It will only happen if we can build quantum computers that are cost-effective," Dzurak said.
A Place in DARPA's Quantum Race
Diraq is one of six companies worldwide that have reached the third stage of the US Defense Advanced Research Projects Agency's quantum benchmarking initiative. The program aims to determine whether any quantum computer can solve real-world problems at a cost that delivers value by 2033.
Diraq says it intends to meet DARPA's definition with its two-million-qubit system by 2031, two years before that deadline. Dzurak added that the AUKUS defence technology partnership between Australia, Britain and the United States has lowered security barriers to sharing quantum technology, which he described as essential to operating across Australia and the larger US market.
Australia's Growing Quantum Cluster
Diraq is not the only Australian name on DARPA's list. PsiQuantum, which is building a different type of quantum computer in Queensland with the support of a 1 billion Australian dollar government investment, began construction in June.
Meanwhile, Silicon Quantum Computing, founded by physicist Michelle Simmons, was recognised by Time magazine for one of the best inventions of 2026, a quantum-enhanced AI chip. Dzurak described the country as going through a Silicon Valley moment in the field.
Cathy Foley, president of the Australian Academy of Technological Sciences and Engineering, said Australia now hosts more than 50 quantum-related businesses. She credited decades of scientific leadership partly to investment in university spinouts, while acknowledging that turning discoveries into successful companies remains one of the country's biggest challenges.
What the Test Signals for Founders and Investors
Placing quantum hardware in a commercial data centre is a meaningful shift from academic demonstration toward real operating conditions. It gives Diraq a chance to show how its silicon approach performs alongside conventional infrastructure.
The bigger questions are scale and economics. Diraq's roadmap requires a leap from eight qubits to millions within a few years, and both its own leadership and Australian industry figures stress that commercial viability, not scientific promise alone, will decide whether quantum computing meaningfully eases AI's energy burden.
Source: KMVU FOX 26 Medford
