Canada's Xanadu achieves worldwide first with error-resistant quantum chip

Globe and Mail

2 days ago

Toronto startup Xanadu Quantum Technologies Inc. is reporting a new milestone in the effort to develop a form of light-based quantum computing that can operate at commercial scale.
For the first time anywhere, Xanadu researchers have created a single chip that embodies a powerful type of error-detection code in a pulse of laser light.
If a number of such chips could be harnessed together, it would open the door to a quantum computer that can deliver reliable results with practical value.
'This is something that's been on our roadmap for a long time,' Zachary Vernon, Xanadu's chief technology officer for hardware, told The Globe and Mail.
A technical description of the chip was published Wednesday in the journal Nature.
The development is significant 'because the chip platform is supposed to be scalable,' said Daniel Soh, an associate professor of optical science at the University of Arizona in Tucson.
'In the future, we will need millions or billions of this kind of devices on a chip. This result is a massive step towards that goal,' said Dr. Soh, who is not affiliated with Xanadu.
Canada 'a sweet spot' for growing quantum computing industry, expert says
Christian Weedbrook, Xanadu's founder and chief executive officer, said the development means it is possible to envision a quantum-computing system operating at the scale of a data centre, with some 5,000 servers fitting into a facility less than 10,000 square metres in size.
'We're also thinking ahead to how we can add more density in there, so that'll change,' he said.
Earlier this year Xanadu published a result showing how its form of quantum computing could be easily modularized. This latest step is aimed at making a machine large enough to solve relevant problems but not so large that it becomes impractical for commercial purposes.
It is the latest example of a shift in the focus and tempo of advancements in the quantum computing world.
Overall, the goal remains to create a computer that runs on qubits – interconnected physical elements that exhibit quantum behaviour – instead of the standard bits of a conventional digital system.
Where a bit can be used to represent a one or a zero in a mathematical calculation, a qubit can be a mixture of both. This dual nature, when combined with many other qubits, is what allows a quantum computer, in principle, to vastly outperform a conventional computer at certain kinds of calculations that are important for data security and other applications.
While various companies, including Google, IBM and Microsoft, have experimented with different types of qubits, all of them face the same challenge: Quantum systems are sensitive to disturbance and difficult to isolate from the rest of the world, which makes quantum computers especially error-prone.
To counter this, qubits can be linked to check each other for signs of failure during a calculation. But the price for such redundancy is that many more qubits are needed to build a reliable computer powerful enough to solve real-world problems.
More recently, teams have sought to exploit various mathematical codes, which are ways of tying qubits together, to make error correction more robust.
Of particular interest are Gottesman-Kitaev-Preskill (GKP) codes. First proposed in 2001, they are challenging to implement but especially amenable for quantum computer builders such as Xanadu, whose machines use qubits made of light moving through a fibre-optic network.
Xanadu's new chip corrals incoming particles of light, called photons, into a quantum state that allows them to work together to form a GKP qubit. The chip has four outputs, three of which are connected to detectors that can reveal whether the fourth is in a state that would allow it to be useful for a quantum calculation.
In a working quantum computer, such chips would provide an initial layer of error detection that would then be further augmented by other error-correction techniques when chips are combined.
Similar strategies are being explored by other companies. Last week, Nord Quantique, based in Sherbrooke, Que., demonstrated that it had successfully encoded microwave photons bouncing around inside a metal cavity with a GKP code.
Meanwhile, Xanadu still has more obstacles to overcome. Chief among them is finding ways to overcome signal loss, which occurs when photons are absorbed by the materials they are moving through.
In addition to making its light-based technology work, Xanadu and direct competitors such as PsiQuantum, Corp. of Palo Alto, Calif., are racing against big tech companies developing computers with qubits that rely on special superconducting materials kept at extremely cold temperatures.
Light-based systems offer a different set of advantages, including the fact that they can operate at room temperature.
While no system has yet emerged as a clear winner, Dr. Soh says light-based quantum computers may end up inching ahead because once the key technical challenges are solved, they will be easier to scale up.