Quantum computers have the potential to revolutionize information technology, processing problems that even our most powerful supercomputers can’t solve.  However, those problems require quantum computers that have millions of qubits (quantum bits), while today's quantum computers operate on a 100-qubit scale.

Connecting more quantum microchips could help overcome the limitations of today’s machines, but they are limited in speed and fidelity. The fastest rate that has ever worked was 180 qubits moved per second with a success rate (fidelity) of about 94 percent. Bringing the fidelity up closer to 100 percent leads to a drop in speed. Researchers from the University of Sussex and Universal Quantum have now demonstrated a new approach to this problem which has broken those records in the most dramatic fashion. 

The teams have designed a new way to connect microchips, that they liken to a jigsaw. In a regular microchip, all the exciting stuff happens in the middle and you can hold it by its side. In their version, the edges are the key parts, which allow them to be connected and transfer qubits with a success rate of 99.999993 percent and a speed of 2,424 per second. 

Their microchip has electrodes placed at its edge that are so good they can control a single atom. Aligning these overhanging electrodes is key to connecting the microchips with incredible success, and they can be aligned on each edge – so there is no limit on how many microchips you can connect together.  

“We kind of change the way you scale quantum computing by coming up with a solution that is as simple as a puzzle you play at home,” senior author Professor Winfried Hensinger told IFLScience. “That enables you to essentially do any arbitrary computation with as many qubits as you like and as complicated as you like. It's a fundamental step change of how we scale quantum computing.”

The fidelity has an error so small that you don’t have to correct for it anymore, and the speed is a whole order of magnitude higher than the current approach, known as photonic interconnect. This method also has room to improve further, although it’s more than good enough to be employed right here, right now.

“We can probably add another order of magnitude [to the rate]. But this is good enough for fault-tolerant quantum computing. This is not a proof of principle or an 'in the future we can…' These numbers are sufficient right now, as they stand. You don't need to change anything,” Professor Hensinger explained to IFLScience. 

This major breakthrough is reported in Nature Communications.