A chip off the quantum block: Microsoft's breakthrough reshapes computing

In the past two years, Artificial Intelligence (AI) has seen breakthroughs turning esoteric research into practical applications. Quantum computing (QC) remains an esoteric field but a new breakthrough from Microsoft may change that.  

QC exploits quantum properties to develop exponentially faster computational capacity. Computing converts everything into binary numbers — current on is “one” and current off, “zero”. A conventional bit can store (and process) either zero or one. 

A quantum bit or qubit can store both values at the same time. When you put an array of qubits together, it can process data exponentially faster than much larger conventional arrays. 

  However, qubit-based computers have very high error rates.  Errors can be set off by heat, or minor vibrations, like those caused by a passing truck. 

  QC generally is done with very small arrays of qubits stored in very cold, physically stable environments. Assembling big arrays of qubits requires huge amounts of physical space in highly controlled environments. 

  Now Microsoft (MS) claims it has broken through several of the barriers imposed by physics. The new Majorana 1 chip from MS promises to scale up to clusters of millions of qubits with far more effective error correction.  Moreover, Majorana quantum chips could be used in normal data centres, reducing costs dramatically while making QC possible at scale. 

  The physics is fascinating. Back in the 1930s, Ettore Majorana, a brilliant young Italian physicist theorised that a certain type of subatomic particle was possible. 

  Majorana particles are their own antiparticle. These are not found in nature but complicated experiments proved their existence in the early 2000s.

  Importantly, a version of Majorana material can be created inside superconductors. This exotic material can be insulated by using topological superconductors, which have the property of being superconductive on one outer layer and insulated inside, creating a two-in-one material that is both superconductor and semiconductor. Like Majorana particles, topoconductors existed in theory for many years before experimental verification and discovery. 

  This combination of exotic materials may be used to create qubit arrays. Microsoft claims that its Majorana chip, which fits physically in one hand, may contain a million qubits plus the control electronics. 

  The chip is kept within a dilution refrigerator that keeps qubits at very low temperatures and a software stack that integrates it with AI and classical computers. Microsoft has designed all those components in-house. 

  Microsoft also says it has worked out a precision measurement approach, which can detect the difference between one billion and one billion and one electrons in a superconducting wire.  

  Instead of silicon, the topoconductor is made of indium arsenide (a semiconductor) combined with aluminium (a conductor). The measurement system can be turned on and off with voltage pulses, like flicking a light switch, rather than requiring fine-tuning for each individual qubit. This digitally controlled approach simplifies quantum computing processes drastically. 

  Microsoft claims it is now on track to build a fault-tolerant prototype of a scalable quantum computer within “years, not decades”.  This would be part of the final phase of the Defense Advanced Research Projects Agency (DARPA) Underexplored Systems for Utility-Scale Quantum Computing (US2QC) programme.

  Microsoft outlined a scalable architecture built around a single-qubit device called a tetron.  Tetrons can be connected in arrays to create a million-qubit chip if all goes according to plan. 

  It could still take years before this proof of concept translates into actual QC chips at scale. Indeed, making the manufacturing process work at scale could itself require the use of QC to optimise the fabrication of exotic materials. But this is a big step forward as a proof of concept that QC chips can work at large scale. 

  QC can, as noted earlier, perform many more calculations than conventional machines. They could, for example, make current cryptography obsolete.  A QC may be able to break passwords by brute force. Allied to AI algorithms, QC could solve many currently intractable problems. 

  Many challenges in material science require billions of dollars in exhaustive experimental searches and lab experiments.  Instead, calculations on QC could reduce time and cost by simulation. 

  Apart from code-breaking, QC may be capable of the complex calculations required to make novel self-healing materials for repairing cracks in buildings and bridges, support sustainable agriculture, speed up chemical discovery, and improve the design of fusion reactors. There’s a real chance QC could make the jump to being a practical tool within a few years.