Looking ahead is rarely a good idea: the act of observation itself tends to collapse probabilities into smaller and smaller feasible options. But the future of an estimated $200 billion market by 2040 must nevertheless be looked at with intense scrutiny – plans and funding on groundbreaking yet specialized technology, such as quantum computing, demands that we do.
This isn’t a technical article – we won’t be taking any deep dives into many of the technologies mentioned, only briefly describing them. But this should serve as a good starting point for looking at the overall quantum technological landscape and its possible developments. We’ll be looking at some (but not all) of the current quantum computing approaches that have had the most promising and consistent developments – their futures inked into corporate roadmaps.
What this roadmap covers
Our approach is structured according to technology “families” among companies with deliverable products or services – think the superconducting qubits we’ve come to associate with IBM, Google, and Rigetti; trapped-ions (which have seen the most solid bets from IonQ and Quantinuum); quantum annealing and its particular track-record in optimization problems (D-Wave); neutral atom tech (Atom Computing, QuEra); and photonics (Xanadu, PsiQuantum).
Then, we cover research that has yet to bear fruit by looking at Intel’s work with silicon spin qubits, and Microsoft’s particularly embryonic topological qubits.
For this article, we’ll only look at superconducting qubits (as interpreted by IBM and Google) and trapped-ion qubits (as designed by IonQ and Quantinuum). Quantum is better taken in slices.
One thing to keep in mind throughout is that, like operating frequencies in processors, which aren’t a direct measure of performance, qubit quantity isn’t the be-all-end-all of quantum computing. The quality of qubits matters more than their quantity, even if quantity does improve quality up to a point.
Superconducting Qubits
Superconducting qubits, as the proximity to “superconductor” implies, take advantage of certain material’s ability to conduct electrical currents with no resistance. Qubits can be built out of these materials through what is called a Josephson junction – essentially, two superconducting layers separated by an insulating, 1-2 nanometer barrier. This junction then induces the emergence of discrete energy levels, which can be used to represent differentiated states (information). Computing is simply what is done to those states, and measuring their outcomes.
Superconducting qubits have the benefit of being compatible(ish) with contemporary 300mm semiconductor wafer fabrication technologies, which significantly improves perspectives on scaling, even if it means that inter-qubit connectivity is an obstacle (think CPU bus designs, and related technologies). However, the required near absolute zero operating temperatures and the degree of hardware cost and maintenance complexity means that superconducting quantum systems tend to be offered to customers via cloud platforms moreso than direct hardware sales.
... continue reading