What to do with a Quantum Computer: We still don’t know
Last year Google loudly proclaimed “quantum advantage” or – if you are feeling aggressive — “quantum supremacy.” This is a controversial point. But whatever your taste in terminology or views on what Google did, there can be little doubt that quantum computers have come a long way. One only has to go back a year or so to the days when earnest articles in quite respectable professional magazines said quantum computers would never happen to see our enormous progress.
The real questions now are (1) how quickly can we get our quantum machines to a point where they can do something useful; and (2) what could these useful things be?
Technology Determinants For Success
We don’t know the answers to these questions. But as a station on the way to these answers, we suggest that partial answers will be found in (a) what type of quantum computing technologies prevail in the medium-to-long term? And (b) how fast the Qubit processing capacity of commercial quantum computers? To an important extent (b) will depend on (a).
To date, there are many different ways to build Qubits. The Exhibit below contains a list of some of the main approaches and several of the companies using them. Our strong suspicion is that ten years from now this list will be shorter and perhaps different. If good room temperature machines prove themselves then superconducting machines will disappear. It is unlikely that the market will really need many different kinds of technology – one or two (perhaps one) will become the technology of choice for end users.
|Type of qubit||Companies building quantum computers with the technology|
|Superconducting||IBM, Google, Rigetti Computing|
|Trapped Ion||ionQ, Honeywell|
Source: Inside Quantum Technology
How fast will Qubit capacities grow: The type of quantum technologies that prevail will help shape and will in turn be shaped by the core technologies that are selected. But in the end, the underlying technology doesn’t matter nearly as much as the results produced. These results are heavily dependent on the number of Qubits along with error rates. Writing actual software and algorithms for a quantum computer is a relatively new field – there is much left to explore. However, one thing is certain – many more Qubits are needed than the 50-60 that are available in the top existing machines to fulfill the promise of quantum computers.
Several different companies have made statements regarding their goals and roadmaps and judging by these statements it seems reasonable that quantum computers will become useful within 4 to 5 years. It is also worth noting that at this very early stage of the game, corporate statements on such things are often bravado and should be taken with a pinch of salt. The very different prognoses made by different vendors at the present time will be revised many times in the coming years, we suspect. See the Exhibit below.
|Company||Statements on Capacity Growth|
|Honeywell:||Honeywell has also reached a quantum volume of 64 and has made far more ambitious statements – increasing that volume 10 times each year, which presumably means Honeywell will hit a volume of 640,000 in four years|
|IBM||IBM plans to double its “quantum volume” (a metric combining qubit count and error rates) every year, and so far has been even slightly ahead of schedule, hitting 64 in August this year. At this rate, it could hit the 1,000 mark in four years|
|PsiQuantum||PsiQuantum is a company working on a photonic quantum computer with a very ambitious goal of a million qubits for their first machine|
Source: Inside Quantum Technology
Conversely, since quantum computers are still in their nascent state, there is also definitely room for new game-changing breakthroughs. Eventually, Inside Quantum Technology expects that an analogous version of Moore’s Law will eventually emerge for Qubits, but probably not as rigorous as Moore’s Law once was. With several hundred or thousands of Qubits, so-called NISQ (Noisy Intermediate-Scale Quantum) technology will become available.
What Comes Next?
Several algorithms have been proposed that will speedup certain applications on quantum computers in comparison to classical analogs and these will be reflected – again certainly within the 4-5 year timeframe – in a large number of practical applications, which we will discuss in a blog yet to come.
For more in-depth discussions on quantum computers and other quantum technologies be sure to participate in the Inside Quantum Technology Online Conference on October 26-30, 2020.