Quantum software industry, is facing a subtle turning point ......

Over the past year or so, the quantum software industry has taken some subtle turns.

Many quantum software companies were founded as "pure" quantum software companies and said at the time that working on classical-based solutions would shift their focus. Now, however, many of them are introducing classical-based solutions (sometimes called quantum-inspired) as part of their product portfolio.

For example, QC Ware's Prometheum quantum chemistry application, SandboxAQ has developed a quantum simulation and optimization solution that will run on GPUs (graphics processing units) or TPUs (tensor processing units), Zapata is promoting their use of generative AI, Strangeworks has said they will soon launch new tools that leverage AI technology, and 1Qbit is developing a new tool that leverages artificial intelligence. and 1Qbit has developed a "black box" approach to reinforcement learning with simulated annealing (SA) ......

Prometheum GPU-based implementation

Zapata's Orquestra is based on generative AI

To be clear, these are not hybrid classical/quantum algorithms, but purely classical (albeit innovative classical) solutions.

For these companies, these are pragmatic changes, and they make a lot of sense. End users usually don't care how their problem is solved - as long as it is solved. If a classic-based solution gets there first, that will be their first choice. For software companies, this will bring in revenue faster and make their investors happier.

And at the same time, they will be able to build relationships with their end users and better understand the problems they are trying to solve. More often than not, these companies are working on solutions based on algorithms such as tensor networks: these algorithms can be implemented on GPUs first, and then relatively easily updated to run on real quantum computers once more powerful machines are available.

One thing to point out is that while companies are using many of the classic underlying solutions today, many are not using the most modern or efficient solutions that may have been developed years ago (although there are probably many organizations still trying to use Excel spreadsheets to solve various optimization problems).

Therefore, moving them to our best and most efficient modern classical algorithms still provides significant value. So while we wait for more powerful quantum computers to emerge and make purely quantum or hybrid quantum/classical based solutions feasible, this is not a bad way to go.

Reference links:

[1] https://www.promethium.qcware.com/

[2]https://www.sandboxaq.com/solutions/quantum-simulation

[3]https://www.zapatacomputing.com/

[4]https://1qbit.com/our-thinking/

[5]https://quantumcomputingreport.com/subtle-pivots-in-the-quantum-software-industry/