Quantum Ready — Part 4 — Buy, Build or Borrow?
The Questions to Ask About Building Your Own Quantum Tech Stack
Earlier this year, I embarked on a consulting engagement with four scaling companies in Canada. Each had either identified a use case or were open to leveraging artificial intelligence within their business, resulting in use cases or applications that were both customer-facing and operational. I introduced each to my companies’ proprietary “Artificial Intelligence Opportunity Assessment”1 incorporating matrices, questions, and graphics driving both a broad understanding of the opportunities and a narrowing of their priorities. The outcome? Some AI would be “built” or developed in-house, while in other cases partners or vendors would be approached. Because of the exercise, all these companies are better positioned to integrate AI now and in the future. One founder commented that the process “has helped us define a strategy where we can apply practical AI technology to solve some near-term tactical needs and over time extend our use to create a significant competitive differentiator.”
For one of these clients, we began looking out a couple of years from now, post-implementing a series of AI tactics, into leveraging quantum computing (QC). The use case? Natural language processing (or NLP). So, the conversation led us into the question of, if you start down the path, would you “buy, build or borrow” when it came to QC? Simple, one might say, at this stage of quantum computing maturity. Surely some would argue to not “build” when one combines the relatively immature or unproven technology and the cost to develop one’s own.
The ‘Quantum Ready’ series of articles have been devoted to those organizational leaders and managers entering into the world of quantum computing, for the first time, or those seeking to dig deeper with some early level of insight into this relatively new technology. In order to map out their unique quantum path, organizations need to consider a series of questions, and in Part 4 we articulate some of those questions (by no means all given space limitations and overwhelming our readers).
The Quantum Questions:
The questions surrounding the development of a quantum computing strategy for any business are numerous. This article reviews the “buy, borrow, build” decisions that organizational leaders will need to undertake as the quantum computing era evolves and accelerates. [For the purposes of this article, “Buy” refers to the purchase of quantum hardware, software or services (QHSS). “Borrow” refers to licensing schemes or alliances. “Build” refers to the internal development of some part of the QHSS tech stack.]
Is 2021 the right time to build a quantum computing full-stack solution in-house? Is it even feasible? Is your organization ready for a quantum computing ecosystem? Have you developed a short, medium and long-term quantum computing strategy? Do you have the right talent on-board? How about trusted partners to deliver on your strategy? Do you have a complex business problem that aligns with current quantum capabilities like optimization? Can you integrate the results of the solution into an action for the business i.e. Is it useful? Does it survive a rigorous cost-benefit analysis?
And exactly what would you want to “buy, borrow, build”? Before we supplement the discussion with additional questions, let’s consider the quantum stack.
The Quantum Stack:
Organizations considering quantum computing as an opportunity need to think about where within the quantum tech stack they will “play”. Each organization must select where their core capabilities exist and where it would be more cost-effective or efficient to invest their efforts, keeping in mind the relative immaturity of the quantum computing field.
Doug Finke, managing editor of the Quantum Computing Report, shared an excellent visual of the quantum tech stack at a recent Quantum London gathering, the contents of which create the foundation for any buy, borrow, build decision.
At the bottom or foundation of the tech stack is the quantum chip, also referred to as quantum processors or QPU’s (quantum processing units). According to Science Alert, QPU’s “perform calculations based on the probability of an object’s state before it is measured — instead of just 1s or 0s — which means they have the potential to process exponentially more data compared to classical computers”. Circuit-based QPU’s include IBM’s superconducting chips which have held such names as Armonk, Ourense, Vigo, London, Burlington, Essex, etc. Annealing QPU’s include D-Wave’s series (2X, 2000Q, Advantage). Photonic or light-based architecture includes Xanadu’s X8, X12, X24. Honeywell and IonQ produce ion trap QPU’s. Released over time, each processor varies in fidelity, speed and qubit capacity.
In terms of hardware, we will explore one example namely quantum annealing. D-Wave, the Burnaby, BC-based manufacturer of the Advantage quantum computer offers tours of their quantum labs and shares videos of annealing computer technology. Their quantum hardware packaging occurs where deep cold removes noise, hard vacuums remove gases, and Farraday cages remove electromagnetic disturbances. As I described in Part 3 of this series, D-Wave’s quantum chip operates in an environment of 15 milliKelvins or less above absolute zero at the bottom of the ‘chandelier’ hardware with several stages above that reducing the temperature from 50 Kelvins (K) to 3K to 1 K and so on. Shielding in the form of a ‘vacuum can’ operates around the system to eliminate external gases/gas molecules from the dilution fridge/chamber. Electromagnetic interference or disturbances are eliminated by a Farraday cage. This is just one example of a quantum computer hardware configuration. Photonics (light, laser generated computing) hardware, on the other hand, does not have the same cooling requirements. So, making an informed decision about the type of quantum hardware can depend upon the diversity of problems they solve, the fidelity (or accuracy) of the output, the speed of the system, the cost of time on the quantum server, etc.
Transpilers come next in the stack. Amelie Schrieber, writer in Towards Data Science, describes transpilers as source-to-source compilers which “take one version of code, optimize it in some way, perhaps by shortening the code, reducing the amount of memory needed to run the code, or reducing the runtime of the code, and the new optimized is the output of the transpiler.” Qiskit defines transpilation as “the process of rewriting a given input circuit to match the topoplogy of a specific quantum device, and/or to optimize the circuit for execution on present day noisy quantum systems.”
Control electronics firmware is the next highest level, being either a standard operating environment for more complex device software or as a complete OS (control, monitoring, data manipulation) for less sophisticated devices. Companies like Q-CTRL provide “a set of protocols that connect quantum hardware with higher, more abstract levels in the quantum computing stack… to bridge the divide between quantum algorithms and imperfect hardware”.
The software framework rests above the transpiler providing generic, reusable functionality which can be changed through additional user-written code, translating into software that is application specific. Specific examples include code libraries, compilers, and API’s that bring components together. Quantum Machines of Israel, for example, promotes its Quantum Orchestration Platform for “any company or institution developing quantum processors…can now purchase the QOP to run the most complex algorithms possible”. Their software interface/quantum assembler (QUA) translates classical code “into a quantum assembler language that can then be run on any quantum processor.”
Software routine libraries are “debugged blocks of code (subroutine, procedure, function etc.), often designed to handle commonly occurring problems or tasks” per ComputerScience.GCSE.GURU. Quantum software libraries include IBM’s Qiskit, Microsoft’s Quantum Development Kit Chemistry Library, TensorFlow’s and Baidu Paddle QML versions.
Cloud infrastructure, by no means a new technology, incorporates the components required for cloud computing such as hardware, storage, and networking enabling hosting services.
Education and User Communities, like those offered by quantumcomputing.com, IEEE, IBM, Waterloo’s Institute for Quantum Computing, Atos, and others bring together customers and users of various quantum ecosystems. Such communities encourage users to share their experiences, reviews and insights independent of quantum vendors or suppliers.
So, that’s the tech stack. Yet, the answer to the “buy, build or borrow” decision of course is “it depends”. The vendor market for quantum hardware and software is growing yet limited in 2021. Quantum hardware vendors (IBM, Rigetti, D-Wave, Honeywell, Google, Microsoft, AWS, etc.) with viable solutions do exist. So, why would a commercial entity attempt to build their own multi-million-dollar quantum computer when several viable options already exist? Where will you invest internally versus externally to be “quantum ready”. Can you ignore the build option and rely on Quantum-as-a-Service at this stage of your quantum readiness?
Look at the statistics in Part 1 of this series if you’re at all uncertain about whether this technology has a practical future. So many quantum use cases and applications are already underway. Yet do you have to go it alone? Can your leadership team comfortably make these decisions today or do you need help navigating the relatively new technology field? Can you maintain and run the purchased infrastructure? Is this a specific business problem/one-off where a partnership or vendor arrangement is more appropriate? Or is this a design program that operates continually in the business and needs QC to be in-house to protect proprietary work?
These are merely a fraction of the questions associated with the ‘quantum ready’ decision. If you’re leading an organization that is ripe with quantum computing talent (a rarity today) you may be open to enveloping more of the quantum tech stack internally. If you are simply opening the door to QC, connecting with an objective source like the Quantum Computing Report or a local meetup community is a better starting point. If you are somewhere in between, reflect back on the questions I posed earlier and connect with us about our “Quantum Computing Opportunity Assessment”2. If you’re like most companies, you’re most likely to focus on Quantum-as-a-service incorporating the various elements above in a third-party full stack solution fueled by your company’s data, but these questions will collectively more closely guide your “buy, build or borrow” decision.
Conclusion:
This four-part series has covered an overview of the state of quantum computing in 2021, how to address the quantum workforce, insights into quantum noise and finally, how organizational management can think about the “buy, build or borrow” decision. Organizations have much to consider in the quantum computing world, yet a structured approach should enhance your probabilities of success.
I hope you have enjoyed the series. Reach out to us for comments or requests for future articles on quantum computing which appear on Brian’s media including LinkedIn, Medium.com, Quantum London and AquitaineInnovationAdvisors.com.
Author’s Note:
I’d like to thank my two editors on this article, namely Esperanza Cuenca Gomez and Stephanie Holko for their insights and challenging me to provide as robust a review of the subject of quantum readiness as possible.
Copyright 2020, 2021 Aquitaine Innovation Advisors
1,2. Contact Aquitaine Innovation Advisors at https://aquitaineinnovationadvisors.com
#quantumcomputing #artificialintelligence #quantuminternet #ai #aiforbusiness #quantumtech #technology #quantumworkforce #dhdp #digitaltwins
Brian’s upcoming book “Quantum Boost: Using Quantum Computing to Supercharge Your Business” is available for pre-order on Amazon.com, Amazon.co.uk, and Amazon.ca
Brian Lenahan is the author of four Amazon-published books on artificial intelligence including the Bestseller “Artificial Intelligence: Foundations for Business Leaders and Consultants”. He is a former executive in a Top 10 North American bank, a University Instructor, and mentors innovative companies in the Halton and Hamilton areas. Brian’s training in quantum computers comes from CERN/University of Oviedo, and Technische Universiteit Delft, and he writes extensively on quantum computing strategy. His new book “Quantum Boost: Using Quantum Computing to Supercharge Your Business” will be released in early 2021.
Email: ceo@aquitaineinnovationadvisors.com
Aquitaine Innovation Advisors: www.aquitaineinnovationadvisors.com
LinkedIn: https://www.linkedin.com/in/brian-lenahan-innovation/
