Quantum Technologies: Innovation and Investment

Q2B25 Silicon Valley, being held this week in Santa Clara, California, is billed as “The Roadmap to Quantum Value,” focusing on practical quantum technologies. This is the eighth consecutive year for the conference.

Quantum computing is not typically an area that I cover, although I do have one other previous article, Leaps in Quantum Computing, that I wrote about two years ago that focused more on the computational complexity implications and BQP.

There’s something exciting about attending an event that’s filled with a lot of hope for the future of quantum computing technology, and which has many small players and startups as well as large corporations exploring several different avenues of research.

One panel session focused on innovation and investment in quantum technologies. The panel was moderated by Peter Olcott, principal at  First Spark Ventures, and featured panelists Scott Buchholz, managing director at Deloitte Consulting; Jamie Garcia, director of strategic growth and quantum partnerships at IBM Quantum; and Raul Camposano, managing partner at Silicon Catalyst. The picture below shows the panelists along with Rich Curtin, managing partner at Silicon Catalyst, who delivered a short introduction before the discussion got started.

Panelists (Back row, left to right): Scott Bucholz, Jamie Garcia, Raul Camposano, Moderator: Paul Olcott, (Front) Rich Curtin

The panel focused on the fact that while new quantum technologies are advancing rapidly, their true value will only be realized through scalable, real-world applications. Discussing the interplay​ between hardware, software, and hybrid quantum–classical systems, as well as the role​ of public–private partnerships, attendees got strategic insights into how quantum​ technologies are moving from proof of concept to practical deployment, and what it will​ take to unlock their projected multi-trillion-dollar economic impact by the next decade.

The participants
First Spark is a deep tech venture capital fund investing in breakthrough technologies. Olcott said that he specializes in investments in AI, biotech, and quantum technologies.

Jamie Garcia has a Ph.D. in chemistry and started doing benchwork. She has a background in organic chemistry and polymers and applied that expertise at IBM to develop photoresist. As it turns out, some of those same materials are used for making qubits. Attending a talk on ground state energies at Yorktown Heights was an “aha” moment for Garcia on how she could connect her own chemistry work and quantum computing.  It was no longer just sci-fi to her, and she took a leap of faith and decided to become a part of it. So when an opportunity opened up at IBM as an operations manager for a computational chemistry and physics team, she took it.

Scott Buchholz’s father was an experimental high-energy physicist, and his mother was a biochemist. Buchholz said he sat through a lot of technical discussions during dinners growing up. Several years ago, he started seeing clients who were going into the quantum computing space. He talked with the company CTO, who agreed that Deloitte should get into this area. The CTO then came back a couple of weeks later with a green light to go ahead.

Raul Camposano started by saying that he’s not sure that he’s totally taken the leap of faith yet in quantum computing. Formerly, as CTO at Synopsys, he focused on how to make things run faster and tackled NP-hard problems. Camposano said he is fascinated by the possibility of exponentially accelerating computations. He read Paul Dirac’s book (The Principles of Quantum Mechanics), noting its description of state spaces in terms of superposition was the most useful book that he’s read in the field. Ten years ago he joined Silicon Catalyst, an incubator/accelerator that specializes in semiconductor companies and partners with other companies that contribute in-kind with tools and foundry shuttles. Silicon Catalyst hadn’t performed any quantum computing evaluations until recently, when they had 16 quantum computing companies apply. Silicon Catalyst hasn’t accepted any into the program yet. The firm has hundreds of advisors on board, and about 25 have a quantum computing background and expertise.

The state of quantum
Olcott noted that it’s hard to get investments, but encouraged people to keep trying.

He asked Garcia about the changes before and after she started working on quantum computing. She said hardware is progressing at a rate that couldn’t have been predicted. Technologies are set to appear in 2029 that many didn’t expect to see in their lifetimes.  But the team wants to see quantum computers that are actually being used for practical problems. In 2016, IBM put a five-qubit device on the cloud for anyone to use. Since then, IBM has deployed dozens of quantum computers around the world in places like Japan, Germany, Spain, and the U.S.  IBM chose superconducting qubits, based on speed and the ability to build them, but not necessarily toward solving chemistry problems. IBM’s next-generation Nighthawk is built for fault tolerance and degree 6 connectivity. Couplers for degree 6 connectivity are targeted for the end of this year.  Nighthawk is 120 qubits, and the next scale-up is to run 200 qubits and is targeted to be available by 2029 on the cloud (in a Poughkeepsie data center).  Goals are to get bigger and bigger and push down error rates.

Olcott asked Camposano to compare the state of quantum computing to his early experience with logic synthesis when it was introduced to the industry.

Camposano focused on the differences between software and hardware. One of the plenary speakers in an earlier session at the conference had identified which types of problems can be solved with quantum computing and then mapped those problems. All interesting problems are NP-hard, and currently tools use heuristics built on years of experience to address them.  With quantum computing, programming of the computers is more like designing complex circuits.  In the U.S., there are approximately 80,000 hardware designers and about 5 million software engineers. Camposano asked how many would write quantum computing software? Tools to enable quantum computing programming will be key, and it’s still an evolving field.

Olcott then asked Bucholz how to best communicate breakthrough advances and get companies ready?

Bucholz said one important factor is understanding that quantum computing is not a better version of today’s computing. Many view quantum computing as if it were some kind of new magic pixie dust that can be sprinkled on their problems to make solving them easier. He said it takes about two years to get competent, roughly about the same amount of time for someone in data sciences to reach competence. Roadmaps indicate that in the next one to three years, commercially relevant use cases will appear. Most of Deloitte’s customers are global 500 companies interested in solving today’s problems and need value today. Having customers ramp up on ML solutions that work today on today’s classical hardware also can be a pathway to getting up to speed on quantum computing.

Olcott noted that businesses tend to work on fear or greed. Is Q-Day, or some optimization problem the driver? What are the kinds of strategies used when pitching here?

Bucholz said you can view this as purchasing a call option if you think that quantum computing might take off, or as an insurance policy if you don’t think it’s going anywhere but don’t want competitors to get too far ahead in case it does. The other choices are to wait or just ignore it. On cybersecurity, fear may not be a driver, but more need to get involved. Some believe the cloud providers will take care of it, but cloud providers don’t always view it as just their problem. Quantum secure techniques are relevant now so that any security solutions currently under development have a lifetime into the future.

Olcott asked Camposano if he had any suggestions for innovators looking to get into this space.

Camposano responded that it’s hard to believe a startup can handle the whole stack, so small companies should focus on a problem while recognizing it also may be integrated into a total solution. The Danish 55North is putting in money, while the University of Munich and Silicon Catalyst are providing incubator opportunities to help startups. Staying connected and focusing on one problem that can be integrated into a total solution is key.

Olcott asked Garcia about all the different possible approaches using photons, ions, superconductors, etc., and for a breakdown.

Garcia said fault tolerance is important to actualize many of the algorithms. Applications like cryptography depend on fault tolerance and a certain number of qubits. Over the last 12 to 18 months, the focus has shifted to locally decodable codes (LDC codes as opposed to surface codes), the number of parameters, how many qubits, and how to lay out the qubits to get to a solution.  This is so that what you’re running now and between 2029, the API will be the same, even though the 2029 quantum computer will have more qubits, gates, and support for larger scales. That’s also why fault tolerance is important (also true outside of quantum as systems get larger).  It’s just much earlier days in quantum computing. Qubits are sensitive, and that’s why they are used.  Temperature, radiation, and other noise sources can have an impact, but today we still can see quantum computing utility. That’s still useful in situations where it just gives an answer that can provide insight into your problem. Fault tolerance is important. The applications, circuits, AI, and AI-HPC are important as we march toward quantum advantage. Quantum computing will be coupled with HPC.

Olcott asked if fault-tolerance can be made visible to the user.

Garcia responded that there’s interesting research on how to balance between QPU, CPU and GPU. IBM is partnering with AMD on quantum computer integration. It’s not yet at scale, but it’s an interesting new avenue to explore in this heterogeneous architecture.

Final questions
An audience member asked Garcia whether using IBM quantum for quantum computing finance software is useful?  He said that Nighthawk saw a large jump in available depth.  How do we know what to expect going forward?

Garcia said that Heron (degree 4) is going to degree 6 for Nighthawk and IBM is working on couplers.  On the innovation roadmap, users will see increases in qubits and number of gates that can be run while they continue to push the errors down.  As new capabilities are developed, they will be integrated into the system.  Users will be able to run more gates, more types of gates, and in 2029 there will be fault tolerance, all “under the hood” from the user’s viewpoint.

Curtin asked the final question to the panel to wrap up the session. He asked whether you should focus on hardware or software. Bucholz responded that as a long-time software guy, the answer is apparent. You can work backward or forward to get there.

Garcia noted that her CEO says 90% of the value in quantum computing is going to come from the applications. Algorithm development is important, and recently Shor’s algorithm being reduced by some clever reorganization helps illustrate that case. Hybrid architectures, software, applications, and algorithms will all play important roles.

Olcott said that addressing hardware scaling loss is important, and he sees a possible bottleneck five to six years out and getting to a depth of millions. Their bet is on the hardware side.

Camposano said he is on the software side. He believes that there will be huge progress in emulation of quantum computing. Quantum computing simulations are already available in 5,000 lines of Java to help enablement, so definitely software.