How much easier it is to be critical than to be correct.Benjamin Disraeli 19th Century Prime Minister of the UK
by David Shaw, Doug Finke, and Andr M. Knig
Scorpion Capital is an activist investor specializing in taking short positions in publicly traded stocks (and therefore stand to gain if the stock price moves down). They recently took such a position in IonQ, a leading trapped ion quantum computing hardware company. IonQ listed on the NYSE in October 2021 in a SPAC assisted floatation. SPACs themselves have been controversial in some circles, where they are viewed as a way to avoid the usual scrutiny of a traditional IPO process.
Scorpion issued a scathing short report aimed to move market sentiment against IonQ [1]. Its important to recall that this style of research does not aim to present a balanced picture or even a structured analysis. Its a scatter gun of bad things that might hurt the stock price. Some of these such as allegations of revenue irregularities are a matter specifically for the company, which has responded with its own press release [2].
However, some of the accusations could mislead debate around the wider quantum industry, and confuse investors more generally. We want to discuss those points here.
Academics have been talking about quantum computing for over forty years [3]. Richard Feynman first speculated about the idea in 1981 [4] and it was formalised by David Deutsch in 1985 [5]. Many would date progress on hardware to Alain Aspects famous experiments in 1982 [6]. Fidelities in the lab slowly improved, notably in the period 2008-2017 [7][10]. Activity has really intensified in the last three years with multiple demonstrations of beyond classical calculations [11][13] (albeit on artificial problems), and tentative logical qubit demonstrations [14][16]. Multiple commercial players have defined roadmaps to build large scale machines [17].
For a recent review of progress see Quantum Outlook 2022.
Just Toys? Up to about 50-60Q we can mostly simulate these quantum devices on conventional computers. In that sense everything less is a toy and the field often learns by working on toy problems. However, this is deadly serious R&D. The IonQ 11Q device deservedly debuted in Nature in 2019 [18]. In has continued to perform well in independent benchmarks [19]. But are any of the current generation of devices powerful enough for commercial computing applications? No. Scaling up is definitely required.
1+1=2? Such calculations are not a target use case for quantum computers (we dont try to do math using wind tunnels). Even so, to a casual observer this might seem like this should be easy for any computing device. It turns out that this isnt necessarily so when working with low-depth NISQ circuits and todays gate sets [20]. Todays hybrid algorithms aim to leave as much work as possible on classical hardware.
Even with envisaged intermediate scale quantum machines, early commercial applications are unproven. Many academics are sceptical (as we pay them to be), pointing to the difficulties facing known NISQ approaches. Some entrepreneurs are battening down for the long term. Others point to the tradition of constructive criticism driving innovation, and of commercial programs making jumps that defied traditional labs. Recent progress in AI is arguably a good example of the latter [21].
A publicly quoted company, with the need to publish results quarter to quarter is a challenging environment in which to manage such an evolving narrative. Only the largest companies can traditionally combine such emerging activities with a public listing.
There are hold-out quantum skeptics, and no one has a roadmap where the physics is completely de-risked (though some are closer than others). However, for the long-term, mushrooming government support around the world reflects the clear majority expert opinion this revolution is going to happen [22][24]. The only question is when.
Trapped ion technology is a technology quite alien to those used to the digital world.
Is it really a 32Q device? This may sound like a fairly straightforward question. But the nature of trapped ion technology blurs the answer. The qubits are individual ions, and you can load a variable number of ions into a typical trap. Some qubits may play an active role in the calculation, others supporting roles. A key question is how many qubits are available for use in the target algorithm? IonQs next generation device does seem to have successfully operated with up to about 21Q in 2021 when it did well in independent benchmark tests [19] (though it clearly wasnt performing at the aggressive targets IonQ had set of 32Q and 4M QV).
Trapped ion quantum computers confound our expectations in many ways. They are set to have very slow gate speeds compared to conventional computers. The point is that quantum computers enable us to use algorithms that complete in exponentially fewer processing steps. Raw trapped ion gate speeds are also set to be slow compared even to other proposed quantum platforms. Here a true comparison is much more subtle. What really matters is which platform can achieve the desired combination of scale, fidelity and speed, and how along the way it keeps down the overheads associated with error correction [25].
Outsourced fabrication Its not necessarily an issue if a trapped ion player outsources the fabrication of the trap and vacuum systems. A basic trap is now a relatively standard component. The really challenging part of the setup for a machine like this is the laser system and the control logic. None of todays commercial players have yet fully recreated the sector-leading 2Q gate fidelities achieved with hero devices in the lab. Trap design and fabrication is set to become more of a focus as players innovate to meet other scaling challenges: miniaturisation and modularisation.
Miniaturization A key challenge for conventional trapped ion setups is gate control. Established approaches use lasers to drive gates. This makes miniaturisation a real challenge. AQT already have a rack-based trapped ion system, but they use optical trapped ion qubits [26]. These require a less demanding setup. Other trapped ion players are typically using hyperfine trapped ion qubits, which in principle offer longer lifetimes and so access to higher fidelities. But working with the special laser setups required looks harder. True large-scale trapped ion systems probably require integrated photonic solutions (if you stick with lasers some are working on ways to control ion traps with microwaves instead [10], [27]). Such systems are at an early stage as conventional photonic platforms dont work well at the required wavelengths [28]. Innovative solutions are emerging.
Modularization The other key scaling challenge is how to interconnect modules. Here trapped ion proponents often point to photonic interconnects. This is more of a challenge than sometimes portrayed. The currently best demonstrated fidelities and speeds dont look good enough [29]. Again, innovative solutions are emerging.
Working with trapped ion based approaches certainly are a bet that some better quantum technology isnt able to get over the line first. IonQ has to innovate to meet the scaling challenge. A positive from the SPAC is that it has an impressive $500M pile of cash to help it drive this process. And they do have ideas. The real question is how quickly can any player move to solve multiple challenges at once: fidelity, miniaturisation and modularity?
You have to be very careful with roadmap promises. Analysts on public equities wont be impressed when they change or are missed. Many R&D phase companies choose to stay private. Some choose to stay in stealth.
These arent old-time software startups where everyone can eat pizza and get things delivered by pulling an all-nighter. These are long term endeavours that have to combine skills from physicists, engineers and computer scientists just to make things work. In the real-world, marketing flair and commercial skills are set to be an equally important part of the mix. Combining such impenetrable disciplines amidst great uncertainty; mixing founding and new senior talent; retaining everyone around a realistic common company narrative (and realistic pay expectations) are going to be challenges many in the sector will face.
The pressures and dynamic of a SPAC process probably doesnt help keep everyone on board. Senior hires bouncing in and out never looks good.
Academic founders face particular challenges In many areas investors traditionally look for founders to fully commit to the new business. However, in the quantum space there are other considerations. Many anticipate that the talent pipeline will be a key issue. Keeping connections with a home institute helps shore-up a natural recruitment pool. It also gives insight into governmental programs of support for the local quantum sectors. An additional pressure is going to be managing to get the best out of academic and corporate lab teams. Each should have contrasting strengths, but also likely different cultures.
(In preparing this piece, a striking feature has been former academic colleagues, but now commercial competitors of IonQ founders Chris Monroe and Jungsang Kim jumping to their defense as physicists. Disagreements naturally continue on whose hardware plans are best.)
Existing and potential investors in quantum technology already face many distractions: an unsettled global economic environment; interest rates across developed economies are on an upward trajectory; inflation stalks the land. But where are investors to invest? Investing in innovative ventures is a vital opportunity to bring uncorrelated exposure into a portfolio.
First, we should point out that all companies have problems. Sometimes it is an engineering program that has slipped its schedule, sometimes it is disagreements within management, sometimes it is unhappy customers, and many other things. A person can certainly look at a company and write a report that only discusses these issues. But a report that only focusses on the bad things, but does not mention any of the good things happening at a company does not give an accurate picture. It also would be incorrect to assume that any problem a company currently faces will be permanent. But again, pointing out that a problem might be temporary wont be mentioned if your sole purpose is to write a report that drives down the price of the stock so you can make a profit.
We scorn the use of hype to create an unrealistic positive picture of how quickly quantum can add value. We equally scorn the use of scatter gun defamation (anti-hype) to paint an equally unrealistic negative picture. Neither benefits the industry or society in general. Both are traps for investors.
Quantum computing is at the very early stages of development and the ultimate proof of whether a company is good or not will be determined by whether a company can deliver on its roadmap and be competitive.
At this stage no one can say for sure which of the companies working in quantum can achieve this. The best that can be done is to bring in a well-qualified team that understands the technology, the market value chain, and real-world company cultures to perform careful due diligence. We dont think a hedge fund that doesnt have people knowledgeable about the technology, and has a motivation to be biased, fits this description.
Many think that SPAC mania in the financial markets is anyway coming to an end (and such vehicles will likely be more thoroughly regulated) [30]. But businesses should think not just about the route to flotation, SPAC or IPO, but also what milestones they need to hit to be ready for life on the public market. Venture investors will expect a plan that allows the business to move on at some point.
Quantum computing is engaged in a long marathon that will take many years to play out. The quantum technology sector overall presents an even wider landscape of opportunities. Business adopters should avoid immediate judgements, but engage with companies that can execute and bring to market competitive products that provide commercial value. Governments should encourage the creative destruction of the innovative process. Investors should weigh the best advice they can find, and make their choices. And IonQ needs to demonstrate that Scorpion Capitals criticismwas indeed nothing but aggressivefinancial posturing.
[1] Scorpion Capital | Activist short selling focused on publicly traded frauds and promotes, Scorpion Capital. https://scorpioncapital.com (accessed May 07, 2022).
[2] IonQ Reiterates Unwavering Commitment to Building the Quantum Future, May 04, 2022. https://www.businesswire.com/news/home/20220504006319/en/IonQ-Reiterates-Unwavering-Commitment-to-Building-the-Quantum-Future (accessed May 07, 2022).
[3]J. Preskill, Quantum computing 40 years later, arXiv:2106.10522 [quant-ph], Jun. 2021, Accessed: May 07, 2022. [Online]. Available: http://arxiv.org/abs/2106.10522
[4]R. P. Feynman, Simulating physics with computers, Int J Theor Phys, vol. 21, no. 6, pp. 467488, Jun. 1982, doi: 10.1007/BF02650179.
[5]Quantum theory, the ChurchTuring principle and the universal quantum computer | Proceedings of the Royal Society of London. A. Mathematical and Physical Sciences. https://royalsocietypublishing.org/doi/10.1098/rspa.1985.0070 (accessed May 07, 2022).
[6]A. Aspect, P. Grangier, and G. Roger, Experimental Realization of Einstein-Podolsky-Rosen-Bohm Gedankenexperiment: A New Violation of Bells Inequalities, Phys. Rev. Lett., vol. 49, no. 2, pp. 9194, Jul. 1982, doi: 10.1103/PhysRevLett.49.91.
[7]J. Benhelm, G. Kirchmair, C. F. Roos, and R. Blatt, Towards fault-tolerant quantum computing with trapped ions, Nature Phys, vol. 4, no. 6, pp. 463466, Jun. 2008, doi: 10.1038/nphys961.
[8]R. Barends et al., Superconducting quantum circuits at the surface code threshold for fault tolerance, Nature, vol. 508, no. 7497, Art. no. 7497, Apr. 2014, doi: 10.1038/nature13171.
[9]J. P. Gaebler et al., High-Fidelity Universal Gate Set for $^9$Be$^+$ Ion Qubits, Phys. Rev. Lett., vol. 117, no. 6, p. 060505, Aug. 2016, doi: 10.1103/PhysRevLett.117.060505.
[10]T. P. Harty, M. A. Sepiol, D. T. C. Allcock, C. J. Ballance, J. E. Tarlton, and D. M. Lucas, High-fidelity trapped-ion quantum logic using near-field microwaves, Phys. Rev. Lett., vol. 117, no. 14, p. 140501, Sep. 2016, doi: 10.1103/PhysRevLett.117.140501.
[11]F. Arute et al., Quantum supremacy using a programmable superconducting processor, Nature, vol. 574, no. 7779, Art. no. 7779, Oct. 2019, doi: 10.1038/s41586-019-1666-5.
[12]H.-S. Zhong et al., Phase-Programmable Gaussian Boson Sampling Using Stimulated Squeezed Light, Phys. Rev. Lett., vol. 127, no. 18, p. 180502, Oct. 2021, doi: 10.1103/PhysRevLett.127.180502.
[13]Y. Wu et al., Strong quantum computational advantage using a superconducting quantum processor, arXiv:2106.14734 [quant-ph], Jun. 2021, Accessed: Aug. 02, 2021. [Online]. Available: http://arxiv.org/abs/2106.14734
[14]L. Egan et al., Fault-tolerant control of an error-corrected qubit, Nature, pp. 16, Oct. 2021, doi: 10.1038/s41586-021-03928-y.
[15]C. Ryan-Anderson et al., Realization of real-time fault-tolerant quantum error correction, arXiv:2107.07505 [quant-ph], Jul. 2021, Accessed: Oct. 09, 2021. [Online]. Available: http://arxiv.org/abs/2107.07505
[16]L. Postler et al., Demonstration of fault-tolerant universal quantum gate operations, arXiv:2111.12654 [quant-ph], Nov. 2021, Accessed: Dec. 02, 2021. [Online]. Available: http://arxiv.org/abs/2111.12654
[17]Quantum Hardware Outlook 2022, Fact Based Insight, Dec. 13, 2021. https://www.factbasedinsight.com/quantum-hardware-outlook-2022/ (accessed Feb. 02, 2022).
[18]K. Wright et al., Benchmarking an 11-qubit quantum computer, Nature Communications, vol. 10, no. 1, Art. no. 1, Nov. 2019, doi: 10.1038/s41467-019-13534-2.
[19]T. Lubinski et al., Application-Oriented Performance Benchmarks for Quantum Computing, arXiv:2110.03137 [quant-ph], Oct. 2021, Accessed: Nov. 01, 2021. [Online]. Available: http://arxiv.org/abs/2110.03137
[20]agaitaarino, How do I add 1+1 using a quantum computer?, Quantum Computing Stack Exchange, Dec. 23, 2018. https://quantumcomputing.stackexchange.com/q/1654 (accessed May 05, 2022).
[21] J. M. Thornton, R. A. Laskowski, and N. Borkakoti, AlphaFold heralds a data-driven revolution in biology and medicine, Nat Med, vol. 27, no. 10, pp. 16661669, Oct. 2021, doi: 10.1038/s41591-021-01533-0.
[22]M. G. Raymer and C. Monroe, The US National Quantum Initiative, Quantum Sci. Technol., vol. 4, no. 2, p. 020504, Feb. 2019, doi: 10.1088/2058-9565/ab0441.
[23]T. Skordas and J. Mlynek, The Quantum Technologies Flagship: the story so far, and the quantum future ahead, Shaping Europes digital future European Commission, Oct. 16, 2020. https://ec.europa.eu/digital-single-market/en/blogposts/quantum-technologies-flagship-story-so-far-and-quantum-future-ahead (accessed Dec. 20, 2020).
[24]S. Mallapaty, Chinas five-year plan focuses on scientific self-reliance, Nature, vol. 591, no. 7850, pp. 353354, Mar. 2021, doi: 10.1038/d41586-021-00638-3.
[25]M. Webber, V. Elfving, S. Weidt, and W. K. Hensinger, The Impact of Hardware Specifications on Reaching Quantum Advantage in the Fault Tolerant Regime, arXiv:2108.12371 [quant-ph], Sep. 2021, Accessed: Sep. 30, 2021. [Online]. Available: http://arxiv.org/abs/2108.12371
[26]I. Pogorelov et al., A compact ion-trap quantum computing demonstrator, PRX Quantum, vol. 2, no. 2, p. 020343, Jun. 2021, doi: 10.1103/PRXQuantum.2.020343.
[27]B. Lekitsch et al., Blueprint for a microwave trapped ion quantum computer, Science Advances, vol. 3, no. 2, p. e1601540, Feb. 2017, doi: 10.1126/sciadv.1601540.
[28]D. Awschalom et al., Development of Quantum InterConnects for Next-Generation Information Technologies, PRX Quantum, vol. 2, no. 1, p. 017002, Feb. 2021, doi: 10.1103/PRXQuantum.2.017002.
[29]L. J. Stephenson et al., High-rate, high-fidelity entanglement of qubits across an elementary quantum network, Phys. Rev. Lett., vol. 124, no. 11, p. 110501, Mar. 2020, doi: 10.1103/PhysRevLett.124.110501.
[30]B. Masters, New reforms should stop failing Spacs in their tracks, Financial Times, Apr. 04, 2022. Accessed: May 07, 2022. [Online]. Available: https://www.ft.com/content/e38125db-caa6-40ce-b16e-1dc0745e1b48
May 7, 2022
More here:
Weathering the First Quantum Short - Quantum Computing Report
