With Jason Potts, first published on Medium

What happens when quantum computing is added to the digital economy?

The economics of quantum computing starts from a simple observation: in a world where search is cheaper, more search will be consumed. Quantum computing offers potentially dramatic increases in the ability to search through data. Searching through an unstructured list of 1,000,000 entries, a ‘classical’ computer would take 1,000,000 steps. For a mature quantum computer, the search would only require about 1,000 steps. 

Bova et al. (2021) describe this capability generally as a potential advantage at solving combinatorics problems. The goal in combinatorics problems is often to search through all possible arrangements of a set of items to find a specific arrangement that meets certain criteria. While the cost of error correction or quantum architecture might erode the advantage quantum computers have in search, this is more likely to be an engineering hurdle to be overcome than a permanent constraint.

Economics focuses on exchange. To our knowledge no analysis of the economic impact of quantum computing has been focused on the effect that quantum computing has on the practice and process of exchange. Where there have been estimates of the economic benefits of quantum computing, those analyses have focused on the possibility that this technology might increase production through scientific discovery or by making production processes more efficient (for example by solving optimisation problems). So what impact will more search have on exchange?

In economics, search is a transaction cost (Stigler 1961, Roth 1982) that raises the cost of mutually beneficial exchange. Buyers have to search for potential sellers and vice versa. Unsurprisingly, much economic organisation is structured around reducing search costs. Indeed, it is the reduction of search costs that structures the digital platform economy. Multi-sided markets like eBay match buyers with sellers at global scale, allowing for trades to occur that would not be possible otherwise due to the high cost of search.

Quantum computing offers a massive reduction in this form of transactions cost. And all else being equal, we can expect that a massive reduction in search costs would have a correspondingly large effect on the structure of economic activity. For example, search costs are one reason that firms (and sub-firm economic agents like individuals) prefer to own resources rather than access them over the market. When you have your own asset, it is quicker to utilise that asset than seeking a market counterpart who will rent it to you. 

Lowering search costs favours outsourcing rather than ownership (‘buy’ in the market, rather than ‘make’ inhouse). Lower search costs have a globalising effect — it allows economic actors to do more search — that is, explore a wider space for potential exchange. This has the effect of increasing the size of the market, which (as Adam Smith tells us), increases specialisation and the gains from trade. In this way, quantum computing powers economic growth.

Typically specialisation and globalisation increases the winner-take-all effect — outsized gains to economic actors at the top of their professions. However, a countervailing mechanism is that cheaper search also widens the opportunities to undercut superstar actors. This suggests an important implication of greater search on global inequality — it is easier to identify resources outside a local area. That should reduce rents and result in more producers (ie workers) receiving the marginal product of their labour as determined by global prices, rather than local prices. In this way, quantum computing drives economic efficiency.

Quantum and the digital stack

Of course other transactions costs (the cost of making the exchange, the cost of contract enforcement etc), can reduce the opportunities for faster search to disrupt existing patterns of economic activity. Here we argue that quantum is particularly effective in an environment of digital (or digitised) trade and production — in the domain of the information economy. 

The process of digitisation is the process of creating more economic objects and through the use of distributed ledgers and digital twins, forming more and more precise property rights regimes. In Berg et al (2018), we explored one of the implications of this explosion in objects with precisely defined property rights. We argued that the increasingly precise and security digital property rights over objects would allow artificially intelligent agents to trade assets on behalf of their users, facilitating barter-like exchanges and allowing a greater variety of assets to be used as ‘money’. Key to achieving this goal is deep search across a vast matrix of assets, where the optimal path between two assets has to be calculated according to the pre-defined preferences not only of the agents making the exchange, but of each of the holders of the assets that form the paths. 

This illuminates one of the ways in which quantum interacts with the web3 tech stack. While some quantum computation scientists have identified the opportunity for quantum to be used in AI training, we see the opportunity for quantum to be used by AI agents to search for exchange with other AI agents; an exchange theoretic rather than production-centric understanding of quantum’s contribution to the economy. The massive technological change we are experiencing is both cumulative and non-sequential — rapid developments in other parts of the tech stack further drive demand for the quantum compute. This is the digital quantum flywheel effect.

Compute as a commodity

Compute is a commodity and follows the rules of commodity economics. Just as buyers of coal or electricity are ultimately buying the energy embodied in those goods, buyers of compute are ultimately buying a reduction in the time it takes to perform a computational task (Davies 2004). There are computational tasks where classical computers are superior (either cheaper or faster), where quantum computers are superior (or could be superior), and those where both quantum and classical computers can satisfy demand. Users of compute should be indifferent as to the origin of the compute they consume, but they have specific computational needs that they wish to perform subject to budget and time constraints. And they should be indifferent to the mixture of classical and quantum computing that best suits their needs and constraints.

This indifference between classical and quantum has significant consequences for how quantum computing is distributed between firms in the economy — and, indeed, between geopolitical powers. At this stage in the development of quantum computing, the major open question is how relatively large the space of computational tasks that are best suited for classical computing are versus that for quantum computing.

For computational tasks where classical computers dominate, compute is already massively decentralised — not just with multiple large cloud services (AWS, Google etc) but in the devices on our desks and in our pockets. There is no barrier to competition in classical compute, nor any risk of one geopolitical actor dominating. Where bottlenecks in classical compute emerge are in the production networks for semiconductor chips — a known problem with a known menu of policy stances and responses. Similarly, no such risk emerges around computational tasks where classical or quantum systems are equally suited.

The salient question is whether there will arise a natural monopoly in quantum compute? This could arise as a result of bottlenecks (say of scarce minerals, or caused by market structure as in the semiconductor chip industry), or as an outcome of competition in quantum computing development. As an example, one argument might be that as quantum compute power scales exponentially with the number of qubits then a geopolitical or economic actor that establishes a lead in qubit deployment could maintain that lead indefinitely due to compounding effects. This is a quantum takeoff analogous to the hypothesised ‘AI takeoff’ (see Bostrom 2014).

Several factors mitigate against this. The diversity of architectures for quantum computing being built suggests that the future is likely to be highly competitive; not merely between individual quantum compute systems but between classes of architectures (eg. superconducting, ion trap, photonics). While quantum compute research and development is very high cost, it is proceeding widely and with significant geographical dispersion. There are at least eight distinct major systems or architectures for quantum computing, seven of which have successfully performed basic computational tasks such as the control of qubits (see the survey by Bremmer et al 2024). 

Nor is there any obvious concern that first-mover advantage implies competitive lock-in. Quantum compute is quite unlike AI safety scenarios, where ‘superintelligence’ or ‘foom’ is hypothesised to lead to a single monopolistic AI as a result of the superintelligence using its capabilities to 1) develop itself exponentially and 2) act to prevent competitors emerging. Quantum computing is and will be, for the long foreseeable future, a highly specialised toolset for particular tasks, not a general program that could pursue world domination either autonomously or under the direction of a bad actor.

One significant caveat to this analysis is that the capabilities of quantum compute might have downstream consequences for the economy, and this could . The exponential capabilities at factoring provided by quantum compute could undermine much of the cryptography that protects global commerce, and underlines the need for the development and deployment of post-quantum cryptography. We have argued elsewhere that the signals for the emergence of quantum supremacy in code breaking will emerge in the market prices of cryptocurrency (Rohde et al 2021). There is a significant risk mitigation task ahead of us to adopt post-quantum cryptography. It is a particularly difficult task because while the danger is concrete, the timeline for a quantum breakthrough is highly uncertain. Nonetheless, the task of migrating between cryptographic standards is akin to many other cybersecurity mitigations that have been performed in the digital economy, and while challenging should not be seen as existential.

Instead, the institutional economic view of quantum computing emphasises the possibilities of this new technology to radically grow the space for market exchange — particularly when we understand the possibility of quantum computing as co-developing alongside distributed ledgers, smart contracts (that is, decentralised digital assets) and artificial intelligence. Quantum computing lowers the cost and increases the performance of economic exchange across an exponentially growing ecosystem of digital property rights. It will be an important source of future economic value from better economic institutions.

References

Berg, Chris, Sinclair Davidson, and Jason Potts. ‘Beyond Money: Cryptocurrencies, Machine-Mediated Transactions and High-Frequency Hyperbarter’, 2018, 8.

Bremner, Michael, Simon Devitt, and Dr Eser Zerenturk. “Quantum Algorithms and Applications.” Office of the NSW Chief Scientist & Engineer, March 2024.

Bostrom, Nick. Superintelligence: Paths, Dangers, Strategies. Reprint edition. OUP Oxford, 2014.

Bova, Francesco, Avi Goldfarb, and Roger G. Melko. ‘Commercial Applications of Quantum Computing’. EPJ Quantum Technology 8, no. 1 (December 2021): 1–13. https://doi.org/10.1140/epjqt/s40507-021-00091-1.

Davies, Antony. “Computational Intermediation and the Evolution of Computation as a Commodity.” Applied Economics, June 20, 2004. https://www.tandfonline.com/doi/abs/10.1080/0003684042000247334.

Rohde, Peter P, Vijay Mohan, Sinclair Davidson, Chris Berg, Darcy W. E. Allen, Gavin Brennen, and Jason Potts. “Quantum Crypto-Economics: Blockchain Prediction Markets for the Evolution of Quantum Technology,” 2021, 12.

Roth, Alvin E. ‘The Economics of Matching: Stability and Incentives’. Mathematics of Operations Research 7, no. 4 (1982): 617–28.

Stigler, George J. ‘The Economics of Information’. Journal of Political Economy 69, no. 3 (1961): 213–25.