Quantum Computing, AI, and Asset Management

We already have examples of the impressive scale of exponentially increasing processing power using pre-quantum processers. Consider SpaceX and the reusable booster. When Elon Musk started SpaceX, he was derided by some veterans in the space industry for even suggesting a booster could be recovered and reused, much less gently landed on a drone ship. As we now know, SpaceX prevailed in their attempt to create the technology. Why was this? Why now? Primarily, because the computing power Musk and his team were able to bring to bear on the problem was greater than ever before realized in the history of that problem. (That and the maverick, outsider-naiveté-based rejection of the conventional wisdom that claimed it wouldn’t be possible meant that it could be.) After all, a booster reenters the atmosphere at something greater than Mach 25, and old hands who had been practicing rocketry for decades knew that the precise firing of thrusters and boosters at just the right times in just the right synchronicities and combinations was impossible, and that a soft touchdown couldn’t therefore be realized. And yet, here we are.

SpaceX did this with classical – albeit very advanced – computing capabilities.

So, the first key realization as we enter the quantum era is that there are innumerable unforeseeable emerging applications, technologies, and capabilities that will result. Insolvable problems will be cracked.

Case Study: Green Alpha Advisors

At Green Alpha Advisors, we are in a unique position to envision the near-term outcomes that might be possible applying quantum computing capabilities within our own industry of asset management. Our economic thesis and stock selection processes are a bit outside of the mainstream, and so far, have not lent themselves to traditional quant-based, formula-ready duplication methods. Several vendors and potential partners have approached us, and together we have attempted to find a rules/algorithm-based way to replicate our processes. So far, none has worked sufficiently. But in a quantum world, that could change.

To explain why, let’s briefly go through Green Alpha’s research processes. Our top-level criterion, the factor that provides the go/no-go decision point for whether to research a given company and its stock, is an internal determination of whether the company is doing more on balance, as an aggregate of its overall business activities, to stabilize the world than to destabilize the world. We define impact investing as sending signals to the market that only the companies providing fixes to our biggest problems have value (and so we decline to purchase companies we define as destabilizing as, by definition, therefore, they will make poor long-term investments).

You might say, “That’s not hard, just plug in ESG scores and boom, that’ll be perfectly quantifiable and replicable by an algorithm.” Except, no. ESG scores are not a reliable way to determine whether a company is helping advance sustainable economics or working to crush them. For example, oil super major Total SA has a good overall ESG score from Sustainalytics. MSCI ESG ratings give both Exxon Mobil and Chevron ‘high average’ ratings. Toyota, responsible for driving a huge quantity of oil demand, and notoriously resistant to a transition to electric vehicles, is an ESG darling, based largely on its very favorable ESG scores created by third-party asset management vendors. In the current ill-defined ESG landscape, there is no way to reach a ‘stabilizing or destabilizing’ conclusion based on these ESG ratings. Green Alpha therefore evaluates each company on its own merits, very carefully, in a similar way to how venture capitalists or private equity managers practice their craft, to determine if – on balance – a corporation’s net activities are resulting in an economy working well for everyone’s benefit.

This involves a lot of work, and in addition requires application of many heuristics that so far have been resistant to algorithmic reduction. As far as we have been able to determine, or have been shown, no algo has yet reproduced one of our strategies. The outputs are just not there.

But what if three or five or 10 algorithms, each optimizing along a different axis, could run in parallel? What if each of these individual algos could be 10 times more powerful than any that preceded it? That is a world not yet experienced by anyone, and a machine capable of that could very well learn to replicate Green Alpha’s research processes, and even apply our philosophies to every publicly listed stock in the world, and therefore derive a kind of super strategy — an equities strategy reflecting all the companies that will define, and indeed constitute, the Next Economy^TM, say, 10 and 20 years hence. That would be great: Not only would it be potentially the highest-impact portfolio investors could own in terms of driving capital towards that economic transition, it may very well also provide a clear path to competitive returns as all the innovative solutions and brilliant approaches uncovered by this quantum-driven algorithm gain market share, and eventually grow to largely constitute the economy itself.

Of course, huge computing power alone won’t result in this portfolio-building juggernaut. Each algo will require a long period of training, and the combination of algos longer still. Before the quantum-based portfolio construction system could completely take over as Green Alpha’s Chief Investment Officer, we would need to operate as a centaur providing inputs and evaluating its decisions over some period of time. But it could replace us. We hope it will.

Green Alpha’s portfolio construction approach depends on a lot of inputs that haven’t been formally or commercially quantified yet. Therefore, it is resistant to automation so far. Many other equities strategies, most obviously index funds, have been run by algorithms for years. The simpler the strategy, the easier it is to automate. Automation ultimately will come for the whole asset management industry, in our view, but the more complex strategies will be among the last to hand over the keys to the machine. However, no amount of complexity will be too great for a properly constructed and optimized series of quantum algos. One of our dreams, in fact, is to work with a leader in quantum development to start the process of creating our own obsolescence. Hey DeepMind, give us a call! OpenAI, same offer.

This all may sound a little whimsical, but it’s not. It’s deadly serious. Channeling capital to the highest-impact, most-scalable, and fastest-growing solutions might make the difference between innovation outrunning the climate crisis or not. Quantum computing will be a key tool in developing and optimizing these maximum-impact portfolios.

Every industry has its own transformational tale to spin about what the quantum computing revolution will mean for it. One of our jobs is to look at companies across industries to see who is planning ahead of this trend taking off and then getting client dollars in front of those compelling investment opportunities. It may seem a bit early, but being early is where investment gains come from.

The importance of the quantum computing revolution is difficult to overstate. Any asset manager or any executive in pretty much any industry who views this as an over-the-horizon thing they don’t have to worry about yet is risking their own obsolescence. The quantum computing world is already beginning to emerge. Put up your tray tables, we may be landing there soon.

 

This article is one section of the report, “Quantum Impact — The Potential for Quantum Computing to Transform Everything.” Click here to learn more and access the full report.