Data, Cloud, and AI: The Future of Capital Markets
In a world where there’s increasing awareness around spending tech stack budgets – efficiency and agility have become a lifeline for capital markets firms. But where do organisations start in terms of understanding how the transformative forces of innovation can best serve their interests?
Cloud integration, data analytics, AI-driven trading strategies, AI-advanced risk management, as well as NLP and market sentiment intelligence – all of these topics were discussed at the recent 26th Annual Conference of the Association of Futures Markets (AFM) held in Bangkok. The “AI Unleashed” panel was made up of Vuyo Mpumza (Manager: Commodities, Johannesburg Stock Exchange (JSE)), Putthipong Skonthawat (Executive Vice President-Head of IT Operations Group, The Stock Exchange of Thailand), Russell Toop (Capital Markets Team Lead, Colt Technology Services), Duncan Symmons (President and Founder, Touch Fire Trading), and moderated by Guy Melamed (CEO and Co-founder, Exberry).
Below are some of the highlights:
Cloud integration and co-location
The panel kicked off by covering the extent of cloud integration by certain exchanges. On one hand, the Stock Exchange of Thailand (SET) has been on their cloud journey for more than 10 years now, starting from 100% on-premises infrastructure, to its current state being about 40-50% cloud-based. SET uses the cloud to take advantage of its flexible compute power to perform data analytics and stream data directly to Thai investment apps. SET also employs an efficient ‘cloud burst’ system that uses cloud computing resources whenever on-premises infrastructure reaches peak capacity. The benefits of using cloud were cited as scalability, the ability to go serverless by programming directly on the cloud through the support of a growing number of cloud-native applications, and reduced cost of operations.
A more typical example of the majority of stock exchanges today, however, is the JSE. The South African-based exchange is still in the early stages of moving to cloud and is in the process of assessing how to best manage this process in terms of its clients as well as the exchange itself. During this transition period, the JSE is committed to providing a market with transparent and efficient systems, along with cost efficiencies.
Cloud has brought a number of benefits to financial markets participants in general, including the ability to store large amounts of data. This is essential when it comes to firms retaining large datasets (typically required by high-frequency traders (HFTs), brokers or exchanges), or meeting regulatory requirements. The cloud has also lowered the barrier to entry, especially for emerging exchanges, because prohibitively expensive co-location charges can be bypassed.
One thing that all panellists agreed on was that co-location was not going away anytime soon. This is due to issues around latency and the prevalence of multicast packet drop in certain infrastructures – though exchanges might be encouraged to know that both AWS and COLT are developing solutions to solve these issues.
Use cases of third-generation AI
The discussion subsequently turned to third-generation AI. Following on from the first-generation AI (the more rule-based education of basic applications to machines) and second-generation AI (incorporating deep learning based on a breakthrough of statistical models), the third-generation of AI is all about the matching of reasoning.
The panel went on to mention the growing number of different use cases of third-generation AI within the capital market environment. For instance, it is quickly becoming definitive when devising trading strategies, especially for HFTs. Predictive or sentiment analytics in the area of space analytics and analysis is another interesting example. AI also has the power to quickly provide understanding of unstructured data from news outlets and financial reports. Some HFTs are even using a specific type of AI to optimise trading strategies across different sites by arbitraging different co-location contracts.
In terms of infrastructure, AI can be used for pattern matching to improve surveillance mechanisms and preventative detection controls to deter insider trading. Infrastructure providers use AI in solutions including robotic processes, network health checking and virtualisation testing.
However, in order to capitalise on third-generation AI, the quality of the underlying data is paramount, meaning access to accurate, real-time, low-latency data. This is crucial because as soon as there is a discrepancy in the data, then any strategies around trading or optimisation basically become redundant.
Moreover, it is imperative to remember that there needs to be traceability to the underlying data – particularly in the area of risk. Answers cannot simply appear out of an AI ‘black box’ with no means in which to explain how you came to those results.
Data + Cloud = AI
Because of the need to maintain accurate, clean and complete datasets so as to benefit from AI-driven strategies, technology providers are constantly trying to improve cloud integration experiences to help everyone get there.
Discussions, both within this panel and across the globe, illustrate the growing recognition within capital markets of the pivotal roles that technologies such as cloud and AI will play in shaping future strategies as enablers of transformation.
The market cap of exchanges in emerging markets should exceed the value of U.S. exchanges as soon as 2030, according to Goldman Sachs economists
Change is afoot when it comes to market infrastructure technology. The old arguments over “buy vs build” have been relegated.
Today’s market infrastructure providers need to navigate increased complexity and unpredictability. These pressures in turn intensify the need to urgently modernise outdated infrastructure and tackle technological debt.