Cloud providers such as Amazon Web Services, Google Cloud and Microsoft are taking a keen interest in the exchange industry. The business of running a trading venue involves huge amounts of data processing, and most of the leading exchange operators are moving parts of their operations into the cloud. That presents an opportunity for cloud providers to leverage their services in partnership with exchanges, especially in data analytics and artificial intelligence.
This trend is also impacting the brokers and clearing firms that interact with the exchanges. These firms are exploring ways to sift through massive amounts of data and use AI to improve their interactions with exchanges and improve the services they offer to their customers. In some cases relying on the exchanges for part of that solution may be better than building it in-house.
All this is a big change from the early days of cloud computing when the focus was on saving money by outsourcing data storage to cloud providers. According to Rohit Bhat, a managing director at Google Cloud, the exchange industry is now entering a "Cloud 2.0" phase that is focused on generating value for the intermediaries, such as broker-dealers and futures commission merchants, and the ultimate clients.
Speaking on a panel at FIA’s Boca conference last month, Bhat said he has seen the evolution go through three phases.
“With cloud, the journey entry point was really around the fundamentals of ‘Well, help me store this stuff and I'll make sense of it later’ – this was a Cloud 1.0 mentality,” said Bhat, who has been working with financial services clients at Google Cloud since 2017.
“Cloud 1.5 then became: ‘You have a great network, maybe I can use you as a distribution engine, potentially for historical or real-time analysis’. Now firms are saying, ‘What capabilities as a service can a [trading] venue provide to the participant community?’ You see that happening now, and in almost near real-time, great value-added services,” he said.
“Venues are starting to think about providing calculation engines for risk or indexing services, bringing the calculation into near real-time to the actual data stream itself, so that by the time they are talking to an FCM, or talking to an endpoint client, these things are provided as a service from the venue, versus every single person having to ingest that information and do it on their own for their own systems,” he added.
“I think the evolution is now going to be shifting from this ‘Cloud 1.0’ to a ‘Cloud 2.0’ mindset, which is a lot more about value generation than trying to deal with the bits and bytes,” he said.
Bhat acknowledged that data quality and consistency is a key challenge for making the most of new technologies.
“That's not to say that there's not work to be done around standardisation and normalisation behind the scenes on the actual data models themselves, be it for risk, be it for margining and so on, but at least you can now have a common platform to play on and build these services from. This is the major shift moving forward,” he added.
Several other experts speaking alongside Bhat echoed his emphasis on the need to have a strong foundation of data management in place to take advantage of the cloud's capabilities.
“To both improve the way you operate and improve the products you're creating, the fundamental structure of your data has become much more important,” said Magnus Haglind, senior vice president for marketplace technology at Nasdaq. “What's pushing and driving this is the hype cycle around AI and machine learning. The promise of AI can only really be unlocked with access to your data.”
The panellists also discussed where they are seeing opportunities for AI.
“The opportunity right now with AI in the listed derivatives marketplace is in operational efficiency,” said Helen Hartwell, head of exchange-traded derivatives client consulting at UBS. “There has been a lot of work in the industry over many years to automate and we're at very high levels of straight-through processing. Clients are looking to us now to find ways to further optimise their flows in those next few percentiles. AI or machine learning elements are really good at finding trends that aren't visible to the human eye.
“We see that already happening in the reconciliation space with improved matching algorithms. We're also seeing it from a client service perspective and improving workflows. For me, this is where the real opportunity is right now to move the industry forward,” she said.
Nasdaq’s Haglind added that he also sees potential in the areas of compliance and trading integrity. “The obvious space from a machine learning and AI point of view is to ensure that our markets are secure, that people are behaving as they should, and to root out fraud. This is where this technology is going to help to identify anomalies when it comes to bad behaviour in markets.”
