Monday, May 31, was a federal holiday here in the US, as it marked Memorial Day, a holiday to honor those who have fallen in defense of this country.

Every Memorial Day weekend, I try and read a story written by Chris Jones, then of Esquire magazine, titled “The Things That Carried Him”. Regardless of what your feelings are toward America’s military apparatus, I highly recommend poring over this, not only because it’s one of the finest pieces of journalism ever penned, but because it’s a sobering reminder of the cost of war. You can find the original here, but it’s behind a paywall, or you can read this annotated version from the Nieman Foundation.

Ok, now for a rough transition … let’s talk market data.

Big Tech’s big market data plans

I’ve been writing this weekly(ish) column for almost one full year and the topic that has come up the most is the idea of application interoperability. This is because I believe that app interop will revolutionize the world of trading. As Lynn Martin of the Intercontinental Exchange once said to me, “We are continuing to evolve our legacy technologies—which is no small lift—to the more modern technologies. Gone are the days of the installed heavy terminal, heavy workstation.”

To oversimply it, banks and asset managers are becoming increasingly comfortable using the major public cloud providers (in a hybrid manner), though they’re farther behind in adoption than some would like to admit. The reason for this is more data is flooding the market, thus the need for more robust storage options and the ability to run high-compute analytics on these datasets so as to find unique insights. At the same time, AI tools—and, specifically, machine learning (ML) and natural language processing (NLP)—are becoming democratized thanks in large part to the open-source tools supplied by Big Tech companies. Essentially, if you want to find unique insights in an ever-growing sea of data, you absolutely must incorporate ML algos and NLP into your tech strategy. This is leading to wider adoption of APIs to deliver data, as end users no longer want closed-off systems—they want trading services and data providers to play more nicely than they have in the past because they want more and more data to overlay with their own internal data and systems.

I’m sure that someone can better encapsulate this interop trend than I can, and I’m sure I’m missing some pieces (such as standardization efforts underway in the industry), but this is the Cliffs Notes version of the way I see the market. So in my opinion, app interop is the underlying trend that pulls together all those other trends, and is leading platform and data providers to change their strategies so as to embrace interop.

This interop push is going to continue, no doubt, but looming on the horizon is another great potential disruptor to the traditional world of financial technology, and that’s the emergence of Big Tech providers as vital to most every bank, asset manager, exchange, and vendor in the space. Yes, we’ve known that they’re disruptive from a hardware/software/infrastructure perspective, but what about market data disruption?

Before we get into this, let me make one thing clear: the idea of Big Tech—and we’ll use this term as a catch-all for Amazon, Google, Microsoft, and IBM—disrupting the capital markets is hardly a novel idea. For years, we’ve seen these companies hire Wall Street talent and say that trading is high on their priority list, only to fire those people later because they lost the drive and staying power due to the bureaucracy and regulation that surrounds the wholesale cap markets.

But there’s reason to think that this time things will be different.

Last week, I wrote about why we should expect to see more consolidation of market data providers in the year to come. For that column, I asked Bob Iati, managing director of research firm Burton-Taylor International Consulting, to provide some color around market share. He gave me that, but also told me that he would not be surprised if it was the Big Tech providers involved in future M&A. This surprised me and I wanted to hear more, so we spoke again this past week. Here’s what he had to say (and I totally endorse this):

“What’s becoming more evident is that the data business will for the foreseeable future grow faster than the other parts of [the industry]. And big data being what it is, and cap markets still being an area where everyone knows there’s big money, the big data companies are going to want to play—that means Google, Microsoft, Amazon Web Services, IBM. They are all going to be drawn into this space.

“If you think about what they’ve done already, to some extent it almost flies under the radar how much business AWS has—or to a lesser extent Google Cloud or IBM Cloud or Microsoft—but they’re dominant [when it comes to cloud services] in our business. What’s next is, how do they get closer to the day-to-day activities of capital markets firms?”

“If you subscribe to the theory that no one does big data better than the Google, IBM, and to a lesser extent Facebook—though I doubt they’ll get in there, though they certainly have cash and data—if you believe that, why wouldn’t they want to get into the business of having cap markets data on top of the data they already have?”

“So, hypothetically, let’s say AWS buys Bloomberg. Think about what they could do with their analytics and Bloomberg’s data? Or any other pair like that. I think it’s unquestionable that it’s out there. The kind of data and analytics that now drives so much of what we do in cap markets, both from a trading perspective and a middle- and back-office perspective, can be enhanced so much by superior analytics. And you’d have to think those big data companies have analytics that are certainly on a par, if not better than, what [traditional capital markets companies] do.”

He’s not alone in this belief. In February 2021, Burton-Taylor released a survey of market data professionals, and it said this: “Expect non-traditional companies, especially Big Tech, to exert greater influence on the industry, shaking up the industry leadership board that’s been relatively static for many years.”

The firm believes this because more than half (54%) of respondents said that “Big Tech will penetrate market data business via partnership and/or acquisition,” with 47% believing that “a leading provider of market data will be acquired by a ‘Big Data’ or ‘Big Tech’ company .”

When I asked Iati if he thought major front-office tech providers like SS&C/Broadridge/Ion/Murex making a “surprise” acquisition of a major market data provider was more likely, or would it be a Big Tech company, he says, “It’s more likely to be the non-traditional players.

“If you look at what the Big Tech companies bring to the table differently than the fintech shops, they are big data machines. They have processing capabilities in data that exceed even the [largest trading platform providers]. They have companies filled with analytics geeks and unquestionably the best data analytics and AI of anyone. And then they’re flush with cash—how many companies could swallow a Bloomberg? There are probably five in the whole world.

“So I think it’s more likely to be the likes of Amazon and Google, which already have a big presence in cap markets. What they don’t have is the desktop presence. Microsoft, on the other hand, does. They have Azure on the back end, and on the desktop they have the Office 365 tools and Teams.

“I actually expect this (a Big Tech company buying a market data provider) to happen. I don’t know if it’s going to be in a month or two, but I do expect to see it, though I don’t know how they going to do it,” he concludes, meaning he’s not sure if they would buy a whole company, or buy certain business units from a large provider.

The funny thing is that this kind of acquisition would not be considered consolidation, as it’s a whole new type of company entering the space. So, as Iati notes, this would represent a true paradigm shift in the industry.

I guess at this point it’s a game of wait and see. There will be more consolidation—that’s inevitable. There are plenty of rumors swirling that a major deal could be imminent, but that’s just rumor. Even if it’s not a Big Tech company making the next big deal, that doesn’t mean this idea is dead. But if and when it’s announced that it’s an Amazon, Google, Microsoft, or IBM buying a major market data provider, the floodgates are likely to open and we’ll be entering into a whole new marketplace.

Think I’m off base? As always, I’d love to hear your thoughts: anthony.malakian@infopro-digital.com.

Deep learning’s slow advancement

Quickly, before I go back to enjoying the long weekend, I wanted to highlight a story about RBC Capital Markets’ deep learning project. Basically, it’s taken a full decade for the Canadian bank, in partnership with Borealis AI, to figure out how best to incorporate deep learning algorithms into a trading platform—in this case, its Aiden platform, which was launched last fall.

You can click here to read the full details of the project and how RBC is looking to expand Aiden’s capabilities to reduce slippage and alpha erosion while executing trades, and identify periods of volatility that could upend the best historical models. The most important takeaway, though, is that these types of advanced projects take a long time to complete from conception to a live product.

In December 2018, I wrote a painfully long (perhaps too long) deep-dive into projects that involved deep learning and neural networks underway at trading firms and vendors across the industry. For that, I spoke with executives from UBS Asset Management, Credit Suisse, Société Générale, Wells Fargo, AllianceBernstein, Bloomberg, and IBM, among others, to understand the benefits and roadblocks that existed around this bleeding-edge technology. The one thing that was clear is that there was a lot of trial and error involved.

This was the last sentence: “The evolution [of deep learning] is here, but there’s still a long way to go.” Granted, not the greatest kicker ever written, but the point is that it’s still very early days when it comes to the true incorporation of deep learning in the capital markets.

The key, though, is to have staying power. It will take time, but breakthroughs will happen. Actually, let me tweak that—there will be LOTS of failure before you realize those breakthroughs, and you absolutely need to know when to cut bait and realize something isn’t working, but you don’t scrap the whole unit in a cost-cutting measure. AI, and specifically machine learning, and, more specifically, deep learning, are the future of analytics and execution; thus, the experimentation and foundation of expertise has to be created today. Actually, the foundation should’ve been created yesterday.

In January, I wrote about the need for banks to start experimenting with quantum computing today, even if tangible benefits are still years away. The same is true for deep learning, but as RBC has shown, benefits are nearer on the horizon … though it will still take some time to get there.

Damn. Another lame last sentence. Chris Jones, I am not.

The image at the top of the page is “Assaut de Deux Fotresses” by Jacques Callot, courtesy of The Met’s open access program.