For the very first “Waters Wrap”—which was published on July 12, 2020—I wrote about how companies like HPR, Trading Technologies, and various other lowercase-t trading technology providers were undergoing ambitious projects to move their legacy platforms to the cloud. Three weeks later, I wrote about how some of the largest exchanges were partnering with the Big Tech cloud providers for increasingly more complex functions.

If I’m doing my job well, the whole point of the “Waters Wrap” column is to highlight trends earlier than other outlets and connect today’s events to those in the past. Cloud adoption is hardly a new concept, but on November 4 of last year, a fundamental shift was felt in the capital markets with the announcement that Google would invest $1 billion into the CME Group as part of a 10-year partnership. It’s the largest investment Google has made to date in the financial services sector. Through the deal, the CME will begin moving its technology infrastructure to Google Cloud, with the exchange’s data and clearing services being the first to migrate.

At the time, I wrote this was likely the first domino to fall and that other exchanges would likely follow suit. Sure enough, at the end of November, Nasdaq announced a similar partnership with Amazon Web Services that would see the exchange migrate its North American markets to AWS in a phased approach. The future of the capital markets will reside in the cloud.

(Quick note: We will be covering the London Stock Exchange Group’s acquisition of Tora, but we’re still talking with industry sources about what it might mean for users and what trends it might point to. But rest assured, we’re on it.)

The CME and Nasdaq will not be the last to make aggressive pushes to the cloud using the tech giants. To get a better feel for how a large migration like this looks at its outset, I recently spoke with Phil Moyer, vice president of strategic industries for Google Cloud. We talked about a range of trends (and crazy theories I have), as well as how the CME rollout will unfold. Let’s start with the latter.

Moyer says that the exchange is committed to migrating their core, non-latency-oriented applications to the cloud in phase one. “This will give those applications global accessibility, as well as on-demand scale and capacity to those workloads,” he said.

Phase two will aim to “accelerate innovation” at the CME. As Big Tech firms look to carve out a niche for their services, AWS has focused on using its massive user base and cheap prices to attract more users; Azure possesses the Microsoft Office suite (and those ubiquitous Excel spreadsheets) along with its Teams collaboration tool; and IBM brings its mainframe and hardware business along with its inroads into the regtech and AI spaces.

Google Cloud, on the other hand, wants to differentiate itself using AI in the front and middle offices. As a leader in the fields of natural language processing and computer vision, its facial recognition tools can pick out and identify your mom in a sea of uploaded images to Google Photos. Its machine-learning tools like BigQuery are used by quant funds of all sizes, as is the Google-developed, open-sourced Tensorflow. For a nominal fee, you can store all your documents, photos, and entertainment in its cloud.

Google is looking to sell itself on the premise that if it can do all these things for everyday people, imagine what new products and services it can create for an exchange.

“We believe they will have many innovative uses of Google’s data analytics and machine learning including our knowledge graph, our natural-language processing capabilities, our Document AI technologies—a whole variety of technologies that they will be using to do innovation for their markets, customers and settlement, clearing and regulatory functions,” Moyer said.

The third phase is more in the realm of space exploration. What can the CME teach Google about the capital markets and how to serve them well, and vice versa?

When I started at WatersTechnology (then just called Waters) in 2009, the discussion of Big Tech centered around colocation and high-frequency trading. (The short-sighted—in my humble opinion—“Flash Boys” would be published five years later.) Location was everything, as alpha was generated by trading faster than the other guy, and so firms would spend millions of dollars to get closer to the exchange and plug in the fastest data distribution tools. For the purposes of trading, cloud was not a latency-friendly tool. What this third phase will look to address is unlocking some of the hardest financial services problems associated with low-latency workloads, Moyer says.

Listening to Moyer, it sounds like this exploratory potential is what’s leading the push by financial-services firms to get further inside the cloud. If you want to use alternative data, you need cloud for storage and analytics. If you want to use more sophisticated forms of AI like neural networks, cloud is going to play an important piece. If exchanges want to improve on multicasting, the next frontier is the cloud. When you have two of the largest exchanges in the world announcing that they are moving their trading infrastructures to AWS and Google, the tipping point for cloud’s Wall Street takeover is already in the rear-view mirror.

So it is that Moyer believes that cloud has brought along with it a paradigm shift in the capital markets.

“I think that the financial industry has always had significant interconnectedness between all of its participants. But in the past, proximity and location determined liquidity. As we look forward, it’s less and less about physical location, and more about the accessibility of the asset, and your ability to attract investors while meeting regulatory requirements,” Moyer said. “To do this at a global scale, you need to be able to settle and clear assets 24/7, and as quickly as possible to support your customers in managing risk. I think that the CME and lots of other exchanges are realizing that accessibility to the capital markets is going to determine the future of liquidity, and not necessarily proximity.”

Crawl, walk, run

Those three phases mirror broader trends in the market. As much as banks and asset managers like to talk a big game when it comes to cloud adoption and migration, the financial-services sector is not as advanced as some might have you believe. To start, it’s a crawl, walk, run scenario, and the crawl is moving the low-hanging—non-mission critical, latency-sensitive—workloads to the cloud.

The second phase is all about innovation and, specifically, using AI to generate alpha, improve risk systems, and more efficiently handle middle- and back-office processes. Trading firms like to make it seem like they’re all Renaissance Technologies, but use glorified robotic process automation (RPA) to masquerade as AI, and are nowhere near as sophisticated with neural networks or making advancements in the field of NLP, where Google’s BERT is exemplary. 

And then there’s the third phase. On Wall Street, you need to be fast. Before (and even still) that meant being first to a trade. That definition is evolving, though, as people want to be fast/first when it comes to both connecting dots and running analytics. Contextualizing data to make informed decisions is becoming the great differentiator for trading firms—cloud and AI are essential to doing that better and more efficiently.

Through this partnership, Google and the CME have a shared fate in how this all plays out over the next decade. Outages or hacks will set not just those two companies back, but the entire industry and how it incorporates the cloud tools of the future (and even quantum computing). But a success here (and at Nasdaq) will lead others to follow similar paths.

Have thoughts on exchanges moving to the cloud? I’d love to read them: anthony.malakian@infopro-digital.com.

The image accompanying this column is “Noah: The Eve of the Deluge” by John Linnell, courtesy of the Cleveland Museum of Art’s open-access program.