At Amazon Web Services’ re:Invent conference in Las Vegas last week, the cloud giant unleashed a flurry of news, including the announcement that Nasdaq will move its matching engine to the cloud service beginning next year, and Goldman Sachs launched the Financial Cloud for Data, a suite of cloud-based data and analytics solutions that leverage AWS. But a third announcement flew largely under the radar: The AWS Data Exchange (ADX), unveiled two years ago, recruited its first major financial data provider—FactSet—to the platform.

Sources say the move could signal the beginning of an industry shift away from traditional data pipelines, databases, and technical costs and maintenance issues that come with bandwidth, security, and scalability, and toward a data-sharing model with potential to shake up the current commercial pricing models for data across financial services.

“In the olden days, people used to ship data around to wherever an infrastructure happened to be,” says Jonathan Reeve, head of content and technology solutions at FactSet. “Increasingly, especially with [AWS’s cloud data warehouse] Amazon Redshift, you no longer have to move data around the planet to serve an application.”

ADX acts as a central hub for users—comprising data analysts, portfolio managers, data scientists, quants, and developers across a range of industries—to access third-party data in an up-to-date cloud environment. Some 266 data vendors contribute to the service, which caters to other industries including healthcare, automotive, geospatial, consumer, and media and entertainment.

FactSet has deployed 30 proprietary datasets to ADX through Amazon Redshift. FactSet’s datasets make up the bulk of the 54 datasets available for consumption through Redshift on ADX, compared to 3,752 available through Amazon S3 (AWS’s static object storage service), and 42 that are available through an API.

In an on-premises setup, moving data from one datacenter to another incurs telecoms charges for the cost of transportation, and a second database at the other end comes with storage costs. Reeve says using Redshift eliminates the costs on both fronts.

Suvrat Bansal, founder of start-up data discovery platform Stellar Data Labs and former chief data officer at UBS Asset Management, says some vendors and end-users still feel that the only way to unite data is to physically bring it together, but the growing patchwork of data vendors, sources, and users makes that belief more of a fantasy than a conviction.

“[The AWS–FactSet pairing] is the 10-year promise. This is only going to grow. Vendors all need to expose their data through some kind of sharing environment, and eventually feeds will go away because they’re so expensive to build, so this will also encourage vendors to publish more data,” Bansal says.

“It reduces your costs and potential failure points; it reduces pipelines, because you can treat these FactSet datasets as if they’re sitting in your own Redshift database; and it opens the door for more data to be used like that. Once firms are familiar with accessing fundamentals in that way, it becomes easier for smaller vendors to make their data available.”

Bansal adds that new delivery methods like these will precipitate changes to licensing and commercial models, rather than prescriptive traditional contracts. “Once you have the data, you should be able to use it wherever you want to use it … and this will bring about a shift to consumption-based pricing models,” he says.

Dream a little stream 

John Kain, head of business development for banking and capital markets at AWS, is one of those responsible for Amazon’s growing influence on Wall Street. He has deep knowledge of the industry’s workings and market infrastructure, having joined AWS five years ago from JP Morgan, where he was responsible for the investment bank’s market surveillance platforms in fixed income, foreign exchange, and derivatives trading. Prior to that, he spent time at Nasdaq, where he ran a service providing sponsored access solutions in the form of low-latency connectivity, pre-trade risk checks, post-trade portfolio risk, and reporting to investment banks so they could sponsor hedge funds into the market.

Kain knows the important role that real-time—or streaming—data plays at institutions, both because of its necessity for trading and the ire it draws from those who must pay for it. ADX hasn’t yet announced the additions of any real-time data providers to the service, and FactSet’s available datasets on the service are a mix of fundamental and reference data that are updated as covered companies report new metrics. But Kain says ADX plans for such vendors to join the hub in the future.

One source expressed concern that AWS messaging, which uses the HTTPS and TCP/IP protocols, wouldn’t be ideal for streaming or for Fix connectivity, which uses multiple TCP/IP sessions. But Kain says Fix is less relevant to streaming market data than it is to other functions such as order entry and reporting.

“Bloomberg, Refinitiv, Nasdaq, and Cboe all have their own unique ways of sending data to their customers, and we work with all of them,” Kain says. “And we do have a few customers, but I’m not sure anyone’s public, using Fix connectivity from their AWS environment to exchanges in order to trade.”

Despite the service’s name and forthcoming features, Kain says AWS doesn’t intend to compete with the financial old guard as an exchange itself, saying it’s more of a business enabler of institutions than anything else.

“In the exchange area, we’ve certainly helped a number of exchanges,” Kain says, pointing to Nasdaq, which originally partnered with AWS in 2008, migrating its low-latency, high-volume US options markets to the cloud; Japan Exchange Group offering historical derivatives datasets to their customers on AWS; and a pilot between the Singapore Exchange and AWS enabling AWS’s multicast service to test run an exchange in the cloud. But he again stopped short of indicating AWS aims to be an exchange.

“Nasdaq isn’t just an exchange operator, it’s a technology provider to a large number of financial institutions. As they change the way they look at technology, that naturally affects all the other exchanges globally, so our ability to help them helps everybody,” Kain says.

Old friends with new doors

ADX is not the first service of its kind.

FactSet, actually, already has its own similar offering, Open:FactSet Marketplace, a hub for FactSet and datasets from 130 partners that can be consumed via API, datafeed, or a partner’s proprietary solution. Reeve acknowledges a small degree of overlap between Open:FactSet and ADX, but says both services, to FactSet, are centered on making its data “hyper-available” to customers. Curated third-party data on Open:FactSet is tightly integrated with FactSet’s own, with the vendor providing linkages and mapping capabilities, and the whole offering is focused on financial markets participants. ADX, on the other hand, offers FactSet exposure to other industries, which aids the company in its efforts to grow its corporates clientele in a newer part of its Research and Advisory business.

There’s also S&P Global Marketplace, another datahub launched by S&P Global Market Intelligence last year. Most datasets available through Marketplace are proprietary to Market Intelligence and other S&P Global affiliates, including Platts, IHS Markit, and Cusip, though it includes some third-party content, as well.

Also participating in the datahub craze is the recently IPOed data warehouse Snowflake and its Snowflake Data Marketplace, of which FactSet is also a contributing partner with its fundamentals data. Snowflake is a competitor of the likes of Amazon Redshift and Google BigQuery.

In an interview last year with WatersTechnology, FactSet CTO Gene Fernandez said FactSet had initially decided to partner with Snowflake over other data warehouses because it took the onus off users, and because Redshift worked very well with AWS, but wasn’t suitable for users of proprietary or private clouds. However, in the past year and a half since the announcement, more institutions and many of FactSet’s clients have gravitated toward the major cloud providers, including AWS. A spokesperson for FactSet says the company maintains itself as cloud-agnostic, and Redshift provides an efficient way to meet clients using Amazon technology where they are, while Snowflake continues to be unique as another agnostic provider and partner. 

A CTO for an asset manager with more than $5 billion under management says this development with AWS “could open doors for us to imagine how we might interact with our FactSet data in new ways,” adding that he would need to understand the development better to flesh out more thoughts on the deal’s potential.

‘It will never substitute’

A senior commercial executive at a market data vendor says they use AWS extensively for what they do and is in active talks with AWS about Data Exchange.

“It will never substitute existing direct channels for data, but it’s an interesting additional channel to market—it depends on the community they can build, and what clients it would allow you to reach who you can’t reach directly,” the executive says.

Extended reach is an attractive factor for FactSet, just as it’s likely that several of the non-financial participating vendors will hope to become sources of alpha-yielding alternative data for deep-pocketed banks and buy-side firms that sign up for the service.

But Barry Star, CEO of Wall Street Horizon, a specialized data vendor in corporate events, says he doesn’t see a strategic advantage to ADX other than infrastructure savings, and as a result, the deal may appeal to a limited audience.

“Hedge funds make a living by beating everyone else, and they beat everyone by having better data. And to manage that, they need better data systems. So why would they use the same platform as everyone else? The big firms all spend money to build proprietary systems,” Star says.

However, he says the broader buy side uses Bloomberg so extensively that it’s clear that segment of the industry isn’t concerned about everyone using the same system. “A buy-side researcher has very different needs from hedge fund guys. When you have 325,000 users of Bloomberg, everyone is using the same system because they know their strategic edge is not the data, but their analysts,” Star says.