Exclusive Q&A with John Kain, Head of Market Development for Financial Services at Amazon Web Services (AWS).
Cloud technology continues to play an increasingly pivotal role in shaping the future of trading infrastructure. With its promise of scalability, resilience, security, and innovation, cloud technology is transforming workflows across the industry. However, the unique demands of electronic financial markets—where ultra-low latency, deterministic performance, and regulatory compliance are paramount—present a complex set of challenges for cloud providers.
In this exclusive Q&A, TradingTech Insight sits down with John Kain, Head of Market Development for Financial Services at AWS, to explore how AWS is addressing these challenges, driving modernisation, and enabling both exchanges and market participants to unlock new opportunities through cloud innovation.
TTI: Welcome John. Can you tell us a bit about your background and your role at AWS?
JK: I lead the market development efforts for financial services at AWS, where I’ve been for over seven and a half years. In my role, my team and I work closely with our customers from an industry perspective to help them adopt AWS technology. Our focus is on understanding the unique needs of financial services firms and helping them to achieve their modernisation and innovation goals with AWS.
Most of my team comes from the financial services industry, bringing deep expertise to our work. Prior to joining AWS, I spent five years at JP Morgan in the market surveillance area, where my responsibilities included monitoring trading activities to ensure compliance and integrity in the markets. Before that, I was at NASDAQ, where I managed a product that provided low-latency trading infrastructure, as well as real-time risk and regulatory solutions for investment banks sponsoring hedge funds in the markets. Earlier in my career, I was involved in startups focused on low-latency, high-performance execution in the trading space.
Drawing from this background, my team and I have a strong understanding of global financial markets and their challenges. In addition to capital markets, our team brings expertise across other key sectors, including banking, insurance, and payments. This breadth and depth enables us to effectively support our customers as they navigate their transition to the cloud with AWS.
TTI: Cloud adoption in financial markets has progressed significantly in recent years. How has AWS addressed the unique challenges of market participants, while enabling firms to leverage cloud innovation? Can you share examples of how AWS solutions have been integrated to support both existing trading environments and future scalability?
JK: From a financial markets perspective, cloud adoption has been advancing rapidly, particularly in post-trade operations, over the past decade. Across all asset classes, workloads supporting trading systems—such as surveillance, real-time market data distribution, billing and analytics, and even using data to develop new order types for exchanges—are running comfortably and at scale on AWS.
Real-time data distribution, for example, is a key area where AWS plays a significant role. Traditional aggregators like Refinitiv and Bloomberg provide data natively on AWS, enabling users to access it almost instantly via API calls. Exchanges such as Euronext, NYSE, CBOE, and NASDAQ also distribute real-time, non-latency-sensitive data globally using AWS infrastructure. By leveraging the AWS backbone, these workloads benefit from variable costs and global scalability. Surveillance systems, including those from NASDAQ, NYSE, and FINRA, have also embraced AWS, gaining enhanced security, resilience, and scalability. This scalability proved critical during peak COVID trading volumes, when NASDAQ processed up to half a trillion records in a single evening—an immense leap from the usual 30 billion.
Core trading workloads have also started moving to AWS. While there’s a perception that exchanges don’t use cloud, many actually do. High-performance platforms like Coinbase in crypto trading and National Australia Bank’s FX matching platform run natively on AWS. These systems are less regulated than traditional equity or derivatives markets, where ultra-low latency and fairness are critical. However, even in regulated environments like fixed-income trading—such as Broadridge’s LTX platform—AWS is being adopted because latency requirements are less stringent. These exchanges operate seamlessly on AWS while meeting full regulatory requirements.
The challenge for AWS lies in highly regulated markets where fairness, ultra-low latency, and deterministic performance are essential. Traditional trading ecosystems are designed for stability and fairness rather than the dynamic scalability that cloud offers. Additionally, there’s a broader ecosystem of trading partners and liquidity providers whose needs must be addressed, alongside the exchange’s requirements.
To bridge this gap, AWS collaborated with NASDAQ to extend AWS infrastructure into NASDAQ’s data centres through our AWS Outposts service. AWS Outposts allows AWS to deploy compute and storage infrastructure directly into customer data centres while maintaining the operational consistency of the cloud, including common CI/CD pipelines, deployment frameworks, and security tools. For NASDAQ, we made specific modifications to suit market needs, such as ultra-low-latency networking, precise time stamping, and multicast support. These adjustments ensured fair and deterministic latency across trading participants, akin to the uniformity seen in traditional market setups.
Importantly, this approach allowed NASDAQ to modernise its infrastructure without disrupting its existing trading environment. By integrating AWS technology seamlessly, NASDAQ could innovate and monetise its markets while maintaining continuity for its trading participants. This model represents our strategy for supporting even the most regulated and latency-sensitive aspects of financial markets.
TTI: Regarding those latency-sensitive trading firms, how is AWS addressing their diverse infrastructure needs, in particular balancing custom co-location setups with cloud adoption?
JK: When you look at the quantitative trading community, many firms continue to bring their own hardware into co-location data centres. These setups are often highly customised, featuring FPGAs, custom NICs, and specialised switches tailored to their trading strategies. In the short term, it’s unlikely that these firms will move away from these hardware advantages or the intellectual property they’ve developed around their platforms. As a result, flexibility is key: the industry needs to support traditional co-location customers while also enabling cloud technology to be integrated closer to the matching engine for less latency-sensitive users.
Some of these firms already use AWS Outposts within data centres, extending their environments into the cloud for certain workloads. However, others still prefer to maintain full control over their hardware. Interestingly, while these firms may rely on custom infrastructure for trading execution, they are significant users of the cloud for other parts of their operations. For example, they leverage AWS for research modelling, back-testing, and algorithm development. Many sophisticated firms even containerise their trading strategies and push them directly to co-location facilities for execution. While the final execution environment may be highly tailored, the development and testing of those strategies often depend heavily on the cloud, requiring seamless integration between cloud-based infrastructure and co-location setups.
This creates a spectrum of needs within the trading community. For some, traditional co-location will remain essential. Others may adopt solutions like AWS Outposts or leverage geographically proximate cloud environments for lower-latency connectivity to exchanges. The key differentiator is deterministic latency.
One example of innovation in this space is One Trading, a crypto exchange that created a “colocation in the cloud” service. They used AWS cluster placement groups, which allow Amazon EC2 compute instances to be located as close together as possible within an AWS data centre. By doing this, One Trading could bring their customers’ trading engines into proximity with their own in AWS. They achieved P99.9 round-trip latencies of less than 100 microseconds, even at message rates of 10,000 per second.
For regulated equity and derivatives exchanges, however, the challenge lies in achieving nanosecond-level deterministic latency for participants. It’s not just about reducing latency—it’s about ensuring that latency is consistent and fair for all participants. This level of precision remains a barrier for moving certain workloads fully into the cloud.
TTI: For low-latency electronic trading, time synchronisation and accurate measurement are critical, both to meet regulatory requirements and to enable granular performance monitoring across trading stacks. What challenges do these issues present in a cloud environment, and how is AWS addressing them?
JK: In recent years at AWS, we’ve made significant investments in enhancing our time services to provide fine-grained timestamping capabilities. Initially, we relied on NTP (Network Time Protocol), driven by GPS and satellite systems in our facilities. Late last year Amazon Time Sync Service introduced the support for Precision Time Protocol (PTP) and enabled time synchronization within microseconds of UTC on Amazon EC2 instances, This is integrated with our custom Nitro networking and hypervisor card, allowing for precise synchronisation and high performance.
Our focus on improving time services extends beyond financial markets—though they are a key driver. These capabilities also benefit areas like global databases, where fine-grained timestamping enables advanced functionality and performance. It’s an area we continue to prioritise, with ongoing investments to refine and expand our offerings.
For co-located trading environments, one of the key co-innovations with Nasdaq’s use of AWS Outposts has been the ability to leverage external time sources. This allows specialised hardware within an AWS Outposts rack to deliver even more precise timestamps tailored to the needs of high-performance trading environments.
We also recognise the importance of visibility and transparency in network performance. This is an area where we’re actively innovating to provide customers with greater transparency and accessibility to these performance metrics.
TTI: Trading venues encompass a wide range of platforms, from regulated exchanges to systematic internalisers, dark pools, MTFs, OTF, ATSs, ECNs and so on. What types of workloads are these disparate venues migrating to the cloud, and how are they approaching the transition?
JK: Some of the best public examples of trading venues using AWS are interdealer brokers like TP/ICAP. Their platforms benefit from traditional cloud advantages such as security and scalability, but more importantly, they leverage insights across asset classes. By providing a unified view of customer activity across different asset classes, they enhance matching, generate trading ideas, and enable more efficient risk management.
In many traditional investment banks and brokers, operations remain siloed—FX teams often don’t interact with credit or equities teams, for example. However, success in today’s markets increasingly depends on bringing together insights across these dimensions, particularly from a risk management perspective. For example, Trumid has utilised machine learning to combine public regulatory filings with internal trading activity data in the fixed income market. By analysing these datasets, they can identify participants likely holding positions and connect them with liquidity, significantly improving matching rates across the venue.
For trading venues with broader asset class footprints, the focus is on harnessing these insights to drive higher matching rates and improve liquidity. Even regulated exchanges are leveraging similar innovations. NASDAQ, for instance, developed the Dynamic M-ELO (Midpoint Extended Life Order) order type, which uses machine learning to enhance midpoint liquidity on its order book. By analysing exchange data, the algorithm better matches buyers and sellers, increasing trading volume and liquidity—a key objective for these venues.
One of the significant advantages of cloud technology for exchanges is the ability to derive value from market data, often seen as a byproduct of their operations. With cloud-based compute power and advanced tools, firms can extract hidden value from this data, enabling new ways to source and push liquidity. This is particularly important given the increasing role of unstructured data—such as satellite imagery, news feeds, social media, and press releases—in trading decisions. For example, some crypto exchanges are now using generative AI to analyse news feeds in 25 languages, identifying sentiment and key topics in under a second.
This approach contrasts with traditional latency-sensitive market makers, who primarily rely on order book data. However, there is a growing segment of trading firms that are slightly less latency-sensitive and increasingly incorporate unstructured data into their decision-making processes. For these firms, the ability to leverage AWS’s scalable infrastructure and cutting-edge technology is critical. It allows them to analyse large volumes of diverse data quickly while staying close enough to the action to react effectively when insights are uncovered.
TTI: Any final thoughts?
JK: For us, the focus is on building a seamless transition from traditional colocation environments to the cloud. Over time, the gap between what can be achieved from a latency perspective on-premises and in the cloud continues to narrow. Each year, advancements in compute, storage, and analytics bring us closer to parity with what the leading firms can achieve in their own data centres. However, we recognise that these markets require a gradual, non-disruptive transition.
The existing market microstructure cannot simply be upended to move to the cloud. Our goal is to provide a path that allows the current infrastructure to evolve into AWS in a way that is smooth and unobtrusive for all participants. We are continuously investing in and exploring technologies that enable this transition, ensuring a seamless experience while meeting the unique demands of the financial markets.
TTI: Thank you, John.
Subscribe to our newsletter
