Could you briefly introduce Adaptive and explain the role it plays in the evolving digital asset trading technology stack?
At Adaptive, we build custom trading technology for firms that want greater ownership and control over their platforms. In digital assets, that often means helping firms address some of the harder infrastructure challenges around performance, resilience, consistency, and 24/7 operations, while giving them the foundations on which to build their own business logic and differentiate. Aeron, our high-performance infrastructure technology, is widely used for low-latency and resilient trading systems across digital asset markets and beyond. Digital asset firms — including Coinbase, Bitvavo, Talos, EDXM, and Bullish — use Aeron as part of their trading infrastructure.
We have also worked with firms to deliver systems spanning execution, connectivity, clearing, and settlement. For example, with EDXM, we built an institutional, cloud-native clearing and settlement platform on Amazon Web Services (AWS) in under 20 weeks.
Digital asset markets operate 24/7 with very different performance demands compared with traditional markets. What are the key architectural challenges this creates for trading infrastructure?
One of the biggest architectural challenges is simply that there is no maintenance window. Exchanges and trading firms need to be able to upgrade systems, fail over infrastructure, and recover from incidents without stopping trading. That makes resilience a fundamental architectural requirement, rather than something that can be addressed operationally afterwards.
Another challenge is evolving a platform while it remains continuously live. Firms still need to ship new features, upgrade components, patch issues, and improve performance, but they have to do so without disrupting trading. That means the architecture has to support rolling upgrades, controlled migrations, and fail-safe deployments while maintaining consistency across orders, market data, risk checks, and positions.
There is also the challenge of scaling without introducing unnecessary complexity. Digital asset platforms can experience sudden spikes in volume, while also connecting to a large number of external venues and services. That is one reason we are seeing greater interest in architectures designed around deterministic behaviour and fault tolerance from the outset. Sequencer architectures are one example of that, particularly in digital assets.
Adaptive has spoken about the next generation of the digital asset trading tech stack. In practical terms, how does this differ from the infrastructure that many exchanges and trading firms rely on today?
We see the next generation of digital asset technology moving away from large, monolithic systems towards more modular architectures. In a market that operates 24/7, firms need to be able to upgrade, replace, or scale individual components without introducing risk across the entire platform. But modularity is only part of it. We are also seeing firms place much greater emphasis on ownership and control of their technology. Digital asset market structure continues to evolve quickly, and the sector is increasingly intersecting with traditional asset classes and institutional workflows. Firms want to be able to adapt their platforms, launch new products, and respond to changes in liquidity, regulation, or client demand without being entirely dependent on a vendor's development cycle.
That is where the next-generation stack really differs from many legacy environments. It is designed not only for flexibility, but also for speed to market, resilience, high performance, and scalability. With no single central exchange in digital asset markets, liquidity is fragmented across hundreds of venues, so firms need infrastructure that can connect to multiple external systems while still delivering the reliability and throughput expected of institutional trading platforms.
Sequencer architectures and zero-downtime systems are becoming increasingly important in digital asset markets. How are these approaches reshaping the way exchanges and trading platforms design their infrastructure?
Sequencer architectures can make it easier to build resilient, high-performance distributed trading systems at scale. In practical terms, they can simplify state management and create more predictable behaviour under load, while reducing some of the complexity involved in scaling a system, recovering from failures, or reconstructing what happened after an event.
That is particularly relevant in digital assets, where systems need to run continuously while maintaining consistency across fast-moving markets. Rather than relying on large numbers of loosely coordinated services and eventual consistency, a sequenced model can provide a clearer way of maintaining an authoritative ordering of events.
With Aeron Sequencer, we are trying to make that architectural model easier for firms to implement without requiring them to build all of the underlying infrastructure themselves. The objective is to give engineering teams a foundation for maintaining consistency and resilience as their platforms scale.
Performance and scalability are critical in crypto markets. What have your cloud performance benchmarks revealed about the limits and opportunities of cloud-native trading infrastructure?
What our Aeron benchmarks have shown is that cloud infrastructure can support very demanding trading workloads when the underlying architecture is designed appropriately.
Our work with AWS and Google Cloud has helped test the assumption that low-latency, high-performance trading systems necessarily need to run on-premises. There are clearly trade-offs involved, but the results show that the cloud can be a viable environment for workloads that historically would have been considered difficult to move away from dedicated infrastructure.
They also show that cloud performance depends heavily on how the system is deployed. Technologies such as DPDK, for example, can materially improve throughput and latency by changing the way applications interact with the underlying network infrastructure.
These benchmarks were also designed to be transparent and reproducible, allowing firms to validate the results in their own environments rather than relying on headline performance claims.
Adaptive works with a wide range of clients, from early-stage startups to some of the largest crypto exchanges. What common lessons or patterns have emerged from building infrastructure across such a diverse set of firms?
One pattern we see consistently is that firms are increasingly clear about which parts of their technology they want to own and where they want to differentiate.
They generally want control over areas such as their business logic, market structure, and client experience, but that does not necessarily mean they want to build every piece of foundational infrastructure themselves. Using established infrastructure for some of those underlying capabilities can allow engineering teams to spend more time on the parts of the platform that are specific to their business.
The underlying technical priorities are also surprisingly similar across firms of very different sizes. Performance, resilience, observability, and the ability to scale and evolve without adding unnecessary complexity tend to come up repeatedly. At the same time, firms increasingly want greater control over their technology roadmap, alongside faster time to market for new features and functionality, and less of the lock-in often associated with vendor-owned products.
The firms that tend to do best are those that combine a clear product vision with architecture built for change. In digital assets especially, that balance between control, speed to market, and resilience is critical.
Your Aeron technology has become widely used across trading environments. Why has it effectively become a de facto infrastructure layer for many crypto exchanges and trading platforms?
Aeron is particularly well suited to digital asset markets because it was designed around high throughput, low latency, resilience, and the flexibility to run effectively in both on-premise and cloud environments. That matters in a market that is always on, fragmented and constantly evolving.
It also gives firms a model that fits where the sector is today. Aeron has a large open-source community, which has helped make it a widely adopted infrastructure layer across trading environments, while Adaptive provides enterprise capabilities and support for firms on top of open-source Aeron to production-harden, scale, and operate these systems with confidence.
For example, EDXM leveraged Aeron as the underlying trading infrastructure to completely re-platform its exchange. The platform was built and deployed in seven months and recorded zero unplanned outages in the more than a year following launch, alongside a median round-trip latency of 73 microseconds.
← Previous interview
Linework
Noel Damien Foti
Next interview →
Neo
Laurent Descout
Image: Adaptive 
