Build vs Buy? For Fixed Income, the Answer Is “Yes, and…”

Architecture becomes the edge

What emerged from the discussion was not a choice but a philosophy. “Build versus buy, it is yes, and.” The desks that feel most confident are those that own the architecture and the IP that truly matters, but rely on external services for undifferentiated heavy lifting, such as data collection, connectivity, and basic workflow components. Adaptive reflects this mindset in its own positioning. Rather than delivering a fully bespoke stack from the ground up, it focuses on providing  the foundational layer that lets clients plug into multiple venues, EMS, and OMS platforms, while retaining ownership of their data models, logic, and code. Platform agnostic by design, this approach “takes out the bottom layer” and gives firms room to innovate on top.

The conversation returned repeatedly to architecture. Old standalone databases that sit outside the main environment and cannot be integrated are now seen as dead ends. So are systems that trap data inside proprietary black boxes. One participant described building an innovative side database that eventually became worthless because it was never integrated into the firm’s primary data architecture. Another highlighted the limitations of all-in-one platforms, where “you do not get any of the data that goes in, you only see your trades and the cover, and if you do not trade it is gone.” The lesson was clear. Point solutions do not scale. “Winning five years from now comes down to good architecture,” one attendee said. “How do you enable everyone in your company to have access to crucial data, even the custodian?”

Cloud, AI, and economics

Cloud adoption is changing the picture. Centralising data storage and analytics in the cloud has given some firms a single source of truth across trading, risk, and portfolio management. It has also made it easier to ingest third-party datasets, LLM-ready research archives, and new venues without having to rebuild pipelines each time. At the same time, it introduces new questions about cost, latency, and governance. Consumption-based pricing can flatten upfront spend, but becomes unpredictable as usage grows. Cross-border data rules, encryption standards, and auditability requirements all need careful design, particularly as AI enters the stack.

AI was treated as both an opportunity and a cautionary tale. Traders and technologists in the room are excited about natural language interfaces that can sit on top of research, filings, analyst calls, and dealer chats, turning unstructured flows into structured, searchable inputs. Several firms are already building large databases specifically to feed LLMs, while others are happy to consume that service from external providers and focus on their own models and interfaces. There is interest in chatbot-style tools that can query positions, surface opportunities, and visualize risk in real-time. Yet there is also a widely shared scepticism. Many models can deliver eighty percent accuracy, but the final twenty percent is where real money is made or lost. Compliance teams are concerned about data leakage, opaque logic, and AI-generated code that nobody fully understands. Early adopters may gain experience, but they also carry reputational risk. The table largely viewed AI as a long-term enabler, rather than a short-term edge.

Economics and culture add another layer of complexity. Large firms may lean build first, with big engineering teams and a willingness to experiment across taxable, munis, and emerging markets. Even then, 2025 style resource constraints mean they need confidence that projects will deliver within budget and on time. Smaller and mid-sized firms increasingly look to the cloud to level the playing field, accepting that they cannot hire ten traders for every strategy and must instead rely on the “right algo price and the right quantamental approach.” As one participant summed it up, “If you have those things, I do not think you need ten traders.”

Change management remains a quiet but powerful theme. Old guard attitudes still exist. New tools are sometimes met with “Why do you need that? It will be fine,” which can slow progress. Successful teams are learning to “sneak in” upgrades by tying them to regulatory requirements or by making technology tangible. One firm described how every potential client who visits their trading desk experiences an “aha moment” when they see the in-house system running in real time. For many, that demonstration is more persuasive than any slide deck, because it turns abstract architecture into a visible competitive advantage.

Throughout the discussion, Adaptive’s role sat in the background as a kind of facilitator. Not the hero system that claims to do everything, but the connective tissue that makes hybrid strategies viable. In a world where no turnkey solution exists for all of a firm’s needs, the ability to integrate, orchestrate, and evolve is more valuable than any single application. “You need tech to stay in the game, and then add beyond it,” one attendee said. The firms that will thrive are not necessarily those that build the most or buy the most. They are the ones who design for change from the start.