UK-based research and data advisory firm Substantive Research and Aiera, the New York-based provider of AI-based content delivery, recently published a joint study finding that almost 80% of the 35 largest global asset managers have enterprise-wide generative AI platforms in place, yet broker and data licensing restrictions are blocking a majority of those same firms from integrating sell-side research feeds directly into their AI systems.
Key Findings from the Substantive Research / Aiera Survey
The main findings from this recent survey reveal a striking gap between AI infrastructure deployment and content integration. While 77% of buy-side firms have rolled out generative AI platforms such as Claude and ChatGPT across their organizations, the content they value most remains largely outside those systems due to broker/data licensing issues. Sixty-nine percent of those surveyed see this as being the biggest barrier to integrating broker research into their internal AI systems.
Seventy-seven percent of survey respondents named broker research as the most valuable input to receive as machine-readable feeds directly into their internal AI systems, ahead of earnings transcripts at 57% and market data at 42%. Compliance and entitlements issues ranked second as a barrier at 54%, behind the licensing restrictions at 69%.
Onboarding timelines compound the problem. Thirty-seven percent of respondents said approving, adopting, and onboarding AI models takes four to six months. Another 20% reported the process takes more than six months. Only 17% completed onboarding new AI models in one to three months. That spread is a major disadvantage for smaller and mid-tier asset managers who lack the dedicated compliance and procurement infrastructure to accelerate the process.
Mike Carrodus, CEO of Substantive Research, described the commercial tension clearly: “The buy-side has mobilised to ensure that they remain competitive as the industry rapidly gears up from an AI perspective. What hasn’t been solved yet is how that affects their commercial relationships with the sell-side. The only easy conversations will be with research providers where they are viewed as top tier clients; with anyone else, content feeds into their LLMs will potentially come with a hefty price.”
Carrodus added that the pricing dynamics cut both ways: “Buy-side research budgets are already under pressure, so it’s understandable that they wouldn’t welcome a new set of licensing accompanied by new costs. And they are banking on a fear from the sell-side that being frozen out of an asset manager’s LLM would be a very cold place to be in future. On the other hand, many on the sell-side feel that capitulation on this issue would be a step too far, especially after Mifid II’s enduring deflationary effects on research pricing.”
The survey also surfaced a secondary strategic question: whether asset managers should build AI infrastructure around general-purpose models or vertically integrated platforms built specifically for investment research. Just over a quarter of respondents are evaluating or have already implemented specialized finance-focused AI platforms. Forty-four percent view those platforms as potentially strategic long-term partners, while another 44% remain in the evaluation stage but undecided.
Gavin Skinner, COO of Aiera, framed the ecosystem challenge: “Buy-side firms aren’t asking whether AI belongs in the research process. They’re asking how to bring their most trusted information sources into AI workflows securely and compliantly. The future of investment research depends on creating an ecosystem where premium content can be accessed intelligently while fully respecting the rights and commercial interests of content owners.”
The survey, which covered firms representing a combined AUM of more than $20 trillion, identifies broker research licensing as the single biggest barrier to AI adoption in the investment research workflow.
About Substantive Research & Aiera
Substantive Research is a UK-based research and market data discovery and pricing analytics provider. The firm serves asset managers representing a combined AUM of more than $20 trillion and total assets exceeding $25 trillion. Its platform allows investment research consumers to compare provider pricing, benchmark consumption habits against peers, and optimize their overall research spend.
Substantive Research subsequently extended its data and analysis into market data, creating what it describes as the industry’s first apples-to-apples Market Data Spend Analytics Service, enabling buy-side and sell-side firms to compare market data payments and budgeting against peers. The firm is owned by Euronext, which acquired it as part of its broader financial data strategy.
Aiera is a consortium-backed, AI-enabled content delivery and access platform serving content providers, content consumers, and the broader content delivery ecosystem. The platform is designed to support development of industry-wide solutions where buy-side and sell-side interests are aligned, with a particular focus on machine-readable content standards and AI-compatible licensing frameworks.
Our Take
The Substantive Research / Aiera survey confirms what many research heads have suspected: the bottleneck in buy-side AI is no longer the AI itself. General-purpose LLMs are already live at most major asset managers. The constraint is content, specifically broker research, which respondents identified as their most valuable source by a wide margin. That is not a rounding error. It means sell-side firms currently control a key variable in buy-side AI effectiveness, whether they intend to or not.
The licensing dispute the survey documents is not new, but the stakes have risen sharply. Broker research has historically been distributed as PDF reports or through proprietary portals designed for human analysts. Those formats were never built for automatic ingestion by enterprise AI systems. As asset managers accelerate their AI buildouts, the mismatch between how research is delivered and how it needs to be consumed is becoming an operational constraint with direct competitive consequences.
For the sell-side, the leverage is a complicated issue. Brokers who hold out on licensing terms may preserve short-term pricing power, but they risk being routed around entirely as asset managers turn to more accommodative brokers, specialized research-oriented AI platforms, earnings transcript providers, and alternative data vendors that have already built machine-readable distribution into their commercial models. The 25% of buy-side firms already evaluating vertically integrated investment-research AI platforms is a number worth watching; if it grows materially over the next twelve months, it becomes a major signal that changes the calculus for every mid-tier broker still negotiating.
The MiFID II parallel Carrodus draws is pointed. A decade of unbundling compressed sell-side research pricing and pushed many firms toward subscriptions and commission-sharing arrangements that left little margin for experimentation. A second wave of structural pricing pressure, this time driven by AI licensing disputes rather than regulatory mandates, would rock an industry that is already running lean. We think the firms most exposed are the second and third-tier brokers who lack the relationship and commercial leverage to command premium licensing terms but also lack the scale to absorb concessions.
It will be interesting to see whether Aiera’s consortium model or a competing standards body moves quickly enough to define machine-readable licensing norms before individual bilateral negotiations between large asset managers and top brokers establish de facto precedents that smaller players cannot match.
For more news about alternative data, investment research, and investment related FinTech, click here to sign up for our free ResearchWatch Weekly newsletter, or click here to subscribe to our proprietary ResearchWatch news and commentary service.