Wall Street’s New Alpha Gold Mine is The Data Locked Inside Their Own Walls – BlackRock VP

For years, the savviest hedge funds and asset managers carved out an advantage by tapping so-called alternative data, credit-card receipts, cellphone-tracked foot traffic, satellite images of parking lots, and crop fields that traditional market feeds could never provide.

That edge has now largely disappeared. The information once considered exotic has become table stakes, available to almost anyone willing to pay the right data vendor.

The new frontier for alpha, according to senior executives at some of the world’s largest money managers, lies inside their own organizations: decades of internal research notes, email threads, meeting transcripts, trade rationales, and the accumulated wisdom of veteran portfolio managers and analysts.

Speaking on a panel at the Future Alpha conference in New York on Tuesday, Jacob Bowers, vice president of quantitative research at BlackRock, said: “AI is great at structuring unstructured data,” and “some of the best unstructured data you have is internal.”

Bowers noted that the publicly accessible data that was once cutting-edge is now “commoditized” by AI. BlackRock, the world’s largest asset manager with $14 trillion in assets under management, has already begun deploying internal AI agents to scour past communications between investment professionals and old reports on opportunities in search of potential investment signals that competitors cannot access.

The idea of mining internal data goes back years. A 2019 report from consultancy Opimas predicted that large funds might one day sell portions of their proprietary data libraries to generate additional revenue. Robert Frey, a former managing director at Renaissance Technologies who now runs a fund of funds, told Business Insider at the time that his old firm’s biggest advantage was its “massive data library” gathered over decades of trading.

What has changed is the technology. Advances in large language models have made it far easier, and far more powerful, to extract meaningful patterns from the messy, unstructured troves of information that sit inside long-running asset managers.

At Balyasny Asset Management, which oversees about $33 billion, quant Andrew Gelfand said the firm had previously tried to monetize unstructured data within its systems, but recent AI advances have made the effort much more fruitful. The firm now requires analysts to type their research and notes into a centralized portal that his team can access, giving the AI models “reams of text to sift through for potential investment signals.”

Mike Daylamani, who runs a team that blends fundamental and systematic investing at Engineers Gate, stressed the importance of high-quality input.

“You need the feedstock to be high quality,” he said, referring to the data feeds quants use to build their models. He added a broader reflection on the nature of the business itself: “At the end of the day, this is a creative endeavor.”

The shift represents a quiet but profound change in how sophisticated investors hunt for an edge. Public alternative data sets have become so widely adopted that they no longer reliably deliver outperformance. Large language models can now scrape and synthesize enormous volumes of publicly available information almost instantly, further eroding any remaining advantage there.

What remains truly proprietary, and nearly impossible for outsiders to replicate, is the institutional memory, the failed investment theses, the off-the-record conversations, and the nuanced reasoning that seasoned professionals have accumulated over years or decades.

This internal data is often scattered across email servers, shared drives, compliance archives, and forgotten folders. Modern AI tools, however, excel at exactly this kind of problem: turning chaotic text, voice notes, and documents into structured, searchable intelligence.

Funds that can effectively organize and interrogate their own history gain something no vendor can sell: context-specific knowledge shaped by their own investment philosophy, risk tolerances, and hard-won lessons from past mistakes.

The challenge, several speakers noted, is maintaining the quality of that “feedstock.” AI agents are insatiable and constantly need fresh, high-caliber input to keep pace with evolving markets. Without continuous contributions from top analysts and portfolio managers, the models risk learning from stale or mediocre thinking.

For an industry that spent the past decade chasing ever-more-obscure external data sets, the realization that the richest untapped vein may lie inside their own walls marks a significant pivot. Analysts believe the winners in the coming years may not be those with the biggest alternative data budgets, but those who best preserve, organize, and mine the collective wisdom already sitting on their servers.