Quantamental 2.0: Why the Quant–Fundamental Divide is Collapsing

The traditional separation between fundamental and quantitative investing is collapsing as generative AI changes how investment professionals interact with data, analysis, and risk.

Drawing on a presentation delivered by Petr Merkuryev at the Exponential Summit, this article examines how natural language interfaces are accelerating convergence across research and risk, why context is becoming the primary source of durable edge, and how investment organizations are beginning to adapt their operating models in response.

A Convergence Years in the Making

For a long time, fundamental and quantitative investing developed as parallel disciplines.

They relied on different tools, spoke different analytical languages, and optimized for different strengths. Fundamental investors focused on company narratives, management quality, and economic intuition, asking why an investment made sense. Quantitative investors focused on patterns, signals, and systematic execution, asking what the data revealed.

The separation was real, and for a long time it was functional.

Today, that separation is collapsing. Not because one approach is displacing the other, but because the practical barriers that once kept them apart are being removed.

The first signs appeared well before the recent wave of generative AI. After the 2007 quant crash and the 2009 junk rally, many firms were forced to confront uncomfortable truths. Signals without context proved fragile in stressed markets. Narrative-driven investing without systematic risk awareness proved equally incomplete.

By the early 2010s, the term quantamental emerged to describe attempts to bridge these gaps. Large fundamental funds embedded data scientists within investment teams. Quant firms hired sector specialists to inject domain understanding into models. The rise of alternative data further blurred the lines.

The Hidden Cost of Integration: The Translation Tax

By the early 2020s, most investment organizations already operated somewhere between the two extremes.

Yet integration remained slow, costly, and uneven. Despite good intentions, structural friction persisted.

Portfolio managers, analysts, quants, risk teams, and data scientists continued to operate across different systems, factor definitions, and workflows. Each new question required interpretation, translation, and coordination across roles.

A fundamental PM might notice an unexpected move in a short position and wonder whether it reflected a factor unwind, forced deleveraging, or a retail-driven squeeze. The question would be passed to a data team, analyzed offline, and returned days later, often after brokers or news had already provided a partial explanation.

This delay is rarely a data problem. More often, it is an organizational one. Insight was slowed not by a lack of information, but by the need for human intermediaries to translate between domains. This "translation tax" kept analytical speed artificially constrained.

Why the Interface Matters More Than the Model

Generative AI did not suddenly make investors better decision-makers. What it changed was the interface.

Before natural language systems, access to advanced analysis was gated by technical skill, tooling, and availability. Even technically proficient investors were often constrained by permissions, fragmented infrastructure, and long development cycles.

Natural language interfaces began to shift that balance. Questions can now be expressed directly, in the language investment professionals already use. The same interface increasingly spans research, risk analysis, monitoring, and operational workflows. Iteration happens in minutes rather than days.

As a result, analytical speed is no longer determined primarily by coding ability. It is increasingly driven by domain understanding, and by the quality of the questions being asked.

The Modern Quantamental Investor

This shift is changing how investment work gets done along three dimensions.

• 1.Analytical acceleration. Complex hypotheses can be explored without writing code. "What if?" and "why?" questions can be tested in near real time, including during live events such as earnings calls.
• 2.Operational independence. Portfolio managers and analysts can increasingly build their own dashboards, monitors, and lightweight research pipelines without constant data-engineering support.
• 3.Functional access. Research, risk, and operational analysis begin to share a common interface and language, reducing friction across roles.

The result is not the disappearance of specialization, but a new baseline: quant speed combined with fundamental depth.

When AI Becomes Standardized, Context Becomes the Edge

As analytical capabilities become more widely accessible, differentiation inevitably shifts.

Frontier models can be purchased. Broad datasets are increasingly commoditized. Human judgment remains essential, but on its own, it is no longer sufficient.

Context increasingly becomes the source of sustainable differentiation.

That context has two distinct layers:

• 1.External market context includes curated, finance-specific information about companies, events, sectors, and their relationships over time. This is what allows AI systems to reason about markets rather than generate generic text.
• 2.Internal firm context reflects each organization's unique DNA: mandates and constraints, institutional memory, decision playbooks, factor definitions, and the heuristics that govern real-world investment decisions.

When these layers are combined and made usable by AI systems, they shape not just what insights are generated, but how those insights are interpreted and acted upon.

From Ideas to Infrastructure: Making Context Compounding

Context does not accumulate automatically. It must be built deliberately.

This typically involves three elements:

• •External context infrastructure, grounded in reliable, AI-native market data structured around entities, events, and time.
• •Internal knowledge capture, turning institutional memory, post-mortems, and decision rules into explicit, machine-readable assets.
• •Feedback loops, where questions asked, decisions made, and outcomes observed are systematically fed back into the system.

Over time, this creates compounding advantages that are difficult to replicate. Not because the tools are unique, but because the embedded judgment is.

The Emerging Role of the Context Engineer

These changes give rise to a new organizational capability, often emerging as a distinct role: the Context Engineer.

This is not a rebranded data scientist. Nor is it a generic AI product role.

A context engineer sits at the intersection of three skill sets:

• •Organizing knowledge and taxonomies
• •Designing workflows across data, AI, and investment teams
• •Understanding markets and what truly drives decisions

Data scientists focus on optimizing models. Context engineers focus on optimizing what models see, and how outputs are used. In consulting and IT services, this capability is already being industrialized. Within much of asset management, it remains informal, even though the underlying need is widely recognized.

From Episodic Analysis to Continuous Insight

The practical implications are already visible.

Earnings calls increasingly trigger real-time peer comparisons and language-based factor analysis. Market shocks prompt immediate scenario exploration rather than delayed post-mortems. New alpha ideas move from data discovery to first-pass validation within the same trading day.

Often the same people, working within the same time constraints, but with materially different analytical throughput and depth. What was once a competitive edge is fast becoming the expected operating standard.

The Window Ahead

The convergence of quant and fundamental investing was inevitable. Generative AI has simply accelerated it.

The next phase of competition will be defined less by who adopts AI, and more by who uses it to encode judgment, context, and experience into durable decision-making systems.

The quant–fundamental divide is collapsing not as a slogan, but as a practical reality. The firms that recognize this early, and build accordingly, will shape the next generation of investment operating models.