The intelligent capital layer

What happens when strategies work together?

In design · Product direction

A portfolio needs more than a collection of individual strategies. It needs evidence about how they behave together, across market conditions and within a set of constraints.

We are designing infrastructure for that problem. Portfolio construction is not available as a product today.

Discuss the portfolio layer →
equitiessectorsmacrocommoditiesdigital assetsvolatilityratesmarket-neutraltrendcarryintradaymulti-monthregime-conditionalCompositionweighting · diversification · constraints · continuous rebalancing · portfolio-level evaluationportfolio-level productobjective: capital preservationportfolio-level productobjective: market-neutral incomeportfolio-level productobjective: long-horizon growththe unit of access moves from a strategy to an objective
  1. Specialized strategiesDifferent markets, horizons, and sources of return
  2. Shared objectives and constraintsEvaluate how strategies behave together
  3. Portfolio compositionSelect, combine, and monitor strategies as a whole
Conceptual architecture: specialized strategies, shared constraints, portfolio-level composition.

From selection to composition

At scale, the question changes.

With a handful of strategies, capital chooses among them. With an ecosystem of hundreds of specialized systems across asset classes, horizons, and sources of return, choosing among them stops being the point. The question becomes how a set of them should be weighted, diversified, constrained, and rebalanced against an objective, and how that composition should change as conditions do.

That is portfolio construction, with strategies as the components.

What specialization looks like

Narrow by design. Valuable in combination.

Specialized intelligence is narrow by design: a system that understands one source of return under one set of conditions. Its value is in combination. A volatility system and a trend system are more useful together than either alone; a regime-conditional agent is only useful if something above it knows which regime it is in. The composition layer is where narrow systems become broad outcomes.

The infrastructure of composition

Composition needs evidence that only a platform can produce.

To compose strategies well, a system needs to know how each behaves: in isolation and in combination, under stress, across regimes, and against specific constraints. That knowledge comes from an ecosystem in which every strategy was evaluated the same way and observed in the same live environment.

Promethean is building the infrastructure for:

  • Behavioural evidence: standardized, comparable, continuously updated records of how each strategy behaves.
  • Composition: weighting, diversification, constraint, and rebalancing across strategies, as reproducible, versioned, evaluable artifacts in their own right.
  • Portfolio-level evaluation: the same methodology applied one level up: point-in-time, out-of-sample, provenance-tracked.
  • Continuous management: composition that responds to conditions, with the same monitoring and controls as any deployed system.

A portfolio of intelligence is itself a piece of intelligence. It is built, evaluated, and deployed on the same platform.

Objectives

Eventually, the unit of access is an objective.

The end state this layer points toward: a party with capital expresses objectives and constraints (horizon, tolerance, existing exposures, what must be true) and systems operating above the strategy layer construct and continuously manage an appropriate composition of intelligence. Selecting individual strategies becomes optional.

What the layer learns

Which forms of intelligence work together.

Every strategy evaluated, every composition tested, every live observation across many systems in one environment adds to an understanding of which forms of intelligence complement one another, under which conditions, for which objectives. That understanding improves every composition built afterward. It accrues to the platform, and it compounds.

This is the largest and most defensible problem in the company, and it is reachable only through the two layers beneath it.

Where this is today

The platform produces the evidence this layer requires, and experiment comparison across strategies is live. The composition infrastructure is in design. We are talking with allocators and managers about what portfolio-level evaluation and construction should look like. Nothing on this page is an offer of any product or service.