terminal VENTIUM

The AI-Native Event-Driven Hedge Fund

Forced Flows.
Found by AI.

Ventium Capital is an AI-native event-driven hedge fund. Our agents trade the predictable dislocations created by $18T+ of passive capital — surfacing alpha signals beyond human insight and running the full trade lifecycle with tighter risk and faster iteration than any human desk.

Alpha Discovery

AI agents surface signals across events, filings, and flows that humans cannot process at scale.

Execution Moat

Adaptive sizing, staging, and market-impact control in real time.

Risk Discipline

Humans approve campaigns. Hard limits enforce downside control.

event_engine

LIVE PILOT: ACTIVE // HUMAN OVERSIGHT: ON

Scroll

What We Do

AI-Native
Event-Driven Trading.

Ventium is an AI-native event-driven fund focused on structurally recurring flows.

We start with index rebalances, where passive funds are forced to trade on a fixed timeline, creating repeatable dislocations.

01

Proprietary Capital

We trade our own book with strict limits and controlled leverage.

02

AI Alpha Discovery

Agents continuously mine events, filings, flows, and microstructure for alpha signals beyond human insight, then route the best ones to capital.

03

Execution as the Edge

Vendor forecasts are commoditized. Our edge is what AI discovers on top, plus sizing, staging, and impact-aware exits.

Why This Works

$18 Trillion in
Forced Flows.

Every quarter, S&P, Russell, MSCI, and hundreds of custom indices rebalance. Passive funds tracking over $18 trillion in AUM must trade at the close. This is a structural mandate, not a discretionary choice.

The Opportunity

Index rebalances create 20-30bps of structural drag, $36-54 billion annually in value transfer from passive funds to participants positioned to capture it.

Every event-driven desk sees this opportunity. Almost none can trade it at scale.

Annual Forced Flow

$18T+

Structural Drag

20-30bps

The Bottleneck

x

A human PM manually manages 5-10 names. A rebalance moves 400+ tickers simultaneously.

x

Rigid schedule-based algos leak information. Counterparties detect and exploit them.

x

Fragmented spreadsheet-to-terminal processes create operational risk on rebalance days.

x

Poor integration between signals, portfolio construction, execution, and risk.

The workflow between prediction and profit is the bottleneck.

Our Edge

Not Prediction.
Workflow.

Vendor forecasts are a commodity, anyone can buy them. The edge is the alpha our AI agents discover on top, and how we size, stage, execute, hedge, and exit trades around forced flows.

visibility

AI Crowding Agent

A dedicated AI agent continuously ingests alt data and 13F filings to estimate how many funds are positioned on the same trade. High crowding triggers reduced sizing or full avoidance.

network_intelligence

AI Alpha Discovery

Specialised agents mine filings, microstructure, and cross-asset flows for non-obvious patterns around each event. The result: a proprietary signal layer on top of vendor forecasts that no human desk has the bandwidth to replicate.

electric_bolt

AI Execution Agent

An AI execution agent adaptively slices orders in response to real-time order book conditions, liquidity toxicity, and counterparty detection risk. TCA feedback minimises market impact and slippage on every fill.

precision_manufacturing

Orchestration Layer

AI agents coordinate signal processing, portfolio construction, execution, and risk management as one integrated system. No spreadsheet handoffs. No manual bottlenecks.

trending_up

Reversal Capture

AI agents position ahead of forced flow, exit into the close, then capture the post-event mean reversion. Two returns per event, with TCA-optimised entry and exit.

cycle

Proprietary Feedback Loop

Every trade generates TCA and execution data that AI agents use to improve future sizing, timing, and crowding estimates. Human overrides feed back into the models. The system compounds its edge.

water_drop

Liquidity & Impact Modeling

AI agents assess order book depth, toxicity scores, and ADV multiples in real time. Position sizing scales dynamically to minimise market impact and transaction costs.

How the System Works

Signal to Portfolio
to Execution to Risk.

AI agents orchestrate the full pod workflow: integrated, automated, and continuously improving. Humans set the parameters and oversee exceptions. The agents handle the rest.

System Architecture
Event Processing Pipeline
HUMAN
OVERSIGHT: ACTIVE
01 Ingestion
INDEX PREDICTIONS + CORP ACTIONS + SEC FILINGS
02 Modeling
FLOW ESTIMATION > DEMAND SCORING > CONVICTION
03 Crowding
ALT DATA > 13F CLUSTERING > CROWD SCORE
04 Sizing
PORTFOLIO CONSTRUCTION > HEDGE > EXPOSURE MGMT
05 Execution
AI AGENT > ADAPTIVE SLICING > TCA-AWARE > ANTI-GAMING
06 Risk
LEVERAGE > VAR > GROSS/NET > CONCENTRATION
07 Reversal
POST-EVENT REVERSION > FEEDBACK > MODEL UPDATE
08 Anomaly
FLASH CRASH > LIQUIDITY EVENT > FAT-FINGER DETECT

Concurrent Tickers

400+

Human Desk Capacity

5-10

Returns Per Event

2

Human Controls

Always On

Return Source 01

Pre-Event Spread

Position ahead of forced demand. Exit into the close as passive funds are compelled to trade at any price. TCA-optimised execution minimises market impact throughout.

Return Source 02

Post-Event Reversal

Artificial demand subsides. Prices mean-revert to fair value. AI agents capture the reversion systematically. Two distinct return streams from a single event.

Beyond Scheduled Events

Anomaly Capture

AI agents detect and respond to unscheduled dislocations that human traders cannot process in real time: flash crashes, liquidity events, fat-finger errors, and abnormal order flow. Always-on monitoring turns transient anomalies into captured alpha.

Risk & Control

AI Agents Automate
the Repeatable.
Humans Control
the Risk.

This is not a black box. Every system decision is explainable, every position is justified, and every campaign requires human approval before going live.

Human Sign-Off Required

Every new event campaign requires human approval before capital is deployed. No exceptions.

Hard Risk Limits

Position limits, leverage caps, gross exposure bounds, and concentration thresholds are enforced automatically and cannot be overridden by the trading system.

Full Explainability

Every AI agent-generated position includes a stated thesis, sizing rationale, crowding assessment, and risk justification. Nothing is opaque.

Human Override at Any Time

Any position can be manually overridden, reduced, or closed at any time. Override data feeds back into model training.

Independent Risk Checks

Pre-trade and post-trade risk validation runs independently of the trading system. Separate infrastructure, separate logic.

Why Start Here

Index Rebalances Are
the Ideal Wedge.

Not every event-driven strategy is equally suited to AI automation. Index rebalances are the best starting point for a reason.

calendar_today

Scheduled & Predictable

Exact dates known weeks in advance. No surprises, the calendar is the trigger.

water

Liquid & Accessible

Large-cap and mid-cap names with deep order books. Sufficient liquidity to enter and exit at scale.

smart_toy

AI-Native Architecture

Structured, quantifiable, repeatable. Built for specialised AI agents that automate systematic decision-making at scale.

lock

Driven by Forced Flows

Passive funds have zero discretion, they must rebalance. This creates genuine price dislocations, not speculative signals.

build

Builds the Full Stack

Crowding, execution, risk, reversal, every capability built for rebalances generalizes to adjacent strategies.

target

Underserved Mid-Cap Niche

Mid-cap and custom-index rebalances offer the highest alpha. Large multi-manager pods are too big to operate efficiently here.

Long-Term Vision

Start Narrow.
Expand Methodically.

Index rebalances are the starting point, not the ceiling.

Current Focus

Index Rebalances

S&P, Russell, MSCI, and custom indices. Pre-position across the full opportunity set, capture spread and reversal, refine the operating system with every event.

Forced Flow / Event

$2-8B

Events / Year

500+

Tickers / Event

400+

architecture

Corporate Actions

Spin-offs, tenders, rights issues. AI agents parse every recapitalization filing and model the pricing dislocations the system can capture.

EXPANSION PHASE 02
account_tree

Merger Arbitrage

AI agents parse filings, model regulatory risk, and assess deal probability. Same workflow engine, different event type.

EXPANSION PHASE 03

One Operating System. Many Event Types.

The same AI-powered core system (signal processing, crowding estimation, execution optimisation, risk management) applies across all event-driven strategies. Each new event type generates data that makes the AI agents smarter. Each expansion is methodical, not speculative.

AI-native specialist pod → broader event-driven fund

Founder / CIO

Built by Someone
Who Trades These Flows Daily.

Alexandru Runcianu

Founder / CIO

Trades index rebalance flow daily on the Systematic Risk Trading and Portfolio Trading desk at Bank of America. Sees firsthand the constraints: manual sizing across fragmented systems, rigid execution that leaks information, and alpha consistently left on the table.

Previously at Arini Capital Management during its earliest stage, learning what it takes to build a world-class fund from zero to one.

That combination of sell-side insight into forced institutional flow and buy-side exposure to elite fund building is the foundation for Ventium.

Systematic Risk Trading

Bank of America

Arini Capital

Buy-Side Foundation

CISI + EUREX

Regulated Professional

Read the Full Thesis arrow_forward

FAQ

Frequently Asked
Questions.

Why not just buy vendor predictions?

add

Predictions are a commodity. Multiple vendors sell the same rebalance forecasts to every event-driven desk on the street. The edge is not in knowing what will happen. Our AI agents size positions, assess crowding, execute with TCA-minimised market impact, manage risk in real time, and capture the post-event reversal. That's Ventium's moat.

Why does AI matter here?

add

A single index rebalance can involve 400+ tickers moving simultaneously. A human PM can manage 5-10. Specialised AI agents don't just make the process faster, they make the full opportunity set accessible. AI agents automate the repeatable workflow (signal processing, sizing, execution, risk monitoring) while humans focus on judgment calls, risk release, and exceptions.

What is the moat?

add

Vendor forecasts are the floor, not the ceiling. Our agents layer proprietary alpha discovery on top, mining filings, microstructure, and cross-asset flows for patterns no human team can monitor at the same breadth. Then crowding estimation, TCA-optimised execution, liquidity modelling, reversal capture, and the live feedback loop compound the edge with every event.

Why start with index rebalances?

add

Rebalances are scheduled, liquid, quantifiable, and driven by forced flows, making them the ideal first strategy for an AI agent-driven system. They also build every capability (crowding, execution, risk, reversal) needed to expand into corporate actions, merger arbitrage, and broader event-driven strategies.

How do humans stay in control?

add

Every event campaign requires human sign-off before capital is deployed. Hard risk limits are enforced independently of the trading system. Any position can be manually overridden at any time. All AI decisions include full explainability, thesis, sizing rationale, risk justification. This is not an uncontrolled black box.

Is this a black box?

add

No. Every position the system generates has an explicit thesis, a sizing rationale, a crowding assessment, and a risk justification. Pre-trade and post-trade risk checks run independently. The system is transparent by design, because real institutional capital requires it.

How is this different from other quant funds?

add

Most quant funds are built around statistical factor models trading broad market signals. Ventium deploys specialised AI agents focused on a specific, well-defined structural edge: forced passive flows around index rebalances. We don't take market views or make macro calls. AI agents trade the mechanics of institutional mandates while minimising transaction costs.

What is the revenue model?

add

Proprietary capital with controlled leverage on structurally recurring events. Two return sources per event: the pre-event spread (positioning ahead of forced demand) and the post-event reversal (capturing mean reversion as artificial demand subsides). No market view. Alpha from the mechanics of forced flow.

Why can't large multi-manager funds do this?

add

Scale works against them. The highest-alpha segment, mid-cap and custom-index rebalances, requires nimble execution and modest capital deployment. Large pods are too big to operate efficiently there. They also face structural constraints: legacy systems, rigid execution tools, and human-dependent workflows that can't process 400+ tickers simultaneously.

Why now?

add

Three structural tailwinds: passive AUM continues accelerating (more forced flows), AI agents are now production-ready for workflow automation and adaptive execution, and incumbent desks are structurally stuck on legacy infrastructure they can't rebuild from scratch.

AI-Powered Execution. Forced-Flow Alpha. Human Control.

Book an Intro Call

We are speaking with a small set of aligned partners. Share details and we will follow up.