Dark Matter venture overview

Context

Dark Matter is a C Corp. Each producer — Megalith, Factor Prime, Joule, Gravity — is a separate corporate entity that lists assets into the Dark Matter schema. This document is about the DMTR layer. Producer businesses have their own materials.

The producers are not hypothetical. Megalith manages $250M in fully deployed venture capital across 30+ SPV co-investments and is raising a $300M follow-on. Joule operates live flare offtake agreements in Argentina. Gravity maintains a proprietary intelligence database (the Blackmoat database) covering 2,500+ private companies with 58K+ funding records. Factor Prime originates collateralized loans against DMTR-encoded positions — it is both a producer (listing lending facilities into the schema) and a consumer (pledging positions as collateral and surveilling coverage).

Dark Matter starts as the ownership and transfer layer for unitizable private positions — beginning with Megalith-originated venture SPVs and Factor Prime secured lending — and expands to any asset class encodable as a Position Object. It is not an exchange, not a fund administrator, and not a tokenization platform. It is the schema and execution layer that those products can be built on.

Part 1: What is broken?

The largest unstructured asset class on earth

Private markets manage over $13 trillion in assets globally — venture, PE, private credit, real estate, infrastructure, energy, digital assets. The asset class has grown at roughly 15% annually for the past decade. Preqin, Bain, and McKinsey all project $20T by 2030.

That is a lot of capital running on fax-machine infrastructure.

Anatomy of a single secondary transfer

Look at what happens when an investor tries to sell an LP interest:

Total: 6–8 weeks. $30,000–$50,000 in legal and administrative costs. Every step rebuilds ownership state from documents. Every step is repeated from scratch the next time the same position trades.

Why

Public securities have ticker symbols, CUSIPs, ISINs, LEIs. A share of Apple is the same object whether held at Fidelity, traded on Nasdaq, or lent at JPMorgan. Private market positions have none of this. The root cause: private market ownership has no canonical digital representation.

Public markets solved this problem decades ago. DTCC settles $2.4 quadrillion annually. SWIFT standardizes cross-border messaging. FIX standardizes electronic trading. Each is an infrastructure layer that made everything above it possible.

Private markets have no equivalent. Without a common schema, there is no API, no automation, no lending market, no agent coordination.

Part 2: What must exist?

Ownership as an executable object

Every private market position — regardless of vehicle, asset type, or jurisdiction — can be represented as a single canonical object. Think of it less as a database row and more as a state machine: versioned, event-sourced, carrying its own rules. This is the Position Object.

The Position Object

Six regions encode everything a system needs to know:

Three questions every position answers

1. What is it? — asset identity, vehicle class, issuer, terms
2. What can happen to it? — transfer restrictions, consent requirements, encumbrances, jurisdictional constraints
3. Who has claims on it? — owner, lender, GP, custodian, regulatory authority

A system that can answer these three questions for any position can automate everything that currently requires counsel and spreadsheets. The target-state improvements:

• Discovery-to-IOI: weeks → minutes (target)
• Transfer cost: $30–50K → near-zero (target)
• Origination-to-collateral-eligibility: months → days (target)
• Feasibility evaluation: 1–2 weeks → milliseconds (target)

Autonomous agents can operate on Position Objects. They cannot operate on PDFs and phone calls.

Part 3: How is it built?

The producer–schema–consumer model

Dark Matter is a three-tier architecture:

1. Upstream producers — domain-specific businesses that originate assets, generate intelligence, or create financial products. They encode outputs as Position Objects and list into Dark Matter via MCP or API. The four current producers are described in the Context section above; the key point is that each is a separate entity with live operations.
2. Dark Matter (the schema layer) — the canonical registry, rights engine, identity graph, collateral engine, and settlement bus.
3. Downstream consumers — applications, terminals, CLIs, and autonomous agents that read, transact, and lend against assets through a unified surface.

Producers don't need to understand each other. Consumers don't need to know which producer originated a position. The Position Object schema is the interface contract.

graph TB
    subgraph Producers
        direction LR
        G["Gravity"]
        J["Joule"]
        M["Megalith"]
        FP["Factor Prime"]
        N["[...]"]
    end
    subgraph "Dark Matter"
        direction LR
        REG["Registry"]
        RE["Rights Engine"]
        IG["Identity Graph"]
        CE["Collateral Engine"]
        GW["MCP / API"]
    end
    subgraph Consumers
        direction LR
        WEB["Web"]
        CLI["CLI"]
        TERM["Terminals"]
        AGENT["Agents"]
    end
    Producers -->|"encode as Position Objects"| GW
    GW --> REG
    GW --> RE
    GW --> IG
    GW --> CE
    GW -->|"read, transact, lend"| Consumers

Five subsystems

• Registry — authoritative record of who owns what. Event-sourced, append-only, fully reconstructible.
• Rights Engine — deterministic evaluation of "can this happen?" for any proposed state transition.
• Identity Graph — KYC, AML, accreditation, jurisdictional eligibility for every participant.
• Collateral Engine — pools, coverage ratios, haircut policies, cure mechanics, fortress periods, atomic pledge settlement.
• MCP/API Gateway — unified ingress and egress. Authentication, rate limiting, idempotency, audit logging.

Data flow

One asset, five subsystems: origination (documents ingested, Position Object created) → registration (canonical entity_id, event log) → enrichment (intelligence linked via entity crosswalk) → transaction (discover, evaluate, consent, settle) → collateral (eligibility, haircut, pledge, surveillance). Every step operates on the same object.

Two interfaces, one schema

• REST API (api.dmtr.ai/v1) — programmatic integration by applications and services
• MCP server — LLM-native integration by AI assistants and autonomous agents

Both operate on the same Position Objects, the same registry, the same rules.

Patient zero: Megalith Fund 2

The first vehicle to run on Dark Matter is Megalith's defense & space fund — a ~$100M vehicle holding 6–8 co-investments, each in its own SPV LLC, encoded in DMTR from day one.

This is not an arbitrary choice of wedge. Defense and space positions have unusually demanding compliance requirements:

• ITAR/EAR sensitivity — international traffic in arms regulations constrain who can hold economic exposure to defense assets
• CFIUS / foreign ownership — Committee on Foreign Investment review requirements create transfer restrictions beyond standard securities law
• Security clearance considerations — certain positions involve investors with clearance obligations that affect eligibility

The Rights Engine is not a convenience here — it is structurally necessary. Manual compliance for these positions is slow, expensive, and error-prone. Machine-enforceable restrictions are the product.

The structure is straightforward: investors subscribe to the top-level Megalith vehicle (standard Reg D offering). The vehicle holds wholly-owned position LLCs. Dark Matter encodes ownership, economics, and transfer rules at the LLC level as internal operating infrastructure. Future secondary transfers flow through the Rights Engine, which enforces the same restrictions counsel would enforce manually — faster, cheaper, with a complete audit trail.

Target: position formation and compliance encoding in under 24 hours from signed subscription docs — vs. 2–4 weeks for traditional SPV setup with manual document review, counsel coordination, and spreadsheet-based tracking.

Patient Zero is greenfield — designed natively on the schema. But the $13 trillion cited in Part 1 is brownfield. Existing positions — LP interests, SPV holdings, fund commitments — enter through an ingest pipeline that extracts ownership state, economic terms, and transfer restrictions from source documents (LPAs, operating agreements, side letters, cap tables) and maps them to the Position Object. The fidelity depends on document quality, but a position encoded at 70% completeness is queryable, discoverable, and collateral-assessable. A PDF in a drawer is none of these.

Dark Matter becomes operational before it becomes universal.

Part 4: Why does it compound?

The network effect from a common schema

Each new producer makes DMTR more valuable to consumers. Each new consumer makes it more valuable to producers. The distinctive property is that producers make each other more valuable through the schema — intelligence from one enriches positions from another, without any direct integration between them.

Cross-producer composability

Three examples:

1. A fund evaluating a secondary venture SPV pulls a structured thesis object from Gravity — conviction at 74%, with one hypothesis flagged at 34%. The buyer adjusts the bid. The thesis and the position were produced by entirely separate businesses. They compose because they share a schema.

2. A 10 MW stranded gas site in Argentina, encoded by Joule as an energy position with capacity, cost-per-kWh, and uptime SLA, becomes collateral-eligible the moment it registers. The collateral engine computes a haircut from the position's liquidity tier and the intelligence layer's counterparty risk assessment. Nobody built that integration. It fell out of the schema.

3. An LP's portfolio agent monitors positions across three producers — venture SPVs from Megalith, energy assets from Joule, credit facilities from Factor Prime — through a single schema query. When a Joule energy position's uptime SLA degrades, the agent flags the coverage impact on a Factor Prime lending facility that includes the position as collateral. Three asset classes, three originators, one query.

Extensibility

The schema is asset-class agnostic. Venture SPVs, energy royalties, private credit facilities, real estate fractions, carbon credits, IP royalties, tokenized fund units — anything encodable as a Position Object gets the full platform on day one. The marginal cost of adding an asset class is the schema mapping, not new infrastructure.

Part 5: Why does the law permit it?

The objection

Private market positions are securities. Operating agreements, LPAs, side letters, consent requirements — these exist for real reasons. Securities law requires registration or exemption. You cannot simply "transfer them programmatically."

This is the right question.

The unitizable LLC

The key structural insight: a co-investment structured as an SPV LLC creates membership interests that are securities — but securities with a well-defined, programmable transfer apparatus. The operating agreement specifies who can transfer, under what conditions, with whose consent. Today those rules live in documents and are enforced by lawyers. With Dark Matter, the same rules are encoded in the Position Object and enforced by the Rights Engine. Same rules. Millisecond execution.

Unitization: an LLC membership interest can be subdivided into smaller transferable units. Each unit is a Position Object with identical rights and restrictions. Fractional liquidity, not regulatory arbitrage.

Tokenization (eventual): once units are Position Objects with deterministic rules, on-chain settlement is a mechanical step. The token is a settlement layer. Rights and restrictions stay in the schema.

The compliance model

We are encoding existing legal structures, not proposing new ones. The compliance stack formalizes obligations that already exist:

• Position Object carries restrictions — lockups, ROFR, consent chains, jurisdictional constraints
• Rights Engine enforces them deterministically — a transfer that would fail legal review also fails the engine
• Identity Graph clears participants before any transfer settles
• Consent Chain requires GP approval with configurable timeouts
• Event Log provides an immutable, auditable history

A lawyer applying these rules manually and the Rights Engine applying them programmatically produce the same output. The difference is cost and latency.

The structure is designed for review by institutional-grade fund counsel before institutional capital is raised against it.

Part 6: Where does it go?

Four phases

1. Schema adoption (live → 12 months) — the Position Object becomes the standard encoding for private market ownership. Patient Zero (Megalith Fund 2) is in formation; the registry and Rights Engine are in development. Critical mass: being outside the schema means being invisible to the network. The seed syndicate is structured as co-launch partners, with each investor listing existing portfolio positions alongside the equity financing — seeding the network with external inventory from day one.
2. Network effects (12–24 months) — multi-sided marketplace dynamics. Cross-producer composability creates value no single producer could build alone.
3. Institutional credit infrastructure (24–36 months) — collateral engine matures into a lending marketplace. Warehouse-to-term securitization. Desk tape as data product.
4. Agent-native capital markets (design horizon) — the schema is machine-readable, which means LLM tooling works today. The longer arc — agents as first-class counterparties negotiating and settling autonomously — follows from the same architecture, with human oversight at policy gates.

Competitive context

Existing platforms address fragments of this problem. Carta handles cap table management. Forge and CAIS provide secondary marketplace access. iCapital and Securitize focus on tokenization and distribution. These are application-layer products; each would benefit from a canonical schema beneath it. None provide that schema.

Dark Matter starts with a structural advantage: $250M in real inventory from Megalith, a live lending arm in Factor Prime, an operational energy vertical in Joule, and an intelligence layer in Gravity already covering thousands of private companies. DMTR is not waiting for its first producer.

Value capture

The fee model follows the infrastructure: per-position registration fees, per-transfer settlement fees, and basis-point fees on collateral under management. Revenue scales with the number of positions encoded, the velocity of transfers, and the size of lending facilities — all of which grow with each new producer. Team, competitive positioning, and go-to-market are covered in companion materials.

The position

Dark Matter sits in the middle: schema, registry, rights engine, collateral engine, settlement bus. The application layer is above it. The client layer is below it. The value of the infrastructure grows with the number of producers encoding into it — and a producer that doesn't exist yet can list into the same schema tomorrow and access the entire network.

When this works at scale, private market positions move the way public securities move today — discoverable, transferable, collateral-eligible, agent-accessible — with the compliance apparatus baked in rather than bolted on.