[ Documentation · 06 / 08 ]
How a position leaves
Four exits have returned $412M. Principal routes back to Reserve, gain routes to Compound. Lock-ups run from same-epoch to 84 months, and outside Reserve there is no early exit.
01
A position leaves the book in one of three ways. Only two of them return money.
There is no fourth route. A member cannot sell their claim to another member — the claim does not trade, there is no order book for it, and no secondary market exists for it.
02
4 positions have been realised, returning $412M. All four were entered in Epoch 00 and exited in Epoch 02.
| Company | Category | Realised | From | Entry → exit |
|---|---|---|---|---|
| Databricks | Data & tooling | $184M | Anchor | Epoch 00 → Epoch 02 |
| Rippling | Fintech | $96M | Anchor | Epoch 00 → Epoch 02 |
| Deel | Fintech | $78M | Shield | Epoch 00 → Epoch 02 |
| ElevenLabs | Models | $54M | Momentum | Epoch 00 → Epoch 02 |
Anchor produced $280M of the $412M on its own. The two largest realisations in the book came out of its most conservative equity vault, which is what a late-stage mandate is for: it buys companies that are close to a liquidity event and it is paid when that event arrives rather than when a story changes.
Four exits is a small sample. Read the realised figure as evidence that the routing mechanism works, not as a track record.
03
Where the proceeds go is mechanical. Nothing about it is discretionary and nothing about it waits on a committee.
Principal returns to Reserve
It settles at par, in the epoch, and is available to redeem or reallocate at the next window. A member's original capital comes back as cash rather than as another position.
The gain routes to Compound
Unless the member opts out at the window. Compound then writes it into mid-stage companies at a maximum of $20.0M per name.
The exit is written to chain
The realisation, the proceeds and the route they took, recorded on the epoch it happens.
Compound holds $84.3M today against $412M realised, because most of that principal went back to members rather than forward into the vault. Its size is therefore a lagging read on how much the rest of the book has exited — it only grows when the other nine vaults do.
A realisation is not automatically cash in hand
Principal is. Gain is reinvested by default, into Compound, and opting out is a member action taken at a window rather than a standing setting. A member who wants every realisation returned as cash has to say so each time. Compound is also the only vault where a loss cannot touch member principal — every dollar in it was a gain before it was a position — which does not make its positions safer, and it is rated R3 on the exposure it actually holds.
04
Every vault states a term, and the term is the strategy rather than a penalty.
| Vault | Tier | Lock-up | Why that term |
|---|---|---|---|
| Reserve | R1 | None. Same-epoch exit | Cash and bills · Instant exit |
| Anchor | R2 | 24 months | Pre-liquidity · Low variance |
| Shield | R2 | 18 months | Hedged against the public index |
| Perpetual | R3 | Evergreen | Core compounders · Never fully realised |
| Structure | R3 | 36 months | Converts, SAFEs, secondaries |
| Compound | R3 | 12 months | Realised proceeds · Recycled |
| Keystone | R4 | 48 months | Ten positions · No eleventh |
| Momentum | R4 | 12 months | Marked-up rounds · Fast follow |
| Frontier | R5 | 60 months | Compute, energy, defence, robotics |
| Genesis | R5 | 84 months | Seed and Series A · Longest hold |
Frontier's 60 months is 60 months because there is no secondary market of any depth for most of that book: a member who needs the money before the term is not going to get it. Momentum's 12 months is the shortest of any equity strategy in the House because a book with a 14-month median holding period should not trap capital for four years. Perpetual has no term at all — it is evergreen, it never fully realises, and it has no scheduled liquidity of any kind.
05
Stated plainly, because it is the question that matters most.
The test to apply before allocating
If there is a real possibility of needing the capital before the term, the allocation belongs in Reserve or it does not belong here at all. A 48-month lock-up on money that is needed in eighteen months is not a risk that can be managed later — the mechanism to manage it does not exist.
06
The lock-up is the smaller constraint. Three larger ones sit underneath it and none of them is removed by putting a claim on chain.
What the chain gives is a verifiable record — of what is held, at what mark, on whose behalf, and what has left. It does not create a buyer. The distinction between a transparent book and a liquid one is the single most important thing to be clear about before allocating. Verification is covered in Custody & attestation.