The Confidential Settlement Layer for private markets

Continuum.

Atomic, private GP-led secondary closes on Canton.

HackCanton · Season 2 Built on Canton + Daml Finance
01 / 12

02 — The problem

continuation close

A continuation close is choreography that can break mid-step.

Assets move old-fund→new-fund, exiting LPs are paid, rolling LPs get new units, secondary buyers wire in — sequentially, not simultaneously. A multi-week, lawyer-and-spreadsheet operation where any leg can fail mid-close.

leg failureA leg fails mid-close and the deal is stranded between two states.
partial closeA partial close is a legal catastrophe — unwinding is contested and costly.
term leakagePer-LP roll-vs-exit terms leak across the syndicate.
T+0Assets transfer old-fund → new-fundsettled
T+2Exiting LPs paid out in cashsettled
T+4Secondary buyer wire delayedcounterparty funding fails to arrivefailed
T+?Rolling LPs receive new unitsblocked
T+?Close certifiedblocked

Three legs settled. One failed. The close is now a legal problem, not an operational one.

03 — Why now

market · regulator · rails

Three curves converged in the last 18 months.

$75B GP-led secondaries in 2024 — roughly 44% of a $160B market scaling toward $400B. Jefferies / Evercore 2024
SEC named the GP information-asymmetry conflict and pushed a third-party fairness-opinion norm. Private Fund Adviser rule
$355M raised by Digital Asset from a16z at ~$2B. DTCC, Goldman, Apollo, Citadel on Canton's institutional rails. Canton adoption, 2024–25

The deal volume is here, the regulator has named the conflict, and the institutional settlement rail is live and funded — at the same moment.

04 — The product

Continuum

Every leg of a continuation close settles in one atomic transaction.

01

Atomic close

Assets, cash, and unit issuance settle all-or-nothing. No leg can strand the deal between two states.

02

Private election

Each LP's roll-vs-exit election stays sealed. No LP — and no operator — sees another's terms.

03

Fairness window

The LPAC and the regulator get a need-to-know disclosure window — provable fairness without leaking the book.

Built on Canton + Daml Finance. The settlement guarantee and the privacy guarantee are the same primitive.

05 — How it works

four steps · one close
1

GP posts terms

The GP publishes the continuation-vehicle terms to the deal contract.

2

Sealed elections

Each LP submits a sealed roll-vs-exit election. No LP sees another's.

3

Blind commitment

The secondary buyer commits a price the exiting LP cannot see.

4

One click → atomic close

Assets → new fund, cash → exiting LPs, units → rolling LPs + buyer. All-or-nothing.

Then the LPAC / regulator window opens for the fairness review. The operator stays blind throughout.

06 — Why Canton — the moat

not a database · not ZK
RequirementPostgresEth + ZKCanton
No shared book across partiesNoYesYes
No party sees every electionNoYesYes
Atomic multilateral DvPNoYesYes
Regulator can see — need-to-knowNoNoYes
Compliant — not anonymityYesNoYes

Atomic multilateral DvP + sub-transaction privacy + selective disclosure is Canton-native. No other rail clears all five rows.

ZK is a non-compliant black box — privacy must be need-to-know, not anonymity.
— Yuval Rooz, CEO, Digital Asset

07 — Demo

five parties · one ledger

Each party sees only its own view — until the atomic close, then the regulator window opens.

GPterms
vehicle NAV$420.0Melections in4 / 4 sealed
LP — Apexprivate
my electionROLL · 100%other LPs
LP — Brevanprivate
my electionEXIT · 100%buyer price
Buyercommit
my bid$118.4Mexiting LP
Regulator / LPACwindow
opens atpost-closescopeneed-to-know
blind party viewsone-click atomic closefairness window opensdeal #2: buyer reused in one click

The flywheel: a buyer pre-cleared on deal #1 reuses its on-chain eligibility credential and commits to deal #2 with zero re-onboarding — onboarded once, reusable across closes.

Storyboard mock on Canton LocalNet · live walkthrough: [ loom link — to record ]

08 — Market — honest

revenue pool, not notional

We price the revenue pool, never the notional.

Notional flow
$115B
GP-led volume in 2025 — the headline, but not the addressable revenue.
Revenue pool · wedge
$30–70M
Per-close fees ($75–250k) × rising deal frequency + buyer-side SaaS. Network lifts share, not take-rate.

The continuation wedge is a real, fundable business on its own. The reusable-participant network is what makes it venture-scale — repeat usage and switching costs, not a bigger fee.

Platform map — engine, then the network moat, then the headline-TAM wedges

Wedge 1 · nowContinuation engineGP-led secondaries / continuation-fund rolls$30–70M
MoatReusable-participant railbuyers onboarded once, reused across deals — never pooled capitalnetwork
Wedge 2SRTsynthetic risk transfer — IMF-named systemic whitespace$100–125M
Wedge 3Electionconfidential buybacks / tendersexpansion
Wedge 4Private creditconfidential lifecycle & settlementexpansion

09 — Business model

fee pool, not notional

Per-close fees plus platform SaaS, distributed through fund admins.

01

Per-close fee — $75–250k

Charged on each organized continuation close. Bound to the fee pool, never the notional.

02

Platform SaaS

Recurring access for fund admins and GPs running continuation-as-a-service.

03

Channel rev-share

Revenue share with the fund administrator who brings the GP relationships.

Comparable regulated-infra utilities

CompanyWhatValue
VersanaConsortium utility · $9T loans$125M+
OSTTRA → KKRPost-trade infra$3.1B
SecuritizeTokenized-asset infra · 15–50bps$1.25B
Discipline

Bind revenue to the fee pool, never the notional. The notional is the headline; the fee pool is the business.

10 — Go-to-market

cold-start-free

No two-sided liquidity to bootstrap. Each close is a discrete organized deal.

The cold-start is asymmetric in our favour: secondary dry powder is at a record ~$300B+, deals are the scarce side — so the deal arrives first and capital follows. We land a fund administrator — Citco, Apex, Alter Domus — as the channel (the Versana rail-first playbook), reach their GP clients, and the network compounds from there. 6–12 month cycles, de-risked with paid pilots.

Phase 1 · rail

One design partner

A single fund-admin design partner and one mocked-then-real continuation close. Deal #1 needs zero network.

Phase 2 · flywheel

Reusable participants

Buyers cleared on deal #1 are reused across the administrator's book of GP closes — onboarded once.

Phase 3 · platform

Platform wedges

Extend the same rail and participant base into SRT and Election.

11 — Moat & competition

incumbents are off-chain

No one does the confidential close.

Coller, Pantheon, Citco run this off-chain in spreadsheets and law firms. iCapital's Canton DLT is a pre-trade data layer — no settlement. Nasdaq's reusable-buyer workflow is off-chain and non-atomic. None settle atomically; none prove fairness without leaking the book.

The moat compounds once Continuum is the system of record — and the reusable-participant network is the switching cost.

01Canton-native atomicity + privacy — the settlement and privacy guarantee are one primitive
02The regulator window — provable fairness, need-to-know, no leakage
03Reusable-participant network — onboard once, reused across deals; network effect, no pooled capital
04Fund-admin distribution — the channel owns the GP relationships
05Switching costs — once it is the system of record for closes

12 — Team · build · vision

the ask
Team
  • Founder + Claude Code product & build
  • Designer institutional UX
  • VC advisor secondaries access
The build
4-week

Canton MVP — Daml Finance Batch close, per-LP private elections, regulator observer window, multi-party UI.

The vision

The Confidential Settlement Layer for private markets. Continuation today; a reusable participant rail; SRT, Election, Private-Credit next.

Privacy must be need-to-know, not anonymity.
— Yuval Rooz · the regulator must be able to see
↑ ↓ · scroll · F