Deterministic computation engine

Your fund's CFO,
computed exactly.

Waterfall calculations, capital calls, LP reporting, and compliance analysis — powered by a deterministic math engine and AI judgment. Fund data never leaves your machine.

Open ATLASExplore Enterprise

Startup OS · Ecosystem

ATLAS lives alongside Startup Atlas — discovery and intelligence for emerging markets. Explore partner materials and sibling products below.

Built for institutional rigor. 10 functional modules. Zero fund data exposure.

10 modules. One system.

Everything a fund CFO does — from waterfall sensitivity to LP quarterly letters — in a single deterministic platform.

Fund Dashboard

Fund metrics and KPIs in one view — committed capital, fee drag, and the path from DPI to TVPI.

Waterfall Engine

European & American waterfalls with full sensitivity analysis. Carry, catch-up, clawback — computed to the penny.

Deal-level version: CapStack (free)

Capital Calls

Projected call schedules with realistic deployment pacing. Fee drag computed across fund life.

Performance

DPI, TVPI, and RVPI from the same model, with unrealized marks kept separate from cash returned.

Scenario Lab

Four-scenario framework: base, top quartile, downside, catastrophic. See where the fund breaks.

Ownership Decay

Round-by-round dilution modeling with pro-rata exercise and exit sensitivity tables.

Deal-level version: CapStack (free)

J-Curve Model

Net cash position tracking, DPI crossover analysis, and recovery driver classification.

Portfolio

Deal-level positions, cost, and current value — the ledger behind the fund-level ratios.

CFO Advisor

Claude-powered advisory for LP letters, compliance analysis, and strategic fund operations questions.

For the quarterly LP letter itself, see LPRP.

Import & Analyze

Paste, upload, or fetch source files, then review what the model can and cannot use.

Live output · $100M fund · default parameters

This is what you get.

Real numbers computed from a $100M fund model. Every figure is deterministic — no estimates, no approximations.

Fund I · $100M committed · European waterfall · 25 deals · 2% mgmt fee · 20% carry
Committed Capital
$100.0M
GP Commit: 2%
Total Fee Drag
$14.2M
14.2% of committed
Deployable Capital
$30.0M
30% of committed
Projected Net TVPI
2.51x
DPI: 1.13x
Scenario Analysis — LP MOIC at fund close
Base Case
1.44x
LP MOIC
Top Quartile
2.53x
LP MOIC
Downside
0.60x
LP MOIC
Catastrophic
0.22x
LP MOIC

Sample from the public default model · fund math in the app runs on your machine

Open with your fund →

Architecture decisions that matter

1

Fund data never leaves your machine

In the app, waterfall, carry, and LP math runs in your browser with no API call. Figures on this page are a sample from the public default model.

2

DPI first, TVPI second, IRR last

Built on the LP priority framework. Every metric is contextualized. Unrealized markups are flagged, not celebrated.

3

AI for judgment, math for certainty

The computation engine handles what can be computed exactly. Claude handles what requires interpretation.

Stop approximating.
Start computing.

Every number your LPs see should be reproducible, auditable, and exact. ATLAS makes that the default.

Launch ATLAS