Sky Forecast Model

A modular Python forecast model for Sky that calculates monthly results and aggregates to quarters. This tool enables scenario planning, sensitivity analysis, and financial projections for the Sky ecosystem.

Source: Laniakea Forecast Model

Overview

The forecast model simulates Sky's financial flows:

Rate Derivations → Agent Valuations → Supply → Revenue → Waterfall

Key Capabilities

• Monthly granularity — Base unit is 1 month, aggregated to quarters
• Scenario inheritance — Bull/bear scenarios can extend base scenarios
• Flexible inputs — Trajectories, persistent changes, one-time impulses
• TMF waterfall — Implements the Treasury Management Function

What It Models

Supply Dynamics

• Base USDS supply growth
• Unrewarded USDS held by users
• Agent-specific supply impacts

Revenue Calculation

• Stability fees based on rates (SOFR + spreads)
• Agent contributions (Spark, Grove, etc.)
• Costs (subsidized borrow, sRUSDS costs)

TMF Waterfall

• Security & Maintenance allocation
• Backstop buffer filling
• Staking rewards distribution

Agent Activity

• Binary active/inactive based on launch month
• Agent-specific valuations
• Farmable vs sellable token types

Configuration Format

Scenarios are defined in YAML files with these components:

Baseline Values

baseline:
  spark_market_cap: 350000000
  base_usds: 9650000000
  sofr: "0.0370"
  savings_rate_spread: "0.0030"
  gross_revenue_rate_spread: "0.0030"
  security_rate: "0.21"

Trajectories (Interpolated Values)

trajectories:
  base_usds:
    points:
      1: 9650000000
      6: 12000000000
      12: 16000000000
    mode: linear    # or "step"

Persistent Changes

changes:
  3:
    sofr: "0.0360"    # Changes in month 3, persists
  7:
    spark_market_cap: 450000000
    security_rate: "0.15"

One-Time Impulses

impulses:
  10:
    backstop_withdrawal: 5000000  # Only applies to month 10

Agent Launches

agent_launches:
  spark: 1
  grove: 4
  keel: 6
  obex: 7

Scenario Inheritance

Scenarios can extend other scenarios:

# bull_2026.yaml
extends: base_2026

baseline:
  spark_market_cap: 500000000  # Override

changes:
  5:
    base_usds: 13000000000     # Add new change

Resolution order:

1. Load parent scenario
2. Merge baseline (child overrides)
3. Merge changes (child adds/overrides)
4. Merge impulses, trajectories, etc.

Key Parameters

Rate Parameters

Supply Parameters

Agent Parameters

Output Format

The model produces markdown reports with monthly and quarterly breakdowns:

## Q1 (Jan-Mar)

| Month | USDS Supply | Gross Rev | Net Rev | Staking |
|-------|-------------|-----------|---------|---------|
| Jan   | 9,800.00M   | 35.12M    | 18.54M  | 10.85M  |
| Feb   | 10,060.00M  | 36.05M    | 19.10M  | 11.18M  |
| Mar   | 10,320.00M  | 36.98M    | 19.67M  | 11.51M  |
| **Q1**| —           | **108.15M** | **57.31M** | **33.54M** |

## Annual Summary

| Quarter | Gross Revenue | Net Revenue | Staking Rewards |
|---------|---------------|-------------|-----------------|
| Q1      | 108.15M       | 57.31M      | 33.54M          |
| Q2      | 121.40M       | 65.22M      | 38.15M          |
| ...     | ...           | ...         | ...             |
| **Year**| **537.65M**   | **294.48M** | **172.28M**     |

Simplifications vs Full TMF

Note: The forecast model intentionally simplifies for practical forecasting:

The full TMF mechanics are documented in the whitepaper appendices.

Running the Model

CLI

# Run a scenario
python -m skyforecast run config/scenarios/base_2026.yaml

# Generate HTML dashboard
python run_dashboard.py

# Generate markdown report
python run_report.py

Output Files

• report.txt — Markdown report
• dashboard.html — Interactive HTML dashboard

Use Cases for /m/skyaccounting

1. Settlement Projections

Use the model to project what future settlements might look like under different conditions.

2. Scenario Comparisons

Compare base vs bull vs bear scenarios to understand sensitivity.

3. Parameter Impact Analysis

"What happens if SOFR drops by 50bps?" — Run a modified scenario to find out.

4. Agent Launch Impact

Model how new agent launches (Grove, Keel, etc.) affect revenue.

5. Backstop Analysis

Track how quickly the backstop buffer fills under different revenue scenarios.

Key Insights

Revenue Sensitivity

Revenue is highly sensitive to:

• SOFR (base rate)
• USDS supply growth
• Rate spreads

Backstop Dynamics

At current scales (~$200M net revenue):

• Phase 1 applies (25% floor)
• Backstop fills relatively slowly
• Most surplus goes to staking rewards

Agent Impact

Active agents contribute:

• Direct revenue from their operations
• Token farming/selling income
• Indirect supply growth effects

Summary

1. Sky Forecast Model is a Python tool for financial projections
2. Monthly granularity with quarterly aggregation
3. Flexible scenarios with inheritance and overrides
4. Models TMF waterfall (simplified for forecasting)
5. Available in Laniakea docs — open source
6. Use for settlement projections, scenario analysis, parameter sensitivity

Related Documents