GOSIX Architecture

Index Weighting Methodology

The mathematical engine balancing thematic purity, physical constraints, and portfolio liquidity in low-Earth orbit compute infrastructure.

1. Strategic Category Allocations

GOSIX structures its exposure across six physical bottleneck categories. Each category represents an indispensable component layer of the space data grid, with its strategic allocation target defined algebraically by variables S_s (where s ∈ {1..6} and ∑ S_s = 100%).

Advanced Semiconductor & Foundry LayerS_1
Sub-System & Laser Link InfrastructureS_2
Space-Grade Compute Edge & PowerS_3
Core Structural ComponentsS_4
Compute Resilience & OrchestrationS_5
Institutional AnchorsS_6
Disclosure Policy: The absolute target values of the segment weights (S_s) are not publicly disclosed to avoid providing specific investment allocations or financial advice.

2. The Weighting Equation

Within each segment s, a constituent i's raw score is calculated using its natural log-scaled USD market capitalization, weighted by qualitative metrics measuring physical bottleneck criticality and thematic focus:

Score_i,s = ln(max(MarketCap_i,USD, 10⁷)) × C_i,s × F_i

Where the input parameters are defined as:

  • Market Cap: The company's total equity valuation converted to USD. A floor of 10⁷ ($10,000,000 USD) is enforced to ensure positive values and filter shell companies.
  • Segment Criticality (C_i,s): A qualitative coefficient between 0.0 and 1.0 representing how vital the company's IP or engineering output is to solving that specific physical segment bottleneck.
  • Thematic Concentration (F_i): A factor between 0.0 and 1.0 reflecting the percentage of enterprise revenue or R&D capital dedicated to space-grade/vacuum environments, favoring pure-plays over diversified conglomerates.

Raw constituent weights are normalized within each segment to sum to 100%:

w_raw_i,s = Score_i,s / ∑_{k ∈ C_s} Score_k,s

3. Multi-Tiered Portfolio Capping Rules

To comply with institutional risk boundaries and prevent thin-liquidity execution errors, GOSIX implements a rigorous four-tier capping framework:

Segment Level

40.00%

Maximum allocation for a single company within its mapped segment, preventing local monopolies from capturing the risk profile.

Standard Global

8.00%

Maximum total index weight for any single constituent in the core semiconductor, optical, or power segments.

Institutional Anchor

4.00%

Maximum global weight for macro-launch or logistics anchors (e.g. SpaceX) to ensure the index focus remains on hardware compute.

Micro-Cap Liquidity

0.50%

Maximum global weight for thin-float small-caps to prevent execution bottlenecks. To avoid turnover whiplash, membership in the micro-cap tier is governed by a two-sided hysteresis band: a constituent not currently uncapped (outside the index, or already capped at 0.50% in the prior composition) must rise above a $350M market cap to escape the cap, whereas a constituent currently uncapped (weight exceeding 0.50% in the prior composition) is only demoted into the tier once its market capitalization falls below $250M.

4. The Biproportional Scaling (RAS) Loop

Capping a company's global weight dynamically reduces its segment contribution, dropping column sums below their targets S_s. To resolve these competing constraints simultaneously, GOSIX runs the iterative RAS Matrix Scaling Algorithm:

1

Initialize Contribution Matrix

Create contribution matrix: M_i,s = S_s × w_capped_i,s

↓
2

Row Scaling (Constituent Caps)

If global sum exceeds its cap C_i (8%, 4%, or 0.5%), scale down its row contributions: M'_i,s = M_i,s × (C_i / W_i)

↓
3

Column Scaling (Segment Targets)

Scale columns back to targets: M''_i,s = M'_i,s × (S_s / ∑_k M'_k,s)

↓
4

Check Convergence & Stabilize

Repeat steps 2 & 3 until contribution matrix elements converge to a stable state within limits.