debt_leverage_profile
Net debt/EBITDA ~3.5x; a 20% rise in interest rates increases annual debt service ~R$80–100M, compressing net income ~8–12%
Inferred
Agent_Inference
interest_rate_sensitivity
~70% of debt indexed to TJLP/IPCA; 20% rate spike raises financing costs ~R$90M annually, reducing ROE by ~2 percentage points
Inferred
Agent_Inference
geopolitical_supply_exposure
High intensity; European gas dependency on Russia exposed structural energy security vulnerabilities.
Medium
GICS-commodity-overlay-v1
supply_chain_dependency
Brazilian chemical inputs (chlorine, aluminum sulfate) from domestic oligopolists; electricity from national grid with Itaipu treaty dependency
Inferred
Agent_Inference
international_expansion_readiness
No material international revenue; sovereign currency devaluation exposure is effectively null — 100% Brazil-domiciled revenues
Inferred
Agent_Inference
geographic_footprint
Operates exclusively in Minas Gerais state, Brazil; zero international revenue exposure; all currency risk is BRL domestic
Inferred
Agent_Inference
commodity_exposure_profile
High intensity; commodities: Natural Gas, Coal, Uranium, Crude Oil, Copper (grid), Lithium (storage); geopolitical: European gas dependency on Russia exposed structural energy security vulnerabilities.
Medium
GICS-commodity-overlay-v1
vendor_lock_dependency_score
Sabesp/CEDAE peer-comparable: electricity (~20% opex) and chemical suppliers each <15%; no single vendor exceeds 30% threshold
Inferred
Agent_Inference
business_model_type_primary
Asset-intensive regulated utility; no material cloud infrastructure dependency — core operations run on legacy OT/SCADA systems
Inferred
Agent_Inference
business_model_type_secondary
Billing and ERP systems (SAP-based) use on-premise or hybrid hosting; cloud termination risk is low-moderate, not operationally critical
Inferred
Agent_Inference
switching_cost_profile
Minimal API coupling risk; legacy SCADA and billing systems are proprietary/on-premise; no significant third-party API dependencies
Inferred
Agent_Inference
howey_test_risk_index
Not applicable — revenue model is regulated water/sewage tariffs, not an investment contract; Howey Test risk is negligible
Inferred
Agent_Inference
regulatory_burden_tier
Very High
Medium
GICS-regulatory-overlay-v1
data_sovereignty_risk
Exposed to Brazil's LGPD (Lei Geral de Proteção de Dados); GDPR/CCPA not applicable; LGPD compliance costs estimated R$10–20M
Inferred
Agent_Inference
antitrust_exposure_flag
State-granted monopoly concession in MG; antitrust risk low but subject to CADE oversight on concession renewals and M&A activity
Inferred
Agent_Inference
regulatory_exposure_profile
Very High burden; regimes: FERC, NERC, EPA, NRC, State PUCs, DOE; Rate-case lag and clean-energy mandates compress returns on regulated asset base.
Medium
GICS-regulatory-overlay-v1
revenue_model_type
~95% recurring — regulated tariff-based billing tied to water/sewage consumption under long-term state concession contracts
Inferred
Agent_Inference
monetization_vector
Volumetric tariff on water consumption and sewage collection/treatment; regulated by ARSAE-MG with periodic tariff revisions
Inferred
Agent_Inference
pricing_architecture
ARSAE-MG sets tariffs via periodic reviews (~4-year cycles); inflation pass-through via IPCA indexation; limited pricing discretion
Inferred
Agent_Inference
pricing_power_rating
Moderate — tariff increases require regulatory approval; IPCA indexation provides inflation protection but limits upside pricing power
Inferred
Agent_Inference
target_gross_margin_bracket
Gross margin ~45–50%; EBITDA margin ~35–40%; constrained by high energy and chemical costs within regulated cost structure
Inferred
Agent_Inference
churn_vulnerability_index
No free-rider problem; monopoly utility with mandatory billing; churn is effectively zero — residential/commercial connections are captive
Inferred
Agent_Inference
headcount_cost_structure
Revenue growth is sublinear to headcount; network expansion capex-driven not labor-driven; doubling revenue requires ~20–30% headcount growth
Inferred
Agent_Inference
marginal_cost_of_growth
Marginal cost of incremental connections is primarily capex (pipes, treatment capacity); opex per new connection decreases at scale
Inferred
Agent_Inference
franchise_compliance_risk
Not a franchise model; regulated concession structure; compliance drift risk replaced by concession renewal risk with Minas Gerais state government
Inferred
Agent_Inference
customer_acquisition_metric
At 10x scale (not feasible in single state); incremental connection cost ~R$3,000–5,000/household; CAC irrelevant in captive monopoly model
Inferred
Agent_Inference
network_effect_present
No traditional network effects; natural monopoly infrastructure — value does not increase with user count; durability is regulatory, not network-based
Inferred
Agent_Inference
asset_efficiency_ratio
AI displacement risk low; physical water/sewage infrastructure cannot be digitally substituted; AI can optimize operations but not replace assets
Inferred
Agent_Inference
recession_resistance_tier
Tier 1 recession-resistant — water/sewage is non-discretionary essential service; volume demand inelastic; historical revenue decline <3% in downturns
Inferred
Agent_Inference
customer_segment_primary
Residential households (~75% of revenue); no single customer >1% of revenue; concentration risk is minimal
Inferred
Agent_Inference
customer_segment_secondary
Commercial and industrial users (~20% of revenue); municipal governments for bulk supply (~5%); diversified, low concentration
Inferred
Agent_Inference
characteristic_occupations
["11-0000 Management Occupations", "13-0000 Business and Financial Operations Occupations", "15-0000 Computer and Mathematical Occupations", "17-0000 Architecture and Engineering Occupations", "23-0000 Legal Occupations", "41-0000 Sales and Related Occupations", "43-0000 Office and Administrative Support Occupations", "47-0000 Construction and Extraction Occupations", "49-0000 Installation, Maintenance, and Repair Occupations"]
High
SOC-2018/GICS-overlay
agent_automatable_labor_share
0.34 (HIL — ~34% of characteristic roles agent-automatable)
Medium
SOC-2018 + agentic-exposure-v1
capital_expenditure_profile
Capex ~R$1.5–2.0B annually; predominantly growth/compliance (sewage universalization mandate by 2033); limited legacy-to-future reallocation
Inferred
Agent_Inference
sec_cik
Not SEC-registered; listed on B3 (Brazil) as CSMG3; no SEC CIK number applicable
Inferred
Agent_Inference
ticker
CSMG3 (B3); trades at ~4–5x EV/EBITDA, discounting regulatory tariff risk and high leverage, not geopolitical commodity supply risk
Inferred
Agent_Inference