debt_leverage_profile
Net debt/EBITDA ~2.5x; 20% rate rise increases annual interest expense ~BRL 300-400M, compressing net margin ~1-1.5pp
Inferred
Agent_Inference
interest_rate_sensitivity
High sensitivity: ~70% of debt CDI-linked; 20% Selic increase meaningfully erodes BRL 1.5B+ annual EBITDA buffer
Inferred
Agent_Inference
geopolitical_supply_exposure
High intensity; OPEC+ supply decisions and Russia-Ukraine conflict drive price volatility.
Medium
GICS-commodity-overlay-v1
supply_chain_dependency
Strait of Hormuz (imported fuel feedstocks) and Panama Canal (LPG tanker routes to Brazil)
Inferred
Agent_Inference
international_expansion_readiness
Minimal international revenue; ~95% Brazil-domestic; currency devaluation risk is BRL/USD internal cost, not cross-border revenue risk
Inferred
Agent_Inference
geographic_footprint
Overwhelmingly Brazil-centric; BRL depreciation raises USD-denominated fuel import costs, squeezing Ipiranga/Ultragaz margins directly
Inferred
Agent_Inference
commodity_exposure_profile
High intensity; commodities: Crude Oil, Natural Gas, Coal, Refined Petroleum Products, Uranium, Steel (equipment); geopolitical: OPEC+ supply decisions and Russia-Ukraine conflict drive price volatility.
Medium
GICS-commodity-overlay-v1
vendor_lock_dependency_score
Petrobras supplies >50% of fuel distributed by Ipiranga; single-supplier dependency is significant and partially non-substitutable short-term
Inferred
Agent_Inference
business_model_type_primary
Physical distribution and retail fuel/LPG network; cloud termination would disrupt logistics IT but not core physical operations
Inferred
Agent_Inference
business_model_type_secondary
Specialty chemicals (Oxiteno divested 2022); remaining model is fuel distribution, LPG, and drugstores (Extrafarma divested 2022)
Inferred
Agent_Inference
switching_cost_profile
Low API coupling risk; operations are physical-asset-intensive; ERP/logistics software switching costs moderate but not existential
Inferred
Agent_Inference
howey_test_risk_index
Fails Howey Test; fuel distribution and retail are clearly commercial commodity transactions, zero securities law risk
Inferred
Agent_Inference
regulatory_burden_tier
High
Medium
GICS-regulatory-overlay-v1
data_sovereignty_risk
Low GDPR exposure (Brazil-only operations post-Oxiteno sale); LGPD (Brazil) compliance required for loyalty/pharmacy data
Inferred
Agent_Inference
antitrust_exposure_flag
Moderate; Ipiranga holds ~25% Brazilian fuel distribution market share; CADE monitors distributor pricing conduct regularly
Inferred
Agent_Inference
regulatory_exposure_profile
High burden; regimes: EPA, FERC, DOE, CFTC, OSHA, SEC; Accelerating emissions mandates and methane rules threaten capex economics.
Medium
GICS-regulatory-overlay-v1
revenue_model_type
~95% transactional (fuel/LPG volume-based); minimal recurring contracted revenue; highly volume-and-margin-spread dependent
Inferred
Agent_Inference
monetization_vector
Volume-spread model: margin per liter/kg on fuel and LPG distributed through owned and franchised station networks
Inferred
Agent_Inference
pricing_architecture
Thin-margin pass-through pricing tied to Petrobras reference prices; limited independent pricing power; margin ~3-6% on fuel revenue
Inferred
Agent_Inference
pricing_power_rating
Low; fuel pricing largely dictated by Petrobras and market competition; Ipiranga competes on service and loyalty, not price
Inferred
Agent_Inference
target_gross_margin_bracket
Consolidated gross margin ~8-12%; Ultragaz slightly higher; fuel distribution inherently low gross margin, high volume business
Inferred
Agent_Inference
churn_vulnerability_index
Moderate free-rider risk at franchised stations; brand loyalty programs (km de Vantagens) help retention but switching is frictionless
Inferred
Agent_Inference
headcount_cost_structure
Revenue growth largely headcount-sublinear; incremental volume absorbed by existing infrastructure; logistics headcount scales modestly
Inferred
Agent_Inference
marginal_cost_of_growth
Marginal cost of growth is capital-intensive (new stations, tanks) but operationally sublinear; ROIC improves at scale within network
Inferred
Agent_Inference
franchise_compliance_risk
Meaningful compliance drift risk across ~8,000+ Ipiranga-branded stations; fuel adulteration and safety violations are recurring CADE/ANP issues
Inferred
Agent_Inference
customer_acquisition_metric
At 10x scale, station network economics face diminishing returns; Brazil market penetration already high; CAC rises in underpenetrated regions
Inferred
Agent_Inference
network_effect_present
Weak direct network effects; loyalty program creates mild data/reward network; physical station density creates localized competitive moat
Inferred
Agent_Inference
asset_efficiency_ratio
AI displacement risk low for physical distribution; route optimization and demand forecasting AI adoption can improve logistics margin ~1-2pp
Inferred
Agent_Inference
recession_resistance_tier
Tier 2 resilience; fuel is near-essential but volumes fall ~5-10% in deep recessions; LPG cooking gas more recession-resistant
Inferred
Agent_Inference
customer_segment_primary
Individual consumers (retail fuel, LPG residential) representing majority of transaction volume across Ipiranga and Ultragaz networks
Inferred
Agent_Inference
customer_segment_secondary
Commercial/fleet customers and resellers (B2B fuel supply, industrial LPG); no single customer >5% of revenue
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", "19-0000 Life, Physical, and Social Science 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", "53-0000 Transportation and Material Moving Occupations"]
High
SOC-2018/GICS-overlay
agent_automatable_labor_share
0.33 (HIL — ~33% of characteristic roles agent-automatable)
Medium
SOC-2018 + agentic-exposure-v1
capital_expenditure_profile
Capex ~BRL 1.5-2B annually; split between maintaining existing station/tank infrastructure and selective new-station/EV charging expansion
Inferred
Agent_Inference
sec_cik
null
Inferred
Agent_Inference
ticker
UGP (NYSE ADR); trades at ~5-7x EV/EBITDA, modest discount reflecting BRL currency risk, Petrobras supply dependency, and thin-margin distribution model
Inferred
Agent_Inference