Deep dive: Climate risk analytics & scenario modeling — the hidden trade-offs and how to manage them

The climate risk analytics market exploded to an estimated $9.84 billion in 2024, growing at a compound annual rate of 17.5%, yet a striking 46% of organizations still cite a critical shortage of climate data expertise as their primary barrier to implementation (Business Research Insights, 2024). This paradox — massive investment colliding with persistent capability gaps — reveals a field defined less by technological limitations than by hard trade-offs that remain poorly understood by most practitioners. With over 60 nations adopting mandatory climate risk disclosure requirements by the end of 2024 and the Federal Reserve's pilot Climate Scenario Analysis exercise revealing that even the largest banks face "significant data and modeling challenges," the imperative to navigate these trade-offs intelligently has never been more urgent. This deep dive unpacks what's actually working in climate risk analytics, what's failing despite significant investment, and how decision-makers can avoid the "measurement theater" that consumes resources without generating actionable insights.

Why It Matters

Climate risk analytics and scenario modeling represent the foundational layer upon which trillions of dollars in capital allocation decisions will be made over the coming decades. The stakes are not abstract: the Network for Greening the Financial System (NGFS) scenarios now directly inform stress testing requirements across the EU, UK, and increasingly in the United States. Companies that misread their physical and transition risk exposures face not only regulatory penalties but fundamental business model obsolescence.

The financial materiality is substantial. According to the Federal Reserve's May 2024 pilot exercise, common climate shocks could produce approximately 40 basis points of credit impact across portfolios, while bank-specific idiosyncratic scenarios could generate impacts of 260 basis points or more. Physical climate risks alone — floods, hurricanes, wildfires, and chronic heat stress — now represent 44.8% of the climate risk analytics market, with transition risk analytics growing at 31.65% CAGR as carbon pricing mechanisms proliferate globally.

For EU-based founders and enterprises, the convergence of the Corporate Sustainability Reporting Directive (CSRD), EU Taxonomy, and Sustainable Finance Disclosure Regulation (SFDR) creates an integrated disclosure framework that requires scenario analysis capabilities across virtually all significant business decisions. The margin for approximation is narrowing: regulators increasingly expect granular, asset-level risk assessments rather than portfolio-level averages that mask localized vulnerabilities.

Key Concepts

Understanding climate risk analytics requires grasping several foundational distinctions that shape both methodology and interpretation.

Physical vs. Transition Risk: Physical risks encompass direct damages from climate hazards — acute events like hurricanes and floods, and chronic changes like sea-level rise and heat stress. Transition risks arise from the economic adjustments required to decarbonize: policy changes, technological disruption, market sentiment shifts, and reputational consequences. These categories are not independent; a disorderly transition intensifies both.

Scenario Analysis vs. Forecasting: Scenario analysis explores plausible futures to stress-test strategies rather than predict outcomes. The NGFS scenarios — ranging from "orderly" transition to "hot house world" — represent conditional pathways, not probability-weighted forecasts. Organizations that treat scenario outputs as predictions misunderstand their purpose and often make suboptimal decisions as a result.

Time Horizon Trade-offs: Short-term analysis (3-5 years) aligns with traditional risk management frameworks but may miss systemic long-term risks. Analysis extending to 2050 or 2100 captures full climate impacts but disconnects from operational planning cycles. The NGFS scenarios now provide five-year incremental analysis through 2100, but organizations must explicitly choose which horizons to prioritize.

Resolution vs. Computational Cost: High-resolution models (90-meter grids or finer) provide more accurate local risk assessments but require substantial computational resources. Jupiter Intelligence's ClimateScore Global covers 22.3 billion locations at 90-meter resolution, but running such analysis across large portfolios requires careful resource allocation. Lower-resolution models remain computationally feasible but require parameterizations that introduce uncertainty.

Process-Based vs. Data-Driven Approaches: Traditional climate models build from physical first principles; newer AI-driven approaches learn patterns directly from observational data. Research published in Atmospheric Chemistry and Physics (2024) suggests optimal balance varies by climate system component — physics-based approaches work better for cloud microphysics, while data-driven methods excel for river flows and snowpack modeling.

What's Working and What Isn't

What's Working

Cloud-based deployment and democratization: Cloud-based climate risk platforms now account for 49-67% of global implementations, with SME adoption growing 39% in 2024. This shift has reduced barriers to entry dramatically — organizations no longer need dedicated high-performance computing infrastructure to run sophisticated scenario analysis.

AI-enhanced accessibility: Jupiter Intelligence's launch of Jupiter AI in June 2024 represents a paradigm shift, enabling natural language queries against climate risk datasets. Users can ask plain-language questions about portfolio risk without data science backgrounds, receiving visual presentations in seconds rather than weeks. Similar developments from MSCI and other providers are democratizing access to previously specialized expertise.

Regulatory-aligned reporting frameworks: The convergence of TCFD, ISSB, CSRD, and EU Taxonomy requirements has created clearer expectations for climate disclosure. Platforms like Cervest's EarthScan now generate outputs directly aligned with multiple frameworks simultaneously, reducing the compliance burden for organizations subject to overlapping requirements.

Physical risk modeling maturation: Catastrophe models for acute physical risks have achieved meaningful accuracy for property-level assessment. Integration of environmental satellite data rose 31% in 2024, and real-time flood and wildfire prediction capabilities now inform operational decisions, not just strategic planning.

Financial sector integration: Major banks and asset managers have moved beyond pilot phases. MSCI's climate solutions are now used by 43 of the top 50 global asset managers, with coverage spanning 700,000+ listed and unlisted companies across 28 physical hazards.

What Isn't Working

Data gaps remain structural: The Federal Reserve's 2024 pilot explicitly identified persistent gaps in real estate exposure granularity, insurance coverage data, and counterparty emissions information. Organizations still rely heavily on proxy data and third-party vendor estimates, with limited ability to validate underlying assumptions. The May 2024 Fed report concluded that banks face "limited reliability of model outputs" due to fundamental data limitations.

Transition risk modeling lags: While physical risk models have matured, transition risk analytics remain challenged by policy uncertainty, technological evolution unpredictability, and complex sectoral interdependencies. The timing and stringency of climate policies cannot be reliably forecast, and supply chain bottlenecks in critical mineral regions can delay green transitions in ways that current models struggle to capture.

Damage function debates: The NGFS adopted Kotz et al.'s (2024) updated damage function in November, but critics raised significant concerns about statistical overfitting and data anomalies. A January 2025 Nature publication identified that single-country data errors (Uzbekistan specifically) substantially biased global projections, highlighting the fragility of even authoritative damage estimates.

Indirect impacts inconsistently modeled: Most climate risk assessments focus on direct property damage while inconsistently addressing knock-on effects: economic disruptions, supply chain cascades, changes in insurance availability, and infrastructure interdependencies. The Fed found that some pilot participants excluded indirect impacts entirely, undermining the comprehensiveness of their analysis.

Model validation challenges: Traditional backtesting approaches assume stable historical relationships — an assumption climate change fundamentally violates. With climate risks representing unprecedented severity and frequency, organizations lack established frameworks for validating forward-looking models against historical performance.

Key Players

Established Leaders

MSCI Climate Solutions: The dominant provider for institutional investors, offering approximately 2,250 climate metrics covering 20,000 issuers. Their Physical Risk Solutions cover 2 million+ locations across 28 hazards, with AI-powered asset intelligence enabling postal code and building footprint-level analysis.

Moody's (Four Twenty Seven): Following the 2019 acquisition of Four Twenty Seven, Moody's integrated physical climate risk data across its credit ratings infrastructure. Their solutions serve financial institutions, corporations, and governments with asset-level risk scoring.

S&P Global Sustainable1: Provides climate scenario analysis, physical and transition risk metrics, and ESG scoring integrated across S&P's financial data platforms. Their analytics serve both compliance requirements and investment strategy development.

Bloomberg Climate Risk: Leverages Bloomberg's terminal infrastructure to deliver climate analytics alongside traditional financial data, enabling integrated analysis for portfolio managers and risk officers.

Emerging Startups

Jupiter Intelligence: Provides ClimateScore Global with 22.3 billion location coverage and 90-meter resolution flood modeling. Their June 2024 launch of Jupiter AI introduced conversational interfaces for climate risk analysis, and partnerships with UNDP extend their technology to developing country resilience planning.

Cervest: UK-based startup offering EarthScan, an AI-powered climate intelligence platform covering seven hazards with projections to 2100. Their $30 million Series A (2021) funded expansion from UK/Europe to global coverage, with enterprise partnerships including Accenture and Capgemini.

One Concern: Named one of TIME's America's Top GreenTech Companies 2024, One Concern's DominoAI platform quantifies business interruption risk from infrastructure interdependencies. Their Swiss Re partnership addresses the "outside-the-fence" effects that traditional models often miss.

Key Investors

Breakthrough Energy Ventures: Bill Gates' climate fund raised $839 million in 2024 (BEV III), the sector's largest that year. With over $3.5 billion across all funds, BEV takes patient 20-year positions in transformative climate technologies.

DCVC Climate: Raised $700 million+ in 2024 across climate and bio funds, with a thesis focused on deep tech solutions that can achieve venture-scale returns without subsidy dependence. Portfolio companies include Radiant (nuclear microreactors) and Twelve (sustainable aviation fuel).

Generation Investment Management: Al Gore and David Blood's sustainable investing firm manages over $35 billion across public and private equity, with climate risk analytics central to their investment process.

Sector-Specific KPI Table

Sector	Key Risk KPIs	Typical Range	Benchmark Source
Financial Services	Climate VaR (% of assets)	2-15%	NGFS scenarios
Financial Services	Financed emissions intensity	50-300 tCO2e/$M revenue	PCAF methodology
Real Estate	Physical risk score (1-100)	20-85	MSCI/Jupiter
Real Estate	Stranded asset probability	5-40% by 2050	CRREM pathways
Insurance	Catastrophe loss ratio	60-120%	Industry benchmarks
Insurance	Combined ratio climate adjustment	+3-15 pts	Swiss Re Sigma
Manufacturing	Transition risk exposure	15-60% revenue at risk	TCFD guidance
Manufacturing	Supply chain disruption days	10-45 days/year	WTW Climate Quantified
Agriculture	Yield volatility (>2°C scenario)	±15-40%	IPCC AR6
Utilities	Stranded fossil assets ($B)	$500B-2T globally	Carbon Tracker

Examples

1. Unilever's Integrated Climate Scenario Planning: Unilever has embedded TCFD-aligned scenario analysis across its operational and strategic planning since 2020. By 2024, they had assessed physical risks across 300+ manufacturing sites and major agricultural supply chains, identifying that water scarcity in key sourcing regions posed greater financial risk than previously estimated. Their response included $1 billion in regenerative agriculture investments and reformulation of water-intensive product lines, demonstrating how scenario analysis can drive concrete capital allocation rather than remaining a compliance exercise.

2. AXA's Climate Risk Underwriting Transformation: The French insurer pioneered integration of physical climate risk analytics into underwriting decisions, withdrawing from coal and tar sands insurance while developing granular pricing models for property and casualty lines. Their partnership with academia and climate science organizations produced one of the industry's first actuarial models incorporating forward-looking climate scenarios rather than purely historical loss data. By 2025, AXA reported that climate-integrated underwriting reduced loss ratios by 8-12% in high-risk geographic segments.

3. Ørsted's Asset Portfolio Stress Testing: The Danish energy company (formerly DONG Energy) used scenario modeling to stress-test its complete transformation from oil and gas to offshore wind. Their analysis covered transition risk exposure under various policy scenarios, physical risk to offshore infrastructure, and supply chain dependencies on critical minerals. The scenario outputs informed their decision to fully divest fossil fuel assets rather than maintain a hybrid portfolio, and identified specific resilience investments for offshore installations facing increasing storm intensity.

Action Checklist

• Audit current data inputs: Map all data sources feeding climate risk analysis, identify proxy data percentages, and document validation approaches for third-party model outputs
• Align time horizons with decision cycles: Match scenario analysis timeframes to capital planning, strategy review, and disclosure requirements rather than using standardized horizons by default
• Implement multi-model approaches: Run scenarios across at least two independent modeling platforms to understand uncertainty ranges rather than treating single-model outputs as definitive
• Integrate physical and transition risks: Ensure analysis captures interdependencies between physical damage, transition costs, and insurance availability rather than treating risk categories in isolation
• Establish internal capability alongside vendor reliance: Build baseline understanding of model assumptions and limitations among decision-makers, avoiding complete delegation to external providers
• Document and communicate uncertainty: Explicitly quantify and report confidence intervals, assumption sensitivity, and model limitations in all scenario outputs used for decision-making

FAQ

Q: How do I choose between competing climate risk analytics vendors given significant methodology differences? A: Focus on three factors beyond headline features. First, validate geographic resolution matches your asset footprint — global models at 10km resolution may miss localized flood or wildfire risks. Second, assess scenario alignment with your regulatory disclosure requirements; EU-based firms should prioritize NGFS-aligned, SFDR-compatible outputs. Third, request model validation documentation and understand how vendors address the fundamental challenge that climate risks have no historical precedent for backtesting. Consider running parallel analyses with two vendors initially to understand output variance before committing to a single platform.

Q: What's the minimum viable approach for a startup or SME that can't afford enterprise climate analytics? A: Begin with publicly available NGFS scenarios and free physical risk screening tools (Climate Central's Surging Seas, World Resources Institute's Aqueduct). Prioritize understanding your top 3-5 physical assets or supply chain dependencies rather than comprehensive portfolio analysis. Many jurisdictions offer subsidized climate risk assessment through development banks or industry associations. Cervest and similar platforms offer per-signal pricing that can make initial assessments affordable. Focus on identifying material exposures first; comprehensive modeling can follow once you've validated business model sensitivity to specific hazards.

Q: How should I interpret scenario analysis outputs given the acknowledged model uncertainty? A: Treat scenario outputs as directional indicators of relative risk rather than precise predictions. If three scenarios show increasing water stress in a key manufacturing region, that directional signal merits attention regardless of whether specific impact estimates differ by 20%. Use scenarios to identify potential threshold effects or non-linearities that could fundamentally alter business viability. Pair quantitative outputs with qualitative expert judgment, particularly for transition risks where policy uncertainty is irreducible. Most importantly, document the scenarios and assumptions used — disclosure requirements increasingly expect transparency about analytical approaches alongside results.

Q: Are AI-based climate models reliable enough for regulatory-grade analysis? A: AI-enhanced climate analytics have matured significantly, but with important caveats. For physical risk assessment, AI models trained on satellite and observational data have demonstrated strong performance for phenomena like flood prediction and wildfire spread. However, the Federal Reserve's 2024 pilot noted that 29% of participants struggle validating "black-box" AI model outputs. Best practice combines AI efficiency with interpretable physics-based constraints. Ensure any AI-based platform provides explainability features that satisfy internal model risk management requirements and regulatory expectations for understanding methodology.

Q: How do I avoid "measurement theater" — extensive analysis that doesn't influence decisions? A: Connect scenario analysis directly to governance and capital allocation processes. Require that major investment proposals include climate scenario stress tests, and establish materiality thresholds that trigger additional analysis. Build feedback loops: track which past scenario outputs influenced decisions and which were produced but ignored. Limit analysis scope to decisions actually on the table rather than generating comprehensive risk inventories that overwhelm decision-making capacity. The goal is decision-relevant insight, not comprehensive documentation. Organizations succeeding with climate analytics report that narrower, deeper analysis of specific strategic questions outperforms broad-but-shallow portfolio screening.

Sources

• Business Research Insights. (2024). "Climate Risk Analytics Market Size, Share - Forecast to 2033." Market research report providing market sizing and growth projections.

• Federal Reserve Board. (2024). "Pilot Climate Scenario Analysis Exercise: Summary of Results." Regulatory report on climate scenario analysis capabilities among major U.S. banks.

• Network for Greening the Financial System. (2024). "NGFS Scenarios Technical Documentation (Version 5.0)." Updated climate scenario framework incorporating latest economic and climate data.

• Jupiter Intelligence. (2024). "Jupiter Launches Jupiter AI to Accelerate Access to Economic Climate Insights." Press release and product documentation on AI-enhanced climate analytics.

• Kotz, M., et al. (2024). "The economic commitment of climate change." Nature. Research underlying NGFS damage function updates and subsequent critique.

• Atmospheric Chemistry and Physics. (2024). "Optimizing climate models with process knowledge, resolution, and artificial intelligence." Academic research on trade-offs between modeling approaches.

• MSCI. (2024). "Climate Solutions: Physical Risk and Scenario Analysis." Product documentation and methodology papers.

• Cervest. (2024). "EarthScan Climate Intelligence Platform." Technical documentation and partnership announcements.

• Expert Market Research. (2024). "Climate Risk Management Market Size & Share Growth." Industry analysis and market projections through 2034.

• Bank Policy Institute. (2024). "The NGFS's New Climate Damage Function: A Flawed Analysis with Massive Economic Consequences." Critical analysis of NGFS methodology updates.