Climate resilience needs granularity (and how we can help you get there)

I. The financial exposure is real

Physical climate risk is translating into measurable financial losses. Fixed asset losses for listed companies are projected at up to $610 billion per year by 2035: this manifests as an annual earnings drop of up to 7.3% annual earnings, rising up to 10.1% by 2045. EBITDA at risk from physical damage could reach up to 25% by 2050. These projections are almost certainly conservative: They model gradual, linear warming and do not account for tipping points, where thresholds trigger self-reinforcing, irreversible changes already within reach at current warming levels.

The mechanisms are straightforward. Extreme weather (floods, wildfires, prolonged heat) damages productive infrastructure and disrupts operations. Chronic heat is increasingly recognized as one of the most costly drivers: As temperatures rise to dangerous levels, working hours must be shortened, compressing output and labor productivity.

These direct costs compound through cost of capital. As exposure to extreme weather becomes financially material, lenders price it in. A 2025 Bloomberg study found that physical risk exposure is associated with an increase in weighted average cost of capital of +22 basis points per ten climate risk points. Unless a business can demonstrate it understands and is managing its climate risk exposure, businesses will face higher borrowing costs, tighter insurance terms, and reduced access to capital.

Insurance premiums are projected to nearly triple to $1.3 trillion by 2040, with roughly a quarter of that increase attributable to climate risk. In the shorter term, MSCI’s analysis found that insurance premiums for physical climate risks were projected to increase 50% by 2030. Some categories of risk may become uninsurable altogether. Businesses that cannot quantify their exposure will be the least equipped to negotiate terms, or find cover at all.

II. Most organizations are exposed without knowing it

Climate resilience begins with understanding where a business is vulnerable. And by that measure, the majority of organizations are operating without the insight they need.

Only 10% of banks have achieved full quantification of climate risk drivers. Only 17% of companies currently disclose the financial impact of identified climate risks. These are not laggards: They are representative of where the market broadly sits.

The common approach today is regional estimates: broad geographic assessments that satisfy the minimum requirements for disclosure but offer little to guide capital allocation, investment planning, or operational response. They tell a business it is "in a high-risk zone." They don't tell it which assets are most exposed, what the expected financial loss is, or what mitigation investment would make sense.

Regulatory requirements demand comparability across time, and go further than that: They seek proof that climate risk analytics genuinely drive risk management and capital allocation, not that they simply exist for compliance's sake. Compliance becomes the output of an existing business process.

III. Granularity is what changes that

For climate risk data to inform a business decision, it must be specific enough to be attributed to a particular asset with an ascribed financial value or quantifiable output. That is what allows a business to calculate expected damage, compare it against the cost of adaptation, and make a defensible case for capital allocation.

Two barriers consistently stall progress. First, access to quality data: Granular physical climate risk analysis depends on high-resolution, downscalable climate models, which are either technically complex to handle in-house or expensive when their insights are purchased pre-packaged from third-party providers. Second, analytics capability: Even where data exists, translating it into repeatable, auditable, decision-useful outputs requires streamlined pipelines that most sustainability and risk teams currently don't have.

For this to stick, it cannot be a one-off project. Finance teams need results they can trace. Sustainability teams, already managing long lists of competing priorities, need workflows that can be rerun without rebuilding from scratch each cycle. And regulatory requirements demand comparability across time.

IV. Dataiku can help you achieve this

We addressed these constraints using Dataiku. The platform's extensibility made it straightforward to build a visual plugin connecting to PhysRisk: an open-source climate risk calculation engine developed by the OS-Climate community, a Linux Foundation-backed initiative.

PhysRisk takes the geolocation and asset type of physical infrastructure, applies user-defined time horizons and climate scenarios, and outputs exposure to key climate hazards alongside an estimate of expected damage expressed as a proportion of the asset's insurable value. That figure — the Average Annual Loss, or AAL — is what makes climate risk financially quantifiable at asset level.

The Dataiku project built around this plugin handles the full workflow: centralizing and preparing data from across an organization, running it through the risk calculation engine, disaggregating outputs for visualisation, and embedding data quality checks at each step. The result is an asset-level physical climate risk assessment that is traceable, explainable, and repeatable, meeting the needs of finance, sustainability and compliance teams.

In short, Dataiku helps a business move from regional guesses to specific, financially-grounded forecasts, through a workflow built to be reused and relied upon.

The next article picks up where this one ends: Given a granular picture of physical climate risk, how should it inform capital allocation decisions?