Intensel

Held deliberately at B because this is a physics and machine learning hybrid where the physics is the differentiator, and the distinction deserves stating rather than smoothing. The strongest technical assets are a proprietary hydrology digital twin and climate physics simulation for typhoons, which are simulation rather than learned models.

Deep learning is named specifically for wildfire analysis, more than 200 damage curve models translate hazard into financial loss, and independent assessments consistently describe the analytics as artificial intelligence driven, with asset-level dollar loss delivered in minutes across portfolios. Remove the learned components and flood and typhoon modelling survives while wildfire analysis and portfolio-scale delivery do not, which is a partial loss rather than a collapse.

The product informs decisions rather than making them, producing risk scores, loss estimates and exposure analysis that investment, underwriting and credit teams act on. Adaptation modelling is the notable feature here because it is prescriptive in a testable way, quantifying how specific physical interventions such as drainage works or raising a building's base would reduce modelled losses, which gives a decision maker something to evaluate rather than a score to accept. Held at B because no confidence interval, uncertainty range or escalation guidance accompanies outputs that feed capital allocation.

Method is disclosed in unusual detail for this index. The financial formulation is stated explicitly as hazard by exposure by vulnerability, which is the standard catastrophe modelling identity and signals actuarial rather than heuristic construction, more than 200 damage curve models are counted, and technique is matched to peril with hydrology-based flood models, physics simulation for typhoons and deep learning for wildfire, which is honest about where learned models are and are not used.

Scenarios follow recognised international climate pathways. Held at B because no validation against realised losses, no accuracy figure and no uncertainty range is published for outputs that inform credit spreads and capital allocation.

Customer types are described with reasonable specificity, spanning commercial real estate owners, sovereign funds, banks and insurers across Asia, the Middle East and globally, and the company is a named partner of the leading real assets sustainability benchmark and is backed by a multilateral development bank's venture arm. No institution is named anywhere, no client count is given, and no outcome or accuracy figure accompanies any deployment.

No boundary statement was located. The platform accumulates asset-level exposure and valuation data across banks, insurers and asset managers who compete for the same properties and underwrite the same risks, and damage curve refinement benefits from observed outcomes across that base. Nothing states whether client portfolio data informs the models, is isolated per client, or is retained after an engagement ends.

No data protection agreement, retention schedule or subprocessor list was located. Personal data exposure is low since the subjects are physical assets rather than people, and the platform does hold precise asset locations and valuations for client portfolios, which is commercially sensitive information about where an institution's capital sits and how exposed it is.

No attestation, certification, trust centre or enumerated framework was located. Institutions submitting complete portfolio asset registers with locations and valuations would examine control documentation during procurement, and nothing is published.

Two established climate disclosure frameworks are named as alignment targets, covering the task force reporting structure and the international sustainability standard for climate, and outputs are explicitly built to satisfy both investors and regulators. The company positions its work as helping clients stay ahead of imminent climate regulation.

Held at B because no financial regulator, prudential stress testing regime or supervisory expectation is named, and banks using climate analytics for capital and provisioning face specific requirements that are not mapped.

No individual is assessed and the adapted exposure is one of the more consequential in this index. Accurately priced physical climate risk withdraws capital and insurance from the places most exposed, which are disproportionately poorer coastal and tropical regions, so a correct model can make coverage unaffordable precisely where it is most needed. The company's stated strength in Asian data makes that effect most pronounced in the markets it serves best. Nothing published addresses distributional consequences, and no analysis of model performance across regions or asset types appears.

No guarantee, indemnity or correction process was located. The affected party is the asset owner whose property is scored, since a high modelled risk figure can raise borrowing costs through a credit spread adjustment or affect insurability, and nothing describes whether an owner learns of the assessment, can inspect the hazard data behind it, or can contest a result derived from modelling rather than observed history.

The scientific basis is identified rather than obscured, with scenarios following recognised international climate pathways under both concentration and socioeconomic frameworks, damage curve models counted, and the hydrology component described as proprietary so a buyer knows it is built rather than licensed. In-house climate scientists and financial specialists are named as part of the capability.

Held at B because no elevation, hazard, exposure or asset data provider is identified, and for physical risk analytics the underlying terrain and hazard datasets determine accuracy as much as the models applied to them.

An interface is offered for portfolio-level screening, which is the mechanism allowing analysis to run across a book rather than asset by asset, and beyond that no named integration, connected system or developer documentation was located. For outputs intended to reach credit, underwriting and disclosure processes, how the results enter those systems is not described.

Delivery is described as a cloud platform with no hosting provider, region selection, residency commitment or private deployment option stated. Clients include sovereign funds and banks across Asia and the Middle East, several in jurisdictions with data localisation requirements, and asset location data for a national portfolio is sensitive in its own right.

Pricing is published, which almost no vendor in this index does: a named solution is stated at between 25,000 and 100,000 US dollars depending on the number of assets analysed, so a buyer knows both the range and the variable that moves it. Held at B rather than higher because only one product line is priced, nothing covers the wider platform or interface access, and no packaging detail accompanies the range.

Buyers span banks, insurers, asset managers, sovereign funds, real estate owners and supply chain operators, with use cases running from due diligence and portfolio screening to stress testing, underwriting and disclosure. Hazard coverage exceeds ten perils under six recognised climate scenarios to 2100, and the company claims particular data accuracy for Asia while operating in the Middle East and globally. Held at B because much of the buyer base sits outside financial services and no market-by-market presence is evidenced.