For a long time, climate and ESG lived in the safe zone of disclosure. Companies reported exposures, published sustainability reports, and responded to questionnaires, but the underlying assumption remained unchanged: climate risk was important, but not yet decisive. It informed narratives more than it made capital differences. It was compliance.

That assumption is now breaking down.

The Context Has Fundamentally Shifted

The last few years have altered the context in which climate risk is understood. According to the World Meteorological Organization, recent years have ranked among the hottest ever recorded, with 2024-2025 marking unprecedented temperature anomalies across multiple regions. Just days ago, the United Nations weather agency confirmed that the decade between 2015 and 2025 was the hottest eleven-year period on record, and that the atmosphere is trapping greenhouse gases at a rate exceeding any prior measurement.

What people often fail to recognise is that rising temperatures, weather unpredictability, and supply chain disruption do not affect only the environment. They affect industries. Agriculture, logistics, energy, manufacturing, each is exposed in ways that can have lasting consequences for food security and economic stability.

Climate risk is no longer a distant, modelled abstraction sitting in 2030 or 2050 scenarios. It is showing up in underwriting losses, supply chain disruptions, infrastructure stress, and increasingly, in the cost and availability of capital. And when time collapses, decision systems must evolve.

From Lagging Indicator to Leading Intelligence

Historically, ESG intelligence was constructed from static and lagging inputs: annual disclosures, ratings, and fragmented datasets that provided a partial and often outdated view of risk. Artificial intelligence is changing that fundamentally, moving climate data from a lagging indicator to a live intelligence input.

By integrating satellite data, hyperlocal weather feeds, ground sensors, logistics flows, regulatory inputs, and unstructured local-language information, AI systems can construct a continuously updating map of risk. More importantly, they simulate consequence. They can estimate how a flood event translates into asset downtime, how heat stress affects labour productivity, how drought alters supply availability, and how all of this flows into financial performance.

In doing so, AI transforms climate from a reporting dimension into a decision variable embedded within capital allocation, procurement, and strategy. It takes risk out of the realm of abstraction and into the ledger.

The Insurance Sector Has Already Moved

The market is beginning to reflect this shift, and nowhere is it more visible than in insurance, a sector that has spent years investing in climate risk modelling to protect its own balance sheet. Despite the United States being among the most prominent opponents of mandatory ESG regulation, its insurance industry has categorised climate risk as a real and quantifiable financial exposure.

Firms such as State Farm and Allstate have publicly withdrawn or restricted coverage in high-risk geographies, particularly regions exposed to recurrent wildfires and extreme weather events. The downstream consequence is a higher cost of capital for businesses and property owners in those areas. This is the real shift: from awareness to enforced risk mitigation, mediated not by regulation, but by pricing.

What Investors Are Actually Demanding Now

Investors are recalibrating what they expect from ESG intelligence. The demand is no longer for better scores or more comprehensive disclosures. Most ESG regulations came into being after major industrial and financial disasters, and embedded within those frameworks is a significant body of non-financial metrics that, when properly modelled, serve as powerful predictors of financial risk.

At Impactree.ai, over three years of research and development have gone into identifying the metrics across industries that most directly drive financial risk exposure. AI tools are now enabling us to quantify those metrics against both financial and non-financial disclosures, across listed and unlisted entities, at a speed that was previously impossible. AI is exceptionally good at taking a defined process and amplifying it at scale.

The Next Wave: Beyond ESG Disclosures

The intelligence being built in this space still requires significant development to enable AI tools to understand the context of the data they ingest. That is where the next generation of startups can make a meaningful difference, taking inputs from multiple sources, using sustainability frameworks to identify crucial drivers, and predicting outcomes and mitigation requirements.

This requires thinking about sustainability beyond ESG disclosures and beyond a single environmental lens. Sustainability operates at a resilience layer: it makes enterprises stronger for tomorrow by addressing environment, social, and governance dimensions together. The social and governance pillars are often underweighted in current systems, yet they are equally crucial drivers of intelligence and risk.

Artificial intelligence sits at the intersection of all three. It compresses detection timelines, increases analytical precision, and translates complexity into decision-ready scenarios. For markets like Southeast Asia, where the next generation is preparing to take over family-led enterprises, this data is critical for shaping internal investment strategy over time.

The Tension That Cannot Be Ignored

The infrastructure required to power advanced AI, particularly large-scale modelling and real-time analytics, is energy intensive. Data centres, which form the backbone of AI systems, are themselves significant consumers of electricity. According to estimates referenced by the International Energy Agency, data centre energy demand is projected to rise materially over the coming decade, driven in part by AI workloads. Beyond energy, the water requirements of large-scale data centres are becoming a material constraint in water-stressed regions.

This creates a tension that cannot be ignored. The same systems that enable us to better understand and manage climate risk may also contribute to the very emissions we are trying to reduce. This does not invalidate the use of AI, but it does impose a discipline. It forces organizations to move from indiscriminate deployment to prioritised intelligence, to ask not just what can be modelled, but what should be modelled.

From Hindsight to Foresight

For decades, resilience was treated as a cost, something that reduced downside but did not necessarily create upside. In a climate-constrained world, that framing is reversing. Resilience, when informed by intelligence, becomes a source of advantage. It allows firms to allocate capital more efficiently, avoid stranded assets, anticipate disruptions, and identify emerging opportunities ahead of the market.

Artificial intelligence is accelerating this shift by making climate intelligence usable at the point of decision. But the outcome is not predetermined. Organizations that treat AI as a reporting upgrade will continue to operate reactively. Those that treat it as a decision system will begin to see returns within two years of investment. The movement from disclosure to decision is, ultimately, a movement from hindsight to foresight.

And in a world where climate risk is no longer waiting its turn, foresight is not optional. It is the only viable strategy.