Accelerating sustainability and resilience with AI-powered innovation

At Climate Week NYC, our teams are deep in conversations about using AI to advance sustainability and advancing the sustainability of AI.

Thousands of global leaders are gathering at Climate Week NYC 2025, one of the largest events of its kind focused on climate action. From business and government to science and civil society, the conversations this year reflect both a shared urgency and a growing sense of possibility.

Across industries, leaders are using AI to turn constraints into catalysts for innovation. Companies are using AI to help balance competing objectives such as reducing costs and environmental impacts while driving growth.

The business case is clear: Morgan Stanley reports that companies investing in climate risk mitigation are seeing average returns of 8X on their initial investment.¹ And, according to the World Economic Forum, every dollar invested in climate adaptation and resilience can generate up to $19 in avoided losses.²

AI in action: How leaders are driving change

Sustainability challenges often span functions, geographies, and time horizons, and AI is proving valuable in helping people collaborate to solve highly complex problems. For example:

• Faced with increasingly frequent flooding, the City of Stuttgart needed a faster way to prepare. Traditional geospatial modeling would have taken months. Instead, by partnering with Esri, Microsoft, and NVIDIA, city planners built a full-scale 3D digital twin in just 24 hours -a 99% improvement in processing time. With AI-powered simulations, these city planners can now visualize rainfall, model water flow, and test mitigation strategies in near real time. The result: faster response, smarter planning, and a more resilient city.

• In Japan, supermarket chain Super Hosokawa and logistics partner Imamura Shoji used Azure Databricks, Azure OpenAI, and Snowflake to build a demand forecasting system that shares two-day-ahead predictions across the supply chain. The impact was immediate: food waste dropped by over 50% for key products, and trial items outperformed expectations, even during periods of declining sales. These results are inspiring broader AI adoption across Japan's food logistics ecosystem.

• In Washington state, researchers at Pacific Northwest National Laboratory partnered with Microsoft to reimagine how we discover next-generation battery materials. Using AI and Azure Quantum Elements, they narrowed down 32 million possibilities to 18 viable candidates in just 80 hours-a process that would traditionally take years. This is a glimpse into the future of scientific research and development: AI-powered discovery, faster time-to-impact, and new frontiers in energy and materials.

Advancing the sustainability of AI with innovations in datacenter cooling

As AI adoption accelerates, managing its resource use is a strategic imperative, to enable customers and partners to scale AI responsibly and competitively. Three areas of focused investment for Microsoft include (1) optimizing datacenter energy and water efficiency, (2) advancing low-carbon materials, and (3) improving the energy efficiency of cloud and AI services.

The latest cloud and AI technologies run on chips that consume more power than previous generations, and the more power that runs through a chip, the hotter it gets. Future generations of chips for AI are expected to become even more powerful, with even greater demands on the cooling systems in datacenters.

To help address this problem, Microsoft has successfully tested a new microfluidic cooling system with up to three times better cooling performance than cold plates, depending on the workloads and configurations involved. Taking the heat signatures of chips, we're able to identify the hot spots and then etch channels into the back of the silicon chips to direct liquid coolant more efficiently to the hot spots, optimizing how we do cooling. These channels are micrometers in size, similar in size to a human hair.

As part of the prototyping effort, the team used AI to help optimize a bio-inspired design to cool chips' hot spots more efficiently than straight up-and-down channels.

Microfluidics is part of our whole-systems approach to optimizing every part of the cloud and AI stack, from datacenters to servers to silicon. Because cooling impacts so many aspects of cloud infrastructure design, from server density to rack density to power management and load balancing between servers, these innovations promise improvements for sustainability as well as other metrics, such as cost, reliability, speed, and consistency.

Getting started with AI for sustainability

Sustainability progress starts with curiosity. Begin by identifying where your organization can reduce risk and improve efficiency, whether through better visibility, smarter supply chain decisions, or new value streams.

To learn more about how AI can help your organization make tangible progress toward your sustainability and resilience goals, read Insight to Impact: AI Use Cases to Advance Sustainability.

¹ Corporate Sustainability: Long-Term Value Creation Opportunity, Morgan Stanley , July 30, 2015.

² With climate risks set to slash earnings, what can CEOs do?, World Economic Forum, December 12, 2024.