Introduction: The New Nexus of Ecology and Intelligence

1The environmental services sector is at a critical inflection point, driven by the convergence of escalating ecological crises and the exponential growth of AI. This expansion is not speculative; it is propelled by a potent combination of urgent climate imperatives, increasingly stringent government regulations such as the European Union's goal to reduce greenhouse gas emissions by 55% by 2030, and the maturation of AI technologies themselves, which are now capable of sophisticated analysis and real-time decision-making.^{1, 2, 3, 4} AI is transitioning from a niche academic tool to a foundational technology for climate change mitigation, resource management optimization, and biodiversity conservation.⁵

2Against this dynamic backdrop, we are positioning ourselves not merely as a participant but as a vanguard of what we define as ecological transformation. In an era where AI often elicits apprehension, our strategy presents a compelling counter-narrative, demonstrating the technology's profound potential for positive environmental impact. Our core thesis is that AI is the indispensable catalyst for achieving our ambitious strategic goals, enabling a fundamental shift in how environmental services are delivered.

3Our AI strategy represents a comprehensive, multi-layered approach that integrates top-down corporate vision with bottom-up operational deployment and a forward-thinking governance framework. It serves as a definitive case study in how we can harness AI not just for incremental efficiency, but to fundamentally re-architect its services for a more resilient and sustainable future. This strategy embodies a crucial evolution in our industry, moving beyond a model of reactive problem-solving toward one of proactive system design. The traditional paradigm involved reacting to environmental challenges: treating wastewater as it arrived, collecting waste on fixed schedules, and repairing equipment after it failed. The new AI-driven model we are pioneering is inherently proactive and systemic. It involves predicting wastewater plant loads to pre-emptively adjust treatment processes, optimizing waste collection routes in real-time based on sensor data to reduce fuel consumption, and performing prescriptive maintenance to prevent asset failures before they occur. This transformation elevates our value proposition significantly, shifting our role from that of a utility service provider to a strategic partner that offers our municipal and industrial clients enhanced sustainability, resilience, and operational intelligence.

The Strategic Blueprint: AI as the Catalyst for GreenUp

4At the heart of our AI-driven transformation is our flagship strategic program for 2024-2027, GreenUp.^{6, 7, 8} This initiative is our definitive response to a global landscape marked by multiple crises, articulating a vision for an ecology that transforms and protects. It is a comprehensive blueprint that embeds AI as a core component for achieving our ambitious environmental and financial objectives.

5The GreenUp program is structured around three central pillars that define its mission: Decarbonize, Depollute, and Regenerate. These are not abstract ideals but are underpinned by concrete, quantifiable targets that demonstrate the scale of our ambition. We have committed a substantial €4 billion investment to the program, with a significant portion prioritized for three growth boosters: Hazardous Waste Treatment, Water Technologies, and Bioenergy, all areas where AI applications are being heavily deployed.^{6, 7, 8} The program's targets by 2027 include treating 9 million metric tons of hazardous waste and pollutants, saving 1.5 billion cubic meters of fresh water, and erasing 18 million metric tons of CO₂ equivalent (Scope 4) compared to 2023 levels,^{6, 7, 8, 9} as part of our commitment to achieve Net Zero by 2050.

6Critically, our GreenUp program is designed with a dual mandate that tightly couples environmental progress with financial performance. The program has delivered significant results, with synergies reaching €435 million and efficiencies achieving €398 million in 2024, both exceeding targets since the program's inception. AI and digital transformation have been cited as key drivers of these efficiencies.^{7, 8} This deliberate linkage reveals a sophisticated understanding that for sustainability initiatives to be durable, they must also be economically sound. Our public statements reinforce this, identifying AI as one of the principal enablers and a crucial catalyst for realizing the GreenUp vision.

7This dual mandate of pursuing both profit and planetary health is central to the resilience and scalability of our strategy. By demonstrating that AI-driven environmental solutions can generate significant operational efficiencies and financial returns, we create a powerful, self-reinforcing loop. The cost savings and new revenue streams generated by AI applications can fund further investment in green technologies and innovation. This approach reframes sustainability not as a regulatory burden or a cost center, but as a core driver of competitive advantage and profitability. This makes our strategy inherently more robust and appealing to shareholders, ensuring its longevity and insulating it from the volatility of political or economic cycles. It establishes a business model where ecological responsibility and financial success are not mutually exclusive but are, in fact, mutually dependent.

Architecting an Intelligent Enterprise: Our Foundational AI Platforms

8To execute our ambitious GreenUp strategy, we are constructing a sophisticated enterprise AI architecture built upon two foundational and complementary platforms. This dual-platform approach combines the power of a centralized, industrial-scale AI engine with the agility of a democratized, human-centric tool, creating a comprehensive ecosystem for innovation and operational excellence.

The Industrial Brain: Talk to My Plant (TTMP) and the Mistral AI Alliance

9The first pillar of this architecture is Talk to My Plant (TTMP), our industrial AI platform conceived to fundamentally reimagine how we manage and analyze data across our vast global footprint. The vision for TTMP is to evolve our thousands of facilities from passively monitored sites into smart, connected ecosystems. The platform facilitates a shift from reactive management to what we call conversational co-piloting, with the ultimate objective of enabling facilities to perform autonomous actions under human supervision to maximize efficiency and environmental performance.

10The technological heart of TTMP's generative AI capability is our strategic alliance with the French AI innovator, Mistral AI, announced in February 2025.^{10, 11, 12, 13, 14} This partnership integrates Mistral's powerful Large Language Models (LLMs) with our immense repository of operational data and domain expertise. This synergy enables groundbreaking functionality: the ability to have a conversation with the plant. An operator, engineer or even a contract manager can query a complex industrial system using natural language and receive precise, data-driven answers and recommendations in real time. This system is already being tested at twelve pilot sites across France, the United Kingdom, Taiwan and other international locations, with our intent for rapid expansion.^{11, 12, 13, 14} The core functionalities this partnership aims to deliver are threefold:

1. Simplified Access to Knowledge, allowing technicians to instantly retrieve critical technical data;

2. Greater Transparency, enabling stakeholders to interact with systems intuitively;

3. Improved Operational Efficiency, driven by optimized monitoring and precise, proactive recommendations.

11The decision to work with Mistral AI reflects our commitment to diversifying our AI partnerships and leveraging the best technologies available globally. Mistral AI's focus on performance, transparency, and compliance with European regulatory frameworks makes it a valuable partner for our operations. This collaboration complements our broader AI strategy, which includes partnerships with multiple technology providers to ensure we can deliver optimal solutions across our diverse global operations. By working with various AI providers, we maintain flexibility in our technology choices while ensuring compliance with local regulations and meeting the specific needs of our clients worldwide. This approach reflects our commitment to innovation and our ability to adapt to the evolving AI landscape.

The Democratic Engine: VeoliaSecureGPT

12Complementing the top-down, industrial scale of TTMP is VeoliaSecureGPT, an internally developed generative AI and agentic platform we designed to democratize AI capabilities across our entire organization. We have made this platform accessible to all Veolia connected employees, serving as a secure, intelligent co-pilot for their daily activities.

13Developed with remarkable agility in just two months by our internal IT teams, VeoliaSecureGPT provides our employees with instant access to the company's vast, specific knowledge base. It streamlines routine tasks such as writing, translation, research, and information synthesis, while improving decision-making with AI-powered insights. Crucially, it operates within a secure, proprietary environment, ensuring that sensitive corporate and client data remains confidential ‒ a key concern when using public generative AI models.

14This dual-platform approach reflects a balanced strategy for advancing digital transformation. TTMP provides a top-down, industrial-scale solution, while SecureGPT enables a bottom-up, more human-centered path for innovation. The aim is not to replace employees, but to augment their capabilities. By providing a secure and accessible AI tool to all employees, we’re enabling a broad base of users to identify operational improvements and explore new use cases. For example, a field operator might use the tool to troubleshoot a local issue, while an analyst in a regional office could synthesize data and draft reports. This creates a feedback loop where frontline insights can help shape and refine larger-scale systems like TTMP ‒ supporting faster iteration and helping build AI fluency across the organization.

AI in Action: A Portfolio of High-Impact Environmental Solutions

15While strategy and architecture provide the blueprint, the true measure of our AI initiative lies in its real-world applications. We are deploying a diverse and growing portfolio of AI-powered solutions that translate high-level ambition into tangible operational and environmental impact across our core business lines. This portfolio demonstrates a mature strategy of applying the right AI tool ‒ from digital twins and deep learning to computer vision and bioacoustics ‒ to solve specific, high-value problems.^{15, 16}

16The following table synthesizes a few of our key AI solutions, providing value under the Veolia Hubgrade umbrella brand, connecting their core technology to their quantifiable benefits, providing a clear overview of the breadth and depth of our deployment.

Business Line	Key Functionality	Core AI Technology	Quantifiable Sustainability/Operational Benefits
Water	Real-time optimization of wastewater treatment processes by sending instructions to PLCs. Predicts loads and volumes.	Digital Twin, Predictive Analytics, Advanced Process Control (APC)	Up to a 60% reduction in carbon footprint; Up to 100% savings on chemicals; Ensures regulatory compliance; Mitigates extreme weather impacts.
Water	Optimizes energy use and control in medium-sized wastewater plants (up to 50,000 PE). Calculates set points for PLCs.	Deep Learning, Remote Monitoring	5% to 15% savings on global electricity consumption; Enhanced operational control and drift detection.
Water/ Industry	24/7 monitoring of water treatment assets to optimize performance and predict failures.	Asset Performance Management (APM), Gen AI, Anomaly Detection	Prevents unplanned downtime; Extends asset life; Reduces waste and resource consumption through continuous optimization.
Waste	Identifies pollution in collected waste on trucks and assesses street litter levels from sweepers.	Computer Vision, AI-powered Cameras	Increases sorting at source; Decreases pollution in MRFs/Incinerators; Enables dynamic resource allocation for city cleaning.
Energy	Manages real-time operations of energy production and distribution systems (e.g., heating networks).	Energy Modeling, AI-based Optimization	Optimizes energy consumption; Reduces carbon footprint; Improves compliance with energy performance contracts.
Biodiversity	Autonomously monitors and identifies over 100 species (bats, birds, etc.) to evaluate ecosystem health.	Bioacoustic Sensing, Species Identification AI	Provides validated biodiversity impact data; Supports ecological health monitoring for Veolia and its clients.

Governance and Vision: Pioneering Responsible AI in Industry

Shaping the Future: Our Leadership in the Coalition for Sustainable AI

17Our leadership recognizes that technological prowess alone is insufficient for the successful, long-term deployment of AI in critical public infrastructure. Building and maintaining public, client, and regulatory trust is paramount. Consequently, we have made a deliberate and visible commitment to establishing an ethical and responsible AI governance framework, positioning ourselves as a pioneer in the industrial application of trustworthy AI.

18Our CEO, Estelle Brachlianoff, has publicly endorsed the Pledge for a Trustworthy AI in the World of Work and aligned Veolia with the Coalition for Sustainable AI,^{17, 18, 19, 20} a multi-stakeholder global initiative launched in Paris in February 2025 by France, UNEP, and international partners.^{17, 18} The Coalition unites governments, technology leaders, and research institutions to align AI development with UN Sustainable Development Goals, focusing on climate action, pollution reduction, and biodiversity protection,^{18, 20} while addressing AI's dual challenge: leveraging its environmental potential while mitigating its own footprint in energy and water consumption. This engagement transcends corporate social responsibility; it is a strategic risk management and positioning maneuver. As AI systems grow more powerful and autonomous in essential services, they face heightened scrutiny around job displacement, algorithmic bias, data privacy, and security. By proactively shaping influential bodies like the Coalition, we position ourselves ahead of the regulatory curve, transforming from passive recipients of future regulations into active participants defining the rules for industrial AI, thereby building a critical social license to operate.

Powering Data Centers: Our Integrated Digital Solutions for Sustainable Operations

19As the Data Center industry serves as a fundamental pillar supporting the global expansion of AI capacity, Veolia strategically embeds digital capabilities to address the critical resource demands and reliability needs of this industry. Veolia’s comprehensive offer for data centers is structured around four pillars: Reliability, becoming Carbon Negative, being Water Positive, and achieving Waste Circularity. Supporting these goals, the company utilizes its Hubgrade platform to offer digital solutions for energy monitoring and optimization, which enhances energy efficiency and facilitates comprehensive sustainability program management. Furthermore, in pursuit of becoming Water Positive, Veolia incorporates advanced water treatment solutions that include real-time performance analytics to optimize cooling efficiency and reduce water consumption. By applying these data-driven tools, Veolia ensures that its clients not only maintain uninterrupted operations but also proactively manage their environmental footprint (including water usage, waste, and energy), which is paramount for ensuring a project's long-term territorial acceptability.

20This proactive stance helps us build what can be termed a social license to operate. It signals to regulators, clients, and the public that we are responsible stewards of this transformative technology. This builds a deep understanding of trust, a critical and intangible asset that can provide a significant competitive advantage over rivals who may be perceived as moving too quickly or without sufficient ethical guardrails. By embracing transparency and collaboration, we turn a potential liability ‒ the black box perception of AI into a source of strength, thought leadership, and enduring brand value.

Conclusion: Our AI-Powered Future and the Blueprint for a Greener Industry

21As I have outlined, our AI strategy is a holistic, multi-faceted, and deeply integrated corporate endeavor. It is not a collection of siloed pilot projects but a coherent ecosystem built upon the clear strategic vision of GreenUp, powered by our foundational platforms of Talk to My Plant and VeoliaSecureGPT, proven through a diverse and impactful portfolio of applications, and guided by a robust and responsible governance framework. We are architecting an intelligent enterprise where data and algorithms are as fundamental to our operations as pipes and treatment plants.

22I am confident that our journey offers a compelling and comprehensive blueprint for the future of the environmental services industry and for heavy industry at large. By successfully marrying ecological ambition with AI-driven operational and financial performance, we are not just optimizing our own business; we are demonstrating a viable, scalable model for how industrial leaders can spearhead the global ecological transformation. We are effectively betting our future on a powerful premise: that in the 21^st century, the most sustainable company will also be the most intelligent ‒ and ultimately, the most successful.