The accelerated buildout of AI data centers, along with growing scrutiny of their electricity, water and land use, is putting pressure on the world’s biggest technology companies to address the sustainability challenges this infrastructure creates. 

Amazon, Google, Meta and Microsoft, for instance, recently teamed up with investors to launch the Data Center Innovation Initiative, which will fund startups working on energy storage, advanced electrical systems, industrial cooling and low-carbon building materials for these large-scale facilities. 

Whatever its eventual real-world impact, the initiative signals something important: AI innovation and corporate sustainability commitments are becoming harder to discuss separately. That applies not only to the companies building the infrastructure but also to businesses using AI across their operations and growth plans. 

AI is not inherently at odds with sustainability. It can support the green transition by improving forecasting, optimizing logistics, managing energy demand, analyzing supply-chain risk and strengthening ESG data. Explaining how those benefits co-exist with the resources AI requires is a hard story to tell. Yet communicating that intersection clearly is just as important as promoting the technology’s potential. 

Companies will need to resist presenting AI as an unqualified sustainability solution or treating its environmental footprint as someone else’s problem. Their innovation and sustainability stories will only hold together if both are grounded in data, transparency and internal alignment.

How Far Does the Evidence Let You Go? 

Companies are already connecting specific uses of AI to sustainability benefits. H&M, for example, has described how AI-supported forecasting can help it better align production with actual demand, with potential benefits for inventory levels, raw material use, waste and emissions. 

That is a credible operational story to explore, but each part needs to be substantiated. Better forecasting may reduce overproduction, but by how much? Did waste fall as a direct result? Were emissions reduced, and over what period? How significant was the improvement within the company’s wider environmental footprint? 

Attribution also matters. A logistics company may use AI to identify shorter delivery routes, but the recommendation alone does not prove that emissions fell. The result also depends on whether drivers follow those routes, on changes in delivery volumes, on traffic conditions and on the types of vehicles being used. Similarly, AI may identify energy waste in a factory, but the environmental benefit only materializes if the business acts on the recommendation. 

Evidence, therefore, does two jobs: It establishes that an outcome occurred and sets the limits of what a company can responsibly say about it. An AI application may enable a specific reduction in waste or energy use, but that alone does not show that AI is making the business more sustainable overall. 

Keeping Public Claims Aligned With ESG Reporting 

Once an AI-related sustainability claim appears in a campaign, investor presentation or executive commentary, it does not exist in isolation. Stakeholders may compare it with the company’s sustainability targets, metrics and formal disclosures. For businesses subject to mandatory reporting requirements such as the EU’s Corporate Sustainability Reporting Directive, there is an even clearer basis for doing so. 

Even a well-supported claim can create confusion when teams describe the outcome in different ways. A retailer may have solid data that AI reduced excess inventory during a six-month pilot and communicate that result accurately. Its communications team, however, may describe the outcome as “waste reduction,” while the sustainability team uses that term specifically for materials sent to landfill across the company’s operations. Both accounts may be factually correct, yet the difference in language makes them appear inconsistent. 

Similar problems arise when teams use different reporting periods or organizational boundaries. A result measured in one business unit over several months may sit alongside annual company-wide data that shows little overall change. Without context, stakeholders are left to reconcile two versions of the story themselves. 

Communications teams, therefore, need access to the same terminology, proof and reporting parameters used elsewhere in the business. That requires coordination with sustainability, technology, operations, procurement and legal teams before the story reaches the public, so a valid result is described consistently across campaigns, investor communications and formal reporting. 

The Value of Admitting What You Don’t Know 

Few companies will have a complete picture of the environmental impact associated with every AI tool they use. Much of that footprint may reside with cloud providers, software vendors, data center operators and hardware suppliers. Energy and water data may be aggregated, estimates may vary, and the impact of a specific application may be difficult to isolate. 

Limited visibility does not make those impacts irrelevant. Companies should resist discussing only the benefits of AI adoption while treating the infrastructure behind it as somebody else’s responsibility. 

A financial services company using several third-party AI platforms, for example, may be unable to calculate the precise electricity or water use associated with those platforms. It can still explain which tools are being deployed, how widely they are being used, what information it has requested from suppliers and how environmental performance may influence future procurement decisions. 

Transparency means separating what the company knows from what it assumes. Admitting uncertainty can feel risky, but it is often more credible than presenting a level of confidence the available facts do not justify. 

As AI becomes more deeply embedded in corporate strategy, companies will face growing pressure to explain both what the technology delivers and what its expansion entails. A polished narrative will not be enough. The innovation and sustainability stories need to draw from the same evidence, or companies risk making claims that sound convincing in isolation but fail to hold together when the full picture comes to light.

Vanessa Horwell

Vanessa Horwell is the Chief Strategy Officer of THINKINK, a specialist communications firm working across sustainability-focused sectors including aviation, hospitality, retail and travel. For over two decades, she has advised companies on brand positioning, messaging, media strategy and thought leadership, helping complex businesses strengthen visibility, build credibility, and connect communications to commercial outcomes.