AI’s Environmental Footprint Moves Onto Consumer Goods Climate Agenda

Artificial intelligence is becoming a new climate accounting issue for consumer goods companies, as retailers and manufacturers increase their use of generative AI, cloud platforms, automated forecasting, marketing tools, and data-driven supply chain systems.

A new report from The Consumer Goods Forum, titled The AI Footprint: What Consumer Goods Companies Need to Know About AI’s Environmental Impacts, finds that generative AI and cloud services currently account for less than 1% of a typical consumer goods or retail company’s Scope 3 emissions. However, the report warns that the sector should not treat the impact as marginal. It estimates that the AI market could grow 25-fold by 2033, while end-user Scope 3 emissions linked to AI could rise by six to eight times.

Why Consumer Goods Firms Should Pay Attention

For consumer goods companies, the issue is especially relevant because most emissions are already concentrated outside direct operations. Purchased goods and services, packaging, logistics, product use, and waste usually dominate the footprint.

AI-related emissions may be small today, but they sit within the same Scope 3 category that many companies already struggle to measure with precision. As AI tools become embedded in business functions, digital infrastructure could become a more visible part of corporate climate reporting.

Where AI Emissions Come From

The environmental footprint of AI is linked mainly to the data centre lifecycle. According to the report, the use phase accounts for around 70% of total lifecycle emissions for a representative data centre, although the figure varies by location, electricity mix, server refresh cycles, and operating model.

Key pressure points include electricity demand, water use for cooling, refrigerants, diesel backup power, embodied carbon in steel and concrete, semiconductor manufacturing, and growing electronic waste from rapid hardware replacement cycles.

Data Centres Face Growing Scrutiny

The findings come as the wider energy implications of AI are receiving greater scrutiny. Rapid growth in cloud computing and generative AI is increasing demand for data centre capacity, electricity, cooling infrastructure, chips, and grid connections.

For consumer goods firms, this creates a governance challenge. AI is often purchased indirectly through cloud providers, software platforms, marketing systems, enterprise tools, or logistics technology. That means emissions may be bundled inside broad IT or cloud procurement categories, rather than tracked as a specific source of climate impact.

Scope 3 Targets Could Be Affected

This matters for companies with science-based targets, supplier engagement programmes, and public net zero commitments. If AI usage grows across customer analytics, demand forecasting, product design, warehouse operations, e-commerce, pricing, and marketing content generation, digital emissions could become more material.

The report notes that spend-based emissions factors may undercount the compute intensity of AI, while many corporate climate roadmaps were developed before the current AI adoption surge. Companies may therefore need to revisit how digital services are represented in their emissions inventories.

AI Can Also Support Sustainability

The report does not argue that companies should avoid AI. It presents AI as both a source of environmental pressure and a potential sustainability tool.

AI can support better forecasting, reduce waste, improve transport planning, optimise energy use, and strengthen supply chain resilience. It can also help companies analyse supplier data, detect inefficiencies, model climate risks, and improve decision-making across complex value chains.

The practical challenge is to ensure that the environmental benefits of AI outweigh the infrastructure impacts associated with deploying it at scale.

Better Measurement is the First Step

The Consumer Goods Forum recommends that companies begin by collecting better internal data on AI use. This may include AI tokens, queries, storage, cloud region, compute location, and the type of workload being used.

This would allow sustainability, procurement, IT, and finance teams to understand whether emissions are coming from high-volume everyday usage, heavier agentic workloads, model training, data storage, or specific vendors.

Better measurement would also help companies compare providers, identify lower-impact deployment options, and avoid treating all cloud or AI services as environmentally equivalent.

Cloud Providers Will Play a Central Role

The report also recommends stronger engagement with hyperscalers and cloud providers. Consumer goods companies should ask for clearer emissions data, including electricity use, renewable energy sourcing, embodied carbon, data centre location, water metrics, and how leased or co-located data centre capacity is treated.

This is likely to become increasingly important as AI procurement moves from experimental pilots to core business infrastructure. Large retailers and manufacturers may be able to use procurement requirements to push technology providers toward greater transparency and lower-carbon infrastructure.

Climate Governance Needs to Include AI

Another recommendation is to update climate governance. Boards, sustainability teams, digital teams, procurement departments, and finance functions may need a shared framework for deciding when AI use is justified, how emissions are measured, and how vendor choices are assessed.

Internal carbon pricing could also be applied to digital and AI services where companies already use carbon prices for capital allocation or supplier decisions.

This would help companies integrate AI into wider climate transition planning rather than treating it as a separate technology issue.

Mitigation Options Are Already Emerging

The report highlights several mitigation routes, including clean energy procurement for data centres, more efficient chips and models, lower-carbon construction materials, immersion or zero-water cooling systems, alternatives to diesel backup generators, and hardware reuse or recovery.

Many of these measures are controlled by technology providers and data centre operators. However, large consumer goods companies can influence adoption through supplier engagement, procurement standards, contract requirements, and climate data requests.

A Growing Issue for Net-Zero Strategies

For retailers and manufacturers, the message is that AI governance is becoming part of climate governance. Companies do not need perfect data before acting, but they do need a clearer view of where AI is used, who provides the infrastructure, how emissions are allocated, and whether digital growth is reflected in decarbonization plans.

AI-related emissions may remain a low single-digit share of total Scope 3 emissions for many consumer goods companies in the near term. But rapid growth, weak measurement systems, and rising data centre resource demand mean the issue is likely to become more material over time.

For companies already under pressure to improve Scope 3 transparency, AI is becoming another test of whether climate targets can keep pace with changing business models.

Source: sustainabilityonline.net