The Problem Nobody in Marketing Is Talking About Yet

Every time an AI model processes a bidding decision, generates a creative variant, or optimises a media schedule, it consumes energy. A lot of it. And as AI tools become embedded across the full programmatic advertising stack, from targeting and creative to bidding and attribution, that energy cost is quietly accumulating inside marketing's carbon footprint.

The scale of the underlying issue is striking. The International Energy Agency projects that global data centre electricity demand will more than double between 2022 and 2026, driven substantially by AI workloads (IEA, 2025). A Goldman Sachs Research analysis from 2025 estimated that around 60% of the additional electricity demands from AI data centres will be met by burning fossil fuels, adding approximately 220 million tonnes of carbon emissions globally (Goldman Sachs Research, 2025). Meanwhile, one analysis suggested the AI carbon footprint alone could reach between 32.6 and 79.7 million tonnes of CO2 by 2025, equivalent in scale to New York City's entire annual emissions (ScienceDirect, 2025).

For marketing teams, the direct implication is this: the more AI is used across campaign planning, programmatic buying and creative production, the more energy is consumed, and the more carbon is attributed to marketing's Scope 3 footprint.

‍

Why This Has Become Urgent in 2026

Until recently, AI's role in advertising carbon accounting was largely theoretical. It was a future concern. That has changed, for two reasons that are happening simultaneously.

The first is regulatory. Under the EU's Corporate Sustainability Reporting Directive, large companies are now reporting their fiscal year 2025 emissions in 2026, with Scope 3 disclosures mandatory where material (European Commission, 2025). Digital advertising sits squarely within Scope 3, Category 1 (Purchased Goods and Services) under the Greenhouse Gas Protocol framework. For any brand or agency spending significantly on programmatic advertising, AI-enabled or otherwise, that spend generates emissions that now need to be disclosed, measured and verified. California's Climate Accountability Package similarly requires Scope 3 reporting from companies with over $1 billion in revenue operating in the state from 2027 onwards (Aligned Incentives, 2024).

The second driver is scrutiny. A February 2026 report backed by Beyond Fossil Fuels found that 74% of industry claims about AI's climate benefits were unproven and could not identify a single case where consumer generative AI systems were delivering material, verifiable and substantial emissions cuts (Greenpeace International, 2026). As investors, procurement teams and regulators become more sophisticated, vague commitments to sustainable AI are no longer credible. Marketing teams need data, not intentions.

‍

What the Advertising Technology Ecosystem Contributes

It is worth being precise about where emissions originate in AI-driven advertising, because the source matters for how you measure and reduce them.

Research from the ad industry estimates that the advertising technology ecosystem as a whole is responsible for roughly 4% of global greenhouse gas emissions, a figure that puts it on a par with the aviation industry (Futureweek, 2025). More than half of advertising's carbon output comes directly from the programmatic supply path: real-time bidding, demand-side platforms, supply-side platforms, data management and the server infrastructure that underpins them all (DMEXCO, 2025).

AI workloads are now a growing proportion of that infrastructure demand. As AI models run bidding algorithms, process audience data at scale and power creative optimisation tools, they add to the computational load sitting behind every campaign. Supply path complexity compounds this. Every unnecessary intermediary in a programmatic chain adds a data transfer, a server call and an energy cost. The industry has understood this problem in the context of waste and efficiency for some time, but the AI acceleration of the last two years has given it new urgency.

The good news is that the relationship between AI efficiency and carbon efficiency is not inherently adversarial. When AI is used well, to eliminate redundant supply paths, remove low-quality inventory and reduce bid duplication, it can reduce campaign carbon footprints alongside improving performance. Research suggests that green media buying and supply chain optimisation can cut carbon footprints by around 12% without compromising campaign outcomes (Accio, 2025). The challenge is that most marketing teams are not yet measuring the baseline, so they have no way of knowing whether the AI tools they are deploying are contributing to the problem or helping to solve it.

‍

The Measurement Gap Is the Core Problem

Here is the practical difficulty: most marketing teams have limited visibility into the carbon impact of their AI-enabled campaigns. Spend-based carbon estimates, which calculate emissions by applying an average emissions factor to media spend, are widely used as a proxy but are highly inaccurate. Research comparing spend-based methods to activity-based measurement found that spend-based methods overstate emissions by up to 451% in some cases (Carbon Intelligence, 2026). That level of inaccuracy is not suitable for regulatory reporting.

The Global Media Sustainability Framework (GMSF v1.2), published by Ad Net Zero in 2025 and now the recognised industry standard, provides activity-based formulae for digital advertising that account for data centre energy, network transfer, device rendering and third-party tracking energy (Ad Net Zero, 2025). These methodologies are aligned with the Greenhouse Gas Protocol and with CSRD requirements, meaning data calculated under the GMSF is suitable for inclusion in verified sustainability disclosures.

For AI-driven campaign activity specifically, the application of GMSF methodology means measuring the actual computational activity behind campaigns, not just the media spend that proxies for it. This requires connecting carbon measurement to live campaign data across all channels in use. EcoMetrics integrates directly with the platforms where this activity happens, including Google Analytics, Meta, YouTube, TikTok, LinkedIn and email marketing tools, pulling campaign-level data and applying GMSF-aligned methodology to produce accurate emissions figures that can feed into Scope 3 reporting. The platform was developed in partnership with Sheffield Hallam University, giving the underlying methodology academic rigour alongside commercial usability.

‍

What Marketing Teams Should Do Now

The window for getting ahead of this is narrowing. Here is a practical sequence for marketing teams and agencies navigating the AI carbon question in 2026.

Establish a campaign-level carbon baseline. The starting point is measurement. Without knowing the current carbon footprint of your digital advertising activity, it is impossible to identify where AI tools are adding to emissions and where they are helping to reduce them. Activity-based measurement using GMSF-aligned methodology gives you the data that is both accurate enough for regulatory reporting and granular enough to drive optimisation decisions.

Audit your programmatic supply path. AI-driven programmatic advertising tends to generate higher emissions when supply chains are complex and opaque. Unnecessary intermediary layers, made-for-advertising inventory and poor targeting hygiene all generate energy without generating value. A supply path audit, informed by carbon data, typically reveals both efficiency opportunities and emission reduction opportunities simultaneously.

Align AI tool selection with sustainability criteria. As AI becomes embedded across the marketing stack, sustainability should be a factor in vendor selection alongside performance capabilities and cost. Questions worth asking include: how energy-efficient is this platform's infrastructure? Is it GMSF-aligned? Does it provide emissions data at a campaign level? Vendors that cannot answer these questions are unlikely to meet the data requirements that CSRD and equivalent frameworks will require.

Build carbon reporting into your marketing governance. Scope 3 advertising emissions are not a bolt-on to sustainability reporting. They sit within the same framework as supply chain emissions and purchased goods. Marketing leaders who engage with sustainability and finance teams now, and build carbon measurement into their reporting cadence, will be significantly better positioned when disclosure obligations mature.

The Opportunity Inside the Problem

It would be easy to frame AI's energy footprint as purely a compliance and risk management challenge. That framing misses the commercial opportunity.

The same data that enables CSRD compliance also enables performance improvement. Research consistently shows that the inventory with the highest carbon footprints tends to be the inventory with the lowest quality: made-for-advertising sites, bloated supply chains, poor viewability. Marketers who use carbon data to clean up their media plans tend to improve campaign performance at the same time. The sustainability case and the commercial case point in the same direction.

In 2026, the most sophisticated marketing teams are beginning to treat carbon as a performance metric, sitting alongside cost per acquisition, return on ad spend and viewability. Those who get there first will not only meet their reporting obligations more easily. They will be making better decisions with their media budgets, running leaner supply chains and building a sustainability narrative that is grounded in verifiable data rather than aspirational commitments.

Sebastian Munden, Chair of Ad Net Zero, described the current moment as the advertising industry's "business opportunity of a generation" (Ad Net Zero, 2025). AI's energy demands make that opportunity more urgent, not less. The brands and agencies that measure their AI carbon footprint digital advertising activity now, and act on what they find, are the ones who will define what responsible marketing looks like for the next decade.

‍

Conclusion: Measure First, Then Optimise

AI is not going to slow down. Its role across digital advertising will deepen, and with it, the energy demands embedded in every campaign will grow. Marketing teams that treat this as someone else's problem are accumulating a liability: in their Scope 3 disclosures, in their sustainability credentials and in the quality of their media investments.

The answer is not to use less AI. It is to measure what AI-driven advertising actually costs in carbon terms, using methodology that meets regulatory standards, and to use that data to make better decisions. That is what EcoMetrics is built to do. The measurement infrastructure exists. The regulatory framework is in place. The commercial case is clear.

The only question remaining is when your team starts measuring.

‍

‍