Marketing’s AI Spending Boom Meets an Accountability Gap in 2026

Marketing has become one of the most AI-saturated functions in the enterprise, from ad targeting to content generation to campaign optimization. But three data points released within days of each other in July 2026 point to the same uncomfortable conclusion: spending on AI is accelerating faster than the industry’s ability to measure what it is buying. Meta’s ad business is booming on the back of AI-driven improvements, Google’s Gemini integration lifted search and ad revenue 17% year over year, and yet the underlying research on both financial return and environmental cost tells a much murkier story — one that echoes how AI Overviews are already reshaping marketing playbooks.

How Much Are Marketers Actually Spending on AI in 2026?

Marketing-adjacent AI spending is now measured in the hundreds of billions of dollars, led by platform owners rather than individual brands. Meta has committed roughly $600 billion to US AI infrastructure and has invested hundreds of billions overall in AI model development and data centers, spending that flows directly into the advertising products marketers depend on.

That spending is showing up in platform results. Google reported search and other ad revenue up 17% year over year, a jump the company attributed partly to Gemini-driven ad improvements and World Cup-related demand, giving marketers a visible signal that AI investment on the platform side is translating into a stronger, if more expensive, advertising product.

Is AI Spending Translating Into Measurable Marketing Returns?

The evidence on return is mixed at best. Meta’s total 2025 revenue was $200.97 billion, with only $4.8 billion coming from sources outside advertising, and analysts estimate the company would need more than a decade to recover its AI infrastructure costs even assuming $100 billion in annual AI subscription revenue — a figure it does not currently generate.

That gap between investment and realized value is not unique to Meta, and it mirrors the same measurement uncertainty marketers face when adapting to answer engine optimization and AI-driven search. A National Bureau of Economic Research study of nearly 6,000 executives found the vast majority reported minimal operational impact from AI tools, directly contradicting the productivity narrative driving much of the current spending cycle. For marketing leaders building next year’s budget, that disconnect is the single most important data point to reconcile before committing further AI spend.

Why Aren’t Marketers Measuring AI’s Environmental Footprint?

Most marketing leaders believe AI is raising their carbon emissions but have not built the systems to confirm it. A June 2026 survey of 200 senior US and UK marketing leaders by 51toCarbonZero found 88% believe AI is increasing their organization’s emissions, with 42% perceiving a significant increase, yet only 36% have comprehensively measured that impact and 8% have not measured it at all.

The same survey found 88% of respondents also report AI is pushing up operational costs, with 35% describing the increase as substantial. As one executive quoted in the research put it, “businesses cannot effectively reduce what they are not measuring” — a warning that applies equally to carbon output and to marketing ROI.

How Do US and UK Marketers Differ on AI’s Environmental Impact?

US marketing leaders report significantly higher concern about AI-driven emissions than their UK counterparts, a gap likely tied to differences in AI adoption intensity and energy grid composition. In the 51toCarbonZero survey, 51% of US marketers perceived a significant increase in AI-related emissions compared with 32% of UK marketers.

Despite that gap, overall sustainability pressure appears to be easing on marketing budgets: only 17% of respondents cited sustainability budget concerns, down sharply from 37% in 2025, and fewer than a quarter see internal alignment as a major sustainability challenge. That suggests AI’s environmental cost is being noticed but not yet prioritized against other budget demands.

What Does the Google-Meta Divergence Reveal About AI ROI?

Google and Meta illustrate two different AI monetization paths that marketers should evaluate separately rather than treating “AI advertising” as one category. Google is converting AI investment into incremental improvements on an already-profitable ad auction, which shows up quickly in reported revenue growth. Meta is simultaneously running a much larger, multi-year infrastructure bet whose payoff — if it arrives — is structurally further away and dependent on new revenue lines outside advertising that do not yet exist at scale.

For a marketing leader deciding where to shift budget, that distinction matters: near-term AI-driven platform improvements can be evaluated with standard campaign metrics, while platform-level AI infrastructure bets are closer to a venture wager that individual advertisers have little ability to influence or audit.

How Should Marketing Leaders Build an AI Measurement Framework?

A credible AI measurement framework needs to track cost, output quality, and environmental footprint together, rather than reporting AI adoption as a standalone success metric — the same discipline that underpins a solid content audit process.

• Separate platform-level AI gains from your own AI tooling spend — improvements in Google or Meta’s ad delivery are not the same line item as your team’s AI content or analytics tools, and each needs its own ROI case.
• Require a baseline-vs-outcome comparison before scaling any AI tool — the NBER finding of minimal operational impact across nearly 6,000 executives suggests many deployments are not being tested against a real baseline.
• Start tracking AI-related energy and compute costs now — with 64% of surveyed marketers still lacking comprehensive measurement, building this capability early is a genuine competitive and reporting advantage, not just a compliance exercise.
• Treat vendor AI roadmap promises as unverified until proven — Meta’s metaverse-to-AI pivot after an $80 billion writedown is a reminder that platform strategy can change faster than a marketing budget cycle.

What Should CMOs Ask Before the Next AI Platform Renewal?

Every major ad platform renewal in the second half of 2026 is effectively a referendum on whether that platform’s AI investment is paying off for advertisers specifically, not just for the platform’s own reported earnings growth. Google’s 17% ad revenue jump is a company-level number; it does not automatically mean an individual advertiser’s AI-assisted campaigns outperformed prior-generation targeting by a comparable margin.

CMOs renewing large platform commitments should request cohort-level performance data that isolates AI-attributed gains from seasonal demand, such as the World Cup advertising surge that also lifted Google’s numbers this quarter. Without that isolation, budget increases risk being justified by macro tailwinds mislabeled as AI performance.

The same discipline applies internally. If a marketing team has adopted AI copywriting, image generation, or campaign optimization tools, the underlying cost — compute, subscription fees, and the emissions each carries — should appear in the same budget review as the output it produces. The 51toCarbonZero finding that 64% of marketing leaders have not comprehensively measured AI’s environmental impact suggests most budget reviews are currently missing half of the picture entirely.

Marketing leaders who build this measurement discipline now will be better positioned when regulators or major clients start asking for AI-related emissions disclosures as a standard part of sustainability reporting, a trend already visible in the sharper focus on operational cost tracking within this year’s survey data.

Frequently Asked Questions

Is AI actually improving marketing ROI in 2026?

Results are mixed: platform-level improvements like Google’s Gemini-driven ad gains show measurable revenue impact, but a broad NBER survey of nearly 6,000 executives found most report minimal operational impact from AI tools overall.

Why do marketers believe AI is increasing costs and emissions but not measure it?

Most marketing organizations have not built the monitoring infrastructure to track AI-specific compute costs or emissions separately from general IT spend, so belief is based on general perception rather than internal data, per the 51toCarbonZero survey.

How much has Meta invested in AI infrastructure?

Meta has committed roughly $600 billion to US AI infrastructure as part of a broader multi-hundred-billion-dollar AI investment program, funded primarily by its advertising business, which generated $200.97 billion of its $205.77 billion in total 2025 revenue.

What is the biggest risk of unmeasured AI spending in marketing?

The biggest risk is budget commitment without an audit trail — without cost, output, and emissions measurement, marketing leaders cannot distinguish genuinely effective AI tools from expensive experiments, making next year’s budget negotiations far harder to defend.

Son Güncelleme / Last Updated: July 27, 2026. Sources: 51toCarbonZero AI sustainability survey (June 2026); National Bureau of Economic Research executive AI impact study; Meta and Alphabet Q2 2026 earnings reporting; Marketing Dive reporting.