Fact check: Carbon footprint of average email

1. Why estimates swing from tiny to substantial — methodological fault lines

The largest factor driving contradictory numbers is which parts of the email lifecycle are counted: purely transmission energy versus storage, viewing, attachments, and the upstream emissions of infrastructure. One analysis gives a per-email range of 0.03 g to 26 g CO2e by varying assumptions about server efficiency and user behaviour, illustrating the sensitivity of any average to methodological choices ^[1]. Other work separates spam from legitimate messages and treats attachments as multiplicative factors, producing a 0.3 g (spam) to 50 g (large attachments) spread that depends on file size and storage duration ^[2]. These diverging frames explain why headline numbers can diverge by two orders of magnitude and why averaging across studies without harmonizing assumptions is misleading ^{[2] [3]}.

2. Spam and filtering: a hidden lever with big claimed impacts

One strand of research focuses on spam’s share and the effect of filtering: spam messages are estimated at about 0.3 g CO2e each, and proponents argue that effective spam filtering saves large quantities of electricity—reported as 135 TWh per year—with claims equivalent to taking millions of cars off the road ^[3]. Those savings are large because spam volume is enormous, but the calculation rests on assumptions about the marginal energy of filtering and the counterfactual of what networks would otherwise process. The emphasis on spam frames policy responses toward better filtering and deletion, which can yield meaningful aggregate gains if the per-message estimates and assumed volumes are accurate ^[3].

3. Attachment-heavy email drives up per-message emissions

Studies that disaggregate email types show attachments dominate per-message footprints: a “standard” email may be estimated at ~4 g CO2e, whereas messages with long attachments can reach ~50 g CO2e, mainly from storage and repeated downloads ^[2]. That distinction elevates behavioral solutions—compressing files, using shared links, or limiting attachment forwarding—because the carbon cost scales with repeated transfers and long-term storage. Emphasis on attachments also shifts responsibility from core infrastructure to user practices and corporate storage policies, suggesting different mitigation levers than simply greening data centers ^[2].

4. Aggregating millions of emails: headline national and global impacts

When per-email estimates are scaled, they can produce large aggregate figures: one synthesis suggested global email emissions could amount to ~150 million tonnes CO2e in 2019, a headline-grabbing total that depends on assumed message counts, average sizes, and system efficiencies ^[1]. Aggregation converts small per-unit uncertainty into large systemic ambiguity: a small change in assumed grams per message multiplies into tens of millions of tonnes. The aggregate framing can justify systemic policy actions, but it can also obscure the sensitivity of the conclusion to methodological choices and data limitations ^[1].

5. Digital activity’s broader footprint — emails in the context of digital emissions

Wider studies of digital content suggest that streaming, conferencing and general web use may account for a large share of per-capita carbon budgets, indicating emails are one part of a broader digital footprint ^[4]. Other analyses point to the potential of digital infrastructure to enable environmental governance while also imposing resource demands, highlighting a dual role: digital tech both consumes resources and provides tools to lower emissions elsewhere ^{[5] [6]}. Placing email emissions in this context suggests mitigation should consider trade-offs across digital uses rather than focusing narrowly on messaging alone.

6. Where researchers diverge — assumptions to watch for

Key assumptions that produce divergent results include the energy intensity of data centers, regional electricity grid carbon intensity, the lifespan and reuse of devices, user viewing behavior, and whether storage is counted as fixed overhead or marginally attributable to each message ^{[1] [2] [3]}. Studies reporting very low per-email footprints typically treat email as marginal on robust, low-carbon infrastructure; higher estimates attribute storage and repeated downloads and use higher grid-intensity factors. Awareness of these assumptions is critical for interpreting any single number and for designing policy or personal behavior changes.

7. Policy implications and practical takeaways from conflicting findings

Given the range of plausible estimates, the most robust policy levers are those that reduce the largest, well-evidenced drivers: cutting storage redundancy, improving spam filtering efficacy, shifting data centers to low-carbon electricity, and educating users about attachments and retention. The evidence signals that aggregate impacts can be meaningful, but the effectiveness of specific interventions depends on which lifecycle components policymakers target ^{[3] [2] [6]}. Different stakeholders—email providers, businesses, and end users—face different responsibilities depending on whether the dominant emissions are operational or user-driven.

8. Final synthesis — no single number, clear choices

The assembled analyses demonstrate no consensus “average” email footprint; instead there is a range reflecting definitional choices and scaling assumptions, from fractions of a gram to tens of grams per message, and potentially large global totals when aggregated ^{[1] [2] [3]}. The practical conclusion is clear: focus on high-leverage interventions (attachments, storage, spam) and clean energy for infrastructure rather than chasing a single per-email gram figure. This approach aligns with multi-study evidence and helps translate uncertain per-unit estimates into actionable policy and behavior changes ^{[1] [4]}.