FIFA World Cup viewers encountered visible branding for products associated with health, financial, or social risks roughly every seven seconds, according to a study released by the University of Bristol. Researchers counted 93,409 appearances during 172.6 hours of live play across all 104 matches. The total covered branding embedded in the centrally distributed match footage—not commercial breaks inserted by individual broadcasters.
The study analyzed branding embedded in centrally distributed live World Cup match footage.
Unhealthy food and drink accounted for 65,722 appearances, or 70.4% of the total. The remaining categories were prediction markets with 9,360, alcohol with 9,266, gambling with 6,429, and cryptocurrency with 2,632. Four brands—the study does not name them in the supplied account—generated 65% of all harmful-commodity logos: a soft drink with 21,893 appearances, a fast-food chain with 13,915, a sports drink with 12,777, and a crisps brand with 12,087.
The practical concern is not simply how many logos appeared, but whether viewers could avoid them. The researchers found branding during 39.3 hours of live play, almost a quarter of match time, and in every match. Because the branding was part of the football footage rather than a separate break, skipping it would also mean missing play. That distinction matters when advertising rules are written around conventional commercial slots rather than logos on hoardings, clothing, or other elements within the broadcast.
The study also raises a cross-border compliance question. FIFA distributed the same centrally controlled footage worldwide, including to countries where particular alcohol, gambling, or prediction-market promotions may be restricted or unlicensed. Virtual board-replacement technology can show different pitchside advertising to different territories, and the researchers say it was not used to remove or replace the relevant branding. Their evidence supports the claim that audiences saw the same embedded logos across markets; it does not, by itself, establish that every display violated a local law, because the supplied material does not reproduce territory-by-territory legal determinations.
For the 53 broadcasts with audience data, the researchers estimated 267 billion advertising impressions in the United Kingdom. “Impression” here is an exposure estimate, not proof that a person looked at or processed a logo. The study counted each visible representation—including logos, standalone symbols, and written names—and recorded the brand, match, second, duration, and number of simultaneous copies. It did not measure attention, recall, purchasing, or subsequent health outcomes.
That limitation is central to interpreting the findings. Existing research may link marketing for unhealthy products, alcohol, and gambling with consumption and harms, particularly among children and young people, but this audit does not test that causal chain during the tournament. It establishes the volume, timing, and distribution of visible branding in the analyzed footage. It cannot show how individual viewers interpreted the exposure or whether a particular appearance changed behavior.
The research team trained a computer-vision model to detect sponsorship and branding in the world feed and ran the analysis on Isambard-AI at the University of Bristol. The model’s classification rules treated food and drink according to World Health Organization nutrition standards, while automatically classifying alcohol, gambling, cryptocurrency, and prediction-market brands as harmful. Those choices make the inventory reproducible in principle, but the category definition is broader than a legal definition of prohibited advertising. A logo’s presence therefore should not be read as a finding that its owner broke a rule.
The researchers propose extending marketing rules to commercial branding within live sports broadcasts, applying meaningful penalties, and requiring FIFA to use an independently developed public-interest standard for commercial partnerships. They also recommend using available technology to remove or replace branding where a product cannot lawfully be advertised or offered. Those are policy recommendations from the study’s authors and affiliated advocates, not conclusions independently established by the logo count.
The evidence is nevertheless useful for regulators, broadcasters, sponsors, and rights holders because it turns an intuitive complaint about saturated sports marketing into a measurable audit trail. It can help identify which brands appeared, when they appeared, and whether a broadcast feed carried the same commercial content across jurisdictions. It cannot, without additional audience, legal, and behavioral evidence, determine the effect on any viewer or the legality of each exposure.
The study was funded by the University of Bristol and conducted by researchers from Bristol and Oxford. Its strongest contribution is therefore a detailed description of branding embedded in the 2026 tournament’s live broadcast feed, together with a clear warning that global distribution can carry commercial messages into markets with different rules. The next evidentiary step is not to treat the count as a verdict on every sponsor, but to pair it with transparent detection methods, territory-specific legal review, and independent measures of audience exposure and harm.