For years, publishers have suspected that AI crawlers were taking far more than they were giving back. This week the suspicion turned into a number, and the number turned into a fight.

On July 1, Cloudflare opened a new dashboard to its Bot Management customers that finally puts a figure on the imbalance between AI crawl traffic and the human readers that traffic is supposed to eventually send back. The ratios documented range from 118 crawls per referral at the low end to nearly 50,000 at the high end, a spread wide enough that two AI companies operating in the same market could be treating publishers in completely different ways without anyone being able to prove it before now, as PPC Land's analysis of the launch lays out. The dashboard sorts crawlers into Training, Search, or Agent categories and lets publishers compare operators side by side, filtering by 24-hour, 7-day, and 30-day windows.

Then, within days, the company whose crawler behavior the dashboard is best positioned to expose pushed back in the most dismissive way available. Google's John Mueller was asked on Reddit about Content Signals, a robots.txt directive Cloudflare introduced last year to let publishers flag how their content may be used. His answer, reported by Search Engine Roundtable on July 6, left no room for ambiguity. "AFAIK none of the crawlers / llms use the content-signal robots.txt directives," Mueller wrote. "It was made up by a CDN, afaik it has no effects whatsoever for any crawler or llm." Adopting the directive, he added, simply adds bloat and ongoing maintenance burden to a robots.txt file, since crawlers follow only the directives they are built to support and quietly disregard the rest. He added that Google does not use llms.txt either, and that he was not aware of any other crawler or large language model that does.

The dismissal matters because of scale. Cloudflare's infrastructure sits behind roughly 21.3 percent of all websites as of January 2026, so a standard it built is not a fringe experiment. It is already embedded across a fifth of the internet, and Google's own crawlers are simply choosing to ignore it. That leaves publishers in an odd position: a site can adopt the signal Cloudflare offers, but the largest AI operator in the world has just confirmed, in public, that doing so changes nothing. The practical effect is that a robots.txt update many publishers spent time implementing over the past year now functions as a symbolic gesture rather than an enforceable rule, at least until crawler behavior catches up with the standard, if it ever does.

Time magazine chose a different response entirely. Rather than trying to negotiate signal by signal with crawlers that may or may not respect them, Time has started converting its entire site from HTML into markdown, a stripped-down format that removes layout, navigation, and styling so that approved AI bots receive only the content itself, as Digiday reported on July 6 in its account of how publishers are rebuilding parts of the web specifically for AI agents. Time blocks all AI bots by default and whitelists the operators it wants to work with, routing them through a marketplace called TollBit that claims the markdown version can be fetched in a quarter of a second, against more than a minute for a full HTML page. It is a quietly significant admission: rather than relying on a shared standard that crawlers might ignore, at least one major publisher is now building a parallel, machine-only version of its own site. The logic is straightforward once stated plainly. If a voluntary robots.txt signal carries no enforcement power, and if the largest search operator has just said so out loud, then the only remaining lever a publisher controls directly is what content it serves and to whom, which is precisely what a whitelist-and-reformat strategy accomplishes without waiting for any crawler operator to change its behavior first.

Where the technical fight leaves ambiguity, a German court has just removed some of it. On July 4, PPC Land published a detailed account of a Munich Regional Court ruling that strips away a liability shield Google's search results had enjoyed under German law for roughly two decades. The court found that Google's AI Overviews had fabricated claims linking two publishing companies to fraud, and it rejected the argument that an AI-generated summary deserves the same legal protection as a plain list of search links. Three findings drove the decision: the overviews spoke in Google's own synthesized voice rather than simply displaying what sources said, the fabricated claims could not be traced to any actual source Google cited, and blocking the injured parties from suing Google over content Google itself generated would leave them with no path to a remedy at all, since the sites the AI drew from never made the disputed statements. Fines for continued violations can reach 250,000 euros.

The commercial and legal threads meet in the click numbers. A randomized study covered by PPC Land found that AI Overviews cut outbound clicks to publisher sites by 39.8 percent, while zero-click searches rose 34.5 percent, with no measurable improvement in user satisfaction to show for it. In other words, the same synthesis that a German court just ruled makes Google legally responsible for its own words is the same synthesis quietly removing the traffic publishers depend on, and PPC Land's reporting on both threads makes clear they are not separate stories so much as two views of the same shift.

Cannes surfaces a gap nobody wants to admit to

While the crawler fight was unfolding in dashboards and court filings, a different but related argument was playing out on the ground at Cannes Lions, and it goes to the heart of who actually controls how fast AI changes marketing.

Agency executives told Digiday, in a July 6 report on conversations held throughout the festival, that the gap between what agencies can now do with AI and what their clients are ready to use is not closing. It is widening. Joe Maglio, chief executive of Cheil Agency Network, put the headline version of the argument plainly: "The macro AI takeaway: agencies are further along than brands; agencies are the early adopters." Several other agency leaders would only speak on condition of anonymity, which is itself revealing, since acknowledging how far behind clients are can read as an implicit criticism of the people paying the bills.

One holding company chief executive, speaking to Digiday without attribution, said their agency's AI platform had grown revenue by triple digits for five consecutive quarters, and that the gap with what clients are prepared for keeps widening rather than closing. By this executive's count, chief marketing officers split roughly into four groups: a quarter actively building with AI, a quarter that wants to but does not know how, a quarter frozen entirely, and a quarter still dismissing the technology as a passing trend. Another agency leader, at a separate large independent shop, heard a similar story from marketers throughout the week, minus the confidence: "Most of them are not ready. Most of them feel like they're behind everybody else." A third executive framed the problem differently again, arguing that clients are asking the wrong question in the first place, fixating on "how many agents does your company have" rather than what any of it is meant to change.

Yet the marketers on the other side of that conversation describe a different set of constraints, not a different set of intentions, and Digiday's reporting gives real weight to both sides rather than treating the agency complaint as the whole story. Data governance, procurement, and legal sign-off move at a pace no agency roadmap accounts for, with one client's AI approval routed through an entire cybersecurity review before a single self-serve tool got greenlit. Budgets for AI experimentation are often squeezed inside marketing spend that is already flat, chasing a return that may not show up for three to six months, if it shows up at all. Jess Dervyn, an analyst at Gartner covering marketing, told Digiday that clients increasingly expect transparency about exactly where and how agencies are using AI on their behalf, rather than taking adoption on faith.

That tension between capability and caution echoed elsewhere at the same festival. Dara Treseder, chief marketing and commercial officer at Autodesk, told Adweek's Marketing Vanguard series, recorded live at Cannes and published July 2, that knowing when not to use AI is just as important as knowing when to reach for it. Her warning was specific: brands that deploy AI indiscriminately across social media and customer communication risk becoming indistinguishable noise, or worse, contributors to what she called AI slop, and trust broken through perceived inauthenticity takes exponentially more effort to rebuild than it took to lose. It is a notably different emphasis from the agency-side complaint that clients are simply too slow. Treseder's version suggests some of that caution is deliberate, not just organizational drag.

A third data point, from Omnicom, suggests the industry's own measurement thinking is shifting alongside the adoption debate rather than waiting for it to resolve. A new Omnicom Media study titled "Connected Content: The Force Multiplier for Maximizing Brand Influence," reported by MediaPost on July 6, argues that the traditional model of building one creative asset and adapting it across channels no longer reflects how consumers actually encounter brands, who now move across multiple platforms, devices, formats, and AI-powered experiences within a single day. The study calls for aligning creative messaging, media context, and audience mindset as one coordinated system rather than a series of separate channel plans. Read alongside the Cannes reporting, the pattern is consistent: the industry's most sophisticated players are moving toward AI-native measurement and creative frameworks, while a meaningful share of the client base is still working out procurement.

A tournament with two economies running side by side

The 2026 FIFA World Cup, the largest edition of the tournament in history at 48 teams, is generating two distinct advertising economies at once, and both were on full display this week.

On June 30, TAG, the nonprofit the advertising industry created to fight ad-related crime, shared a list of 1,376 pirate domains streaming or hosting stolen World Cup content with advertising supply chain intermediaries, PPC Land reported on July 5. The goal is not to take the sites offline. It is to cut off the advertising revenue that funds them, by getting demand-side platforms, agencies, and exchanges to automatically exclude the flagged domains from programmatic buying. TAG identified a further 176 domains already on its exclusion list as also trafficking in stolen World Cup content, bringing the running total past 1,500 sites addressed at some point during the tournament, which runs through the July 19 final at MetLife Stadium. Rachel Nyswander Thomas, TAG's chief operating officer, described major global sporting events as prime targets for criminals who redirect legitimate advertising dollars into pirated streams.

TAG's list works alongside a separate and more forceful action. The US Department of Justice announced Operation Offsides on June 26, seizing roughly 400 pirate domains outright through the National Intellectual Property Rights Coordination Center, with law enforcement partners across the United States, Peru, Bulgaria, Croatia, Romania, Poland, and Colombia, per PPC Land's reporting. The two efforts are complementary rather than identical: one removes domains from the internet entirely, the other starves the money behind them, and together they represent two different levers being pulled against the same underlying piracy economy within the same short window.

The legitimate side of that economy, meanwhile, is setting records. The United States men's national team's 2-0 win over Bosnia and Herzegovina on July 3 drew a preliminary 24.4 million viewers on the Fox Television Network, making it the most-watched English-language broadcast of a soccer match in American television history, ahead of the 22.3 million viewers the 2015 Women's World Cup final drew, also on Fox. Combined with Spanish-language coverage on Telemundo and Peacock, which added 9.1 million more viewers, the total combined audience of 35.3 million topped the 26.7 million combined total for that 2015 final. Fox has averaged 18.9 million viewers across the four matches the US team had played through that point in the tournament, MediaPost's figures show, and the trend line suggests the round of 16 and quarterfinal matches still to come could push those numbers higher still.

Brands are chasing that audience through a channel that looks nothing like a traditional media buy. Rather than simply purchasing airtime, brands are building entire creator-led activation programs spanning every host city, Digiday reported on July 2. Scott Sutton, chief executive of influencer marketing platform Later, told Digiday that a major in-person activation now runs in the eight to nine figure range of budget, and offered a description of what that spending is actually for: "Experiential is really just a vehicle to create published content on social." Charlene Patten, chief marketing officer at beauty brand Not Your Mother's, which built a themed pop-up in Miami and brought creators to a match in the same city, was direct about what she was and was not optimizing for: "It's a conversation play, it's a connection play. It's not about hardcore ROI at all." TikTok sent 30 creator correspondents to games this cycle, and Rollo Goldstaub, the platform's global head of sport, told Digiday that FIFA's brief was to grow awareness, drive tune-in, and expand the global fanbase, with success measured in cultural resonance rather than raw view counts alone. US-based searches for the term World Cup have risen more than 320 percent on TikTok since kickoff.

That creator spending sits in an odd relationship to the broadcasts themselves. AdExchanger's July 6 roundup noted that Fox Sports pays two creators fifty thousand dollars each to watch every match from a transparent cube in Times Square, yet the campaign has not appeared inside Fox's own broadcasts, and the much larger battalion of credentialed influencers working inside stadiums rarely surfaces in live programming either. Creator marketing has, within a single tournament, earned its own substantial budget line. A place inside the broadcast it orbits remains, for now, unearned, and that gap between the money creators command and the airtime they are given is likely to be one of the more interesting open questions once the tournament ends and the industry starts writing its post-mortems.

Two leadership changes point in the same direction

Two executive moves within the coverage window, both touching how companies plan to run AI and data together, are worth reading side by side.

Meta named Denise Moreno as chief marketing officer while Alex Schultz stepped into a newly created role as the company's first chief data officer, both reporting to chief operating officer Javier Olivan, Adweek reported on July 1. Moreno is a seventeen-year veteran of the company, and Schultz has held the top marketing job since 2020; between them they carry more than 35 years at Meta. Moreno's own framing of the move, posted to LinkedIn and cited in AdExchanger's July 6 roundup, captures a tension that runs through nearly every conversation happening in the industry right now: "The teams that win won't be the ones that hand everything to the machine. They'll be the ones that pair AI's speed and scale with human judgment and taste." Schultz, in his own announcement, described data, research, and experimentation as remaining among the most important strategic capabilities any company can hold, while adding that they need to be transformed for the current moment.

Paramount Skydance made a parallel move on the corporate side rather than the marketing side. The company is consolidating its ad tech and product teams under Hugh Williams, a newly hired executive vice president and former Google leader, per a memo reported by Business Insider. Five divisions emerge from the reorganization: product management under a team of four executives, engineering under new hire Rich Orne, advertising solutions under longtime Paramount executive Dayna Wasilefski, client relations under new field chief technology officer Travis Scoles, and a data division that does not yet have a named leader. Chief executive David Ellison has said the goal is closing a gap with competitors including Netflix and NBCUniversal on ad tech capability, and the timing carries an added layer of insurance: an ongoing United Kingdom regulatory challenge to Paramount's acquisition of Warner Bros. Discovery could delay access to that company's NEO ad platform, and a stronger in-house team would limit the damage if that access is delayed. The reorganization also follows Paramount's earlier work unifying the separate ad tech stacks behind Paramount+ and Pluto TV, a project executives have discussed publicly since the 2025 merger with Skydance, so this is less a sudden pivot than the next stage of a rebuild already underway.

Read together, the Meta and Paramount moves describe the same underlying pattern from two different angles. One company is elevating a dedicated data leadership role specifically to sit alongside marketing, rather than beneath it. The other is consolidating scattered ad tech functions under a single new hire pulled directly from the company most associated with programmatic infrastructure in the first place. Both are bets that the organizations best positioned for what comes next are the ones that stop treating data, AI, and advertising as separate departments.

Also noted