1 Cannes Lions entries fell 25% as the integrity reset bit
Cannes Lions 2026 drew 20,050 entries from 92 countries — down 25.46% from last year’s 26,900 — after the festival tightened its awards-integrity standards. Coverage of the week framed it as a serious reset: fewer entries, tougher scrutiny, and a widening gap between the work that wins trophies and the work brands actually run. The recurring verdict from the jury rooms was that the strongest entries answered older questions — did it solve a real problem, did it change behavior, did it respect the audience’s intelligence.
A quarter of the entries vanishing the same year the rules got stricter tells you how much of the “case study” economy was built for the trophy, not the business. For marketers the signal is genuinely useful: the festival is trying to re-anchor on work that moved a customer to do something, not work that spiked attention for a week. If your 2026 creative plan is shaped around what plays well on the Croisette, run it through the same test the juries did — would it survive if the only question were “did it change what a customer did?”
2 Claude’s anti-AI ads won the Film Grand Prix
The Film Grand Prix — Cannes’ oldest category — went to Anthropic’s Claude for two Super Bowl LX spots, “Can I Get a Six Pack Quickly?” and “How Can I Communicate Better With My Mom?”, created by Mother London and directed by Jeff Low through Biscuit Filmworks. The campaign used humor to needle AI advertising and argue there’s “a time and a place,” also taking Gold and Silver in Film and a Silver in Creative Strategy — and it pulled a public jab from OpenAI’s Sam Altman, who called Anthropic “deceptive” on X.
The most-awarded film of the year was an AI company arguing against AI-made ads, and it worked precisely because it was funny, human, and had a point of view — the exact things a model can’t reverse-engineer from a brief. That’s the lesson hiding inside the trophy. As AI floods every feed with competent, on-brief, forgettable creative, the scarce asset isn’t production volume; it’s a take. The brands winning attention right now aren’t out-producing the machines — they’re saying something the machines never would.
3 Publicis agreed to buy LiveRamp for $2.2B — and rivals attacked the “neutrality”
Publicis Groupe agreed to acquire the data-collaboration platform LiveRamp for roughly $2.2 billion, with the deal expected to close before year-end 2026 and LiveRamp CEO Scott Howe reporting to Arthur Sadoun. Publicis pledged that LiveRamp would stay neutral and interoperable, with open ecosystem access and no use of client data beyond existing contracts — but rivals Omnicom and Stagwell publicly questioned whether an identity backbone owned by one holding company can credibly stay neutral.
LiveRamp’s identity graph sits underneath huge swaths of ad targeting and measurement, including for agencies that compete directly with Publicis — so “neutral” is doing a lot of work in that press release. If you run media through LiveRamp, this is the week to re-read your contract and trace where your first-party data flows and who can see it. The wider pattern is the real story: holding companies have stopped buying creative shops and started buying the data plumbing, because in an agentic market the side that owns identity owns the bid.
4 Walmart Connect pitched a global “commerce media” era
On Cannes’ opening day, Walmart Connect laid out its next act: folding the Sam’s Club Member Access Platform deeper into the offer, piping first-party audiences into Google’s Display & Video 360 for YouTube, and pitching global brands on closed-loop measurement tied to in-store and online sales. Walmart Connect’s ad revenue has roughly tripled in five years to about $6.4 billion in 2025.
Retail media’s pitch used to be “buy our shelf ads”; now it’s “use our purchase data to target and measure everywhere you already advertise, including YouTube.” That closed loop — ad seen here, product bought there, stitched to a single shopper record — is exactly what search and social can’t fully match in a cookieless world. For D2C and brand marketers the question shifts from how much to spend on Walmart to how much of your measurement you’re willing to let Walmart’s data define.
5 OpenAI courted Cannes as Criteo passed 2,000 brands on ChatGPT
OpenAI worked the Croisette hard this year, and the proof point came from Criteo, which said it has now passed 2,000 brands transacting on ChatGPT as shoppable placements scale inside the assistant and shoppable TV spreads alongside. The week’s adtech subtext was hard to miss: the conversational interface is becoming a commerce surface with a measurable advertiser roster behind it.
“2,000 brands” is the number that turns ChatGPT ads from a 2025 experiment into a 2026 channel with a rate card. When discovery moves into a chat, the product the assistant recommends inside the answer wins the sale before the shopper ever sees a results page. If you sell anything online, the homework is the same as it was for Amazon a decade ago: get your catalog, reviews, and feed clean enough that the assistant can confidently surface you — because being absent from the answer is the new being on page two.
6 DISQO launched AI Search Lift to put a meter on the AI black box
DISQO released AI Search Lift, billed as the first product to measure the incremental effect of advertising on brand activity inside LLMs and AI-powered search. Built on a consented research panel with exposed-versus-control testing, it ran in beta from Q4 2025 across automotive, insurance, beauty and personal care, CPG, and travel — and found higher-consideration categories like auto, insurance, and travel generated more brand-related activity inside the models.
As shoppers ask an AI instead of clicking ten blue links, the marketer’s oldest question — did my ad actually do anything? — went dark inside the model. AI Search Lift is an attempt to put the meter back on, and it pairs neatly with AEO and GEO: optimization tells you how to show up in the answers, lift tells you whether showing up changed anything. Expect “did it move us inside the LLM?” to become a standard line item on media plans the way brand lift became table stakes for social.
7 Google pushed AI Max for Search out of beta — agentic is now the default
At the center of Google’s post-Marketing-Live rollout, AI Max for Search moved out of beta and became the company’s preferred direction for campaign management — an agentic model where Gemini handles more of the planning, targeting, and execution that paid-media teams used to do by hand. The week’s coverage put it bluntly: Google is formalizing agentic advertising as the default way to run search.
When the biggest ad platform makes its agent the recommended path, “should we let the machine run it?” stops being a choice and becomes the new normal you actively opt out of. The leverage for marketers moves up a level — away from manual keyword and bid tuning, toward feeding the system better inputs: clean conversion data, sharp creative, and a clear business goal. The teams that win won’t be the ones fighting the agent; they’ll be the ones who give it the best brief.
8 Meta brought Live Video Ads to Instagram and Post-view ads to Reels
Meta expanded its Live Video Ads format to Instagram and added livestream-commerce tools like virtual cards, while rolling out “Post-view” ads for Reels. Across formats, the model is identical: brands hand over product catalogs and creative assets, and Meta’s AI decides which products and which formats to assemble into an ad in real time.
Meta keeps shrinking the advertiser’s job to “give us your assets and your goal” and keeps growing its AI’s job to “decide everything else.” That’s great for reach and brutal for control — you’re tuning inputs to a system you can’t fully see. The new craft is asset strategy: a deep, well-tagged library of creative and products the machine can recombine, because in an AI-assembled feed the brands that feed the model the most raw material get the most at-bats.
9 BlueConic bought Blueshift to fuse the data layer with the decision layer
Customer-data platform BlueConic acquired cross-channel engagement and decisioning vendor Blueshift, combining a CDP with AI-driven decisioning so brands can capture first-party behavior, decide the next best move, and execute across web, email, push, in-app, and SMS in one system. The companies framed it as closing the gap between what a brand knows about a customer and what it does next; the combined business now serves more than 600 customers across CPG, retail, DTC, and travel.
The pattern is now unmistakable across MoEngage, Zeta, and now BlueConic: nobody wants to be just the database or just the campaign tool — everyone is buying their way to “we hold the data and we make the call.” For buyers the upside is fewer integrations; the risk is lock-in, because once one vendor owns both your data and your decisioning, switching costs go vertical. Evaluate these stacks on how portable your data stays, not just how slick the agent demo looks.
10 Zero-click search hit 68% — and AI citations split from rankings
New figures for the week put zero-click search at roughly 68% of queries, as AI Overviews and AI answers resolve more searches without a visit, and analysts confirmed that AI-citation share has formally decoupled from traditional Google rankings — landing in the top 10 no longer reliably gets you cited in the AI answer.
Two-thirds of searches now end without a click, and the old SEO trophy — rank #1 — no longer guarantees you’re in the answer the model reads aloud. That breaks the decade-old assumption that ranking equals traffic, and it’s why AEO, first-party audiences, and owned channels surfaced in nearly every story this week. There’s the throughline for all ten items: discovery, measurement, and buying are all moving inside machines — so the brands that win are the ones building assets the machines have to reckon with, and a voice the machines can’t fake.