It’s that time of year again—where the ad industry swaps New York boardrooms for yachts in the south of France and pretends “collaboration” means sipping spritzes with a DSP logo on the glass. But beneath the sweaty desperation to stay relevant and the overwhelming scent of “eau de venture capital,” a few real conversations emerged from the chaos. At the center of many of them? CTV and retail media finally moving from awkward flirtation to full-on PDA.

And thanks to our intrepid reporter Gavin Dunaway from MediaTrust, who elbowed his way through half-naked influencers and half-baked panels, we got a front-row view of what really went down—including from a place called the Samsung Ads Floating Penthouse (which, let’s be honest, is just a yacht trying to get on the blockchain).

Let’s start where the real merger talk happened: Samsung’s floating showroom of synergy, hosted by Snowflake’s Paul Winsor and featuring a sharp lineup of Christian Russ (Samsung Ads), Alice Anson (Nectar360), and Christian Raveaux (REWE Group).

Russ kicked things off by stating the obvious—but in Cannes, the obvious still gets you applause. TV, he said, remains the best medium for brand building. But then came the real confession: “Retail media is the secret sauce.” Gavin, watching from the side with a raised eyebrow, noted that “Retail media can learn the benefits of scale from CTV—there’s a danger of going too narrow with retail data.” Translation: retail media's data is sharp, but it still thinks too small.

The bigger question, according to Russ, is where CTV’s data story goes next—and his answer was enough to make a few agency execs choke on their oysters. Imagine, he said, if payment data got layered into targeting and measurement. Gavin summed it up best: “Imagine the potential of payment data in the targeting/measurement workflow.” It wasn’t a casual idea—it was a subtle shot at the future, and everyone knew it.

Then Alice Anson brought the cold water. There’s no magical interoperability wand. Gavin quoted her noting that “clean room coverage is mixed in the UK, and requires a lot of integrations.” And while that might sound like backend hand-waving, it’s exactly the kind of technical glue that makes—or breaks—campaigns in a fragmented ecosystem. Anson also reminded the yacht crowd that data quality is still the bottleneck. “There is no AI execution strategy without data quality,” she said, then gave Snowflake a little wink: it’s “pretty great for making that data cook.”

But this wasn’t just about tech. It was also about organizational territory. As Gavin reported, Anson pointed out that “retailers need to get further up the campaign planning process—because that’s where the CTV/TV plan is being developed.” Retail media, she said, can’t be an afterthought tacked on at the end—it needs a seat at the planning table, not the kiddie table.

Christian Raveaux of REWE was maybe the most blunt of the trio. As Gavin recounted, Raveaux made it clear that “CTV and retail media are merging efficiently and it’s the best way to go up funnel and get more budget.” That’s code for: follow the money. Brands are moving dollars out of linear and into retail because they can finally see what they’re getting. But they still want the CTV “lean back” experience—just with receipts attached.

And just when you thought things couldn’t get more honest, Gavin delivered the Cannes line of the week: “CTV feels like the gym—everyone signs up, only a few show up, and most are doing it wrong.” If that doesn’t describe the current state of media investment, nothing does.

Away from the yachts and brand pontification, Dell Technologies was quietly showing the rest of the industry how it’s done. At the Nectar360 Retail Media Summit, Briana O’Dea laid out how Dell spent the last 18 months executing one of the most thoughtful global in-housing rollouts we've seen—no drama, no fireworks, just ruthless efficiency.

Dell didn’t do this to save money, though it did. They did it to move faster. By bringing data, buying, billing, and optimization under one roof, they effectively cut out the lag time agencies introduce. O'Dea made it clear this wasn’t a war against agencies—just a rebalancing. In her words, they realized that having a media agency handle billing and reconciliation “lasted five minutes.” They needed tighter control, more agility, and—most importantly—the ability to pivot when the landscape changed.

Hiring wasn’t easy. Dell had to retrain recruiters on what digital media talent even looks like, blend college grads with repurposed talent, and build custom training programs to get them up to speed. And yes, AI was already a part of the process—but not the sexy kind. They’ve been using it in planning and search for years. Now, they’re looking at how agentic systems could reshape what media buying even is.

As Gavin reported from the summit, O’Dea was refreshingly candid: they don’t know exactly what media will look like post-AI, but they’re doing the work to find out.

Back aboard the yacht, interoperability continued to rear its head like an uninvited Cannes guest. Everyone loves the idea of clean rooms—but very few are willing to admit that the integrations are still brutal. Gavin summed it up with a shrug: “It can be painful at times, but we need to meet the clients where they are.”

And that’s the tension at the heart of this whole Cannes subplot. The tools are evolving, but the planning cycles haven’t. The data is smarter, but the strategy is still trapped in linear-era org charts. Everyone wants AI to “do more,” but no one wants to clean up the data layer it depends on.

The industry is moving—just not all at once, and not in the same direction. CTV isn’t dying, it’s shape-shifting. Retail media isn’t a fad, it’s a reallocation. And clean rooms? They’re still trying to become the thing everyone pretends they already are.

And as Gavin put it best, “CTV feels like the gym.” It’s aspirational, mostly misunderstood, and littered with expensive equipment that no one knows how to use properly.

If you were walking the Croisette thisweek, you probably saw AI everywhere—on panel titles, pitch decks, and the glistening foreheads of execs trying to say “generative content” without twitching. But here’s the real headline: the actual AI deals weren’t happening on panels. They were happening in villas, behind closed doors, off-calendar, and completely off-brand for Cannes’ performative circus act.

While CMOs sat onstage talking about creative empowerment and “letting creators be authentic,” the real power players were driving up into the hills to meet with companies like Perplexity, who weren’t there to give keynotes—they were there to steal the future.

This wasn’t about AI-generated banner ads or ChatGPT copywriting demos. This was about what real AI is already doing behind the scenes—rewriting the value chain while half the industry is still obsessing over whether an AI can write a Super Bowl joke.

Perplexity’s head of advertising and shopping, Taz Patel, didn’t need a branded cabana or beach takeover. According to Digiday’s reporting, he was meeting directly with advertisers and agencies in off-grid villas, asking one question: “What kind of ads do you want to see on our AI-powered search engine?” That wasn’t an open brainstorm—it was a trapdoor into the next platform pivot.

No agenda. No preloaded answers. Just an implicit message: help us shape the monetization now, or don’t cry when we don’t need you later.

Digiday’s scoop revealed what many were whispering after the third round of cocktails: AI platforms aren’t here to compete with Google’s search business—they’re here to reroute it entirely. And the ad dollars will follow.

One exec admitted they expect Perplexity to go full ad monetization in five years, and they’re not wrong. The revenue model might start with subscriptions, but it ends—like everything else on the internet—with ads, targeting, and data-fed media optimization.

There’s a dangerous myth floating around the industry right now: that AI is mostly about content. Speeding up image generation. Writing more copy with fewer writers. Trimming costs on production.

But that’s not where the blood is pooling.

The real shift is under the hood—in planning, measurement, media allocation, and predictive modeling. The stuff that agencies used to gatekeep with spreadsheets and a poker face.

Now? AI’s got the data, the algorithms, and none of the overhead. And agencies know it.

Digiday reported that most holding company execs wouldn’t speak on the record about AI, because they’re terrified of how fast it’s moving. In public, it’s all cautious optimism and “excited for the possibilities.” Behind the scenes? They’re whiteboarding layoffs.

While everyone else was slathering on sunscreen and muttering “disruption,” Omnicom showed up with an actual roadmap. In a Cannes full of AI platitudes, they rolled out a tactical blueprint that feels almost uncomfortably relevant.

Their pitch? AI-powered influencer and livestreaming ad planning on YouTube, in partnership with Google. The move lets Omnicom clients buy around live inventory—sports, gaming, concerts—in ways that are targeted, real-time, and pre-integrated into its Omni orchestration platform.

In plain English? It’s programmatic meets Twitch, minus the cringe.

The kicker? The system’s learning which influencers are better unscripted, which ones convert better during live moments, and which ones should just stick to filtered posts. It’s machine learning + performance analytics + human nuance—at scale.

Omnicom’s Megan Pagliuca said it best (though probably not realizing how terrifying it sounded): “Any marketer not asking its media agency for a live investment strategy is already behind.”

That’s not advice. That’s a warning.

The split in the industry is now visible in neon. There are those using AI to reengineer workflows, flatten organizational layers, and optimize for margin. And then there are those still talking about “storytelling” and “human connection” like it’s 2012 and Vine just launched.

The builders are in the villas. The bystanders are on the panels.

We didn’t get a Gavin dispatch for this one (he was too busy networking responsibly), but if he had seen this play out, we’re confident he’d say something like: If your AI strategy starts with Midjourney prompts, it ends with unemployment. Brutal, but not wrong.

This isn’t AI vs. the ad industry. It’s AI replacing parts of the industry that forgot how to evolve. Platforms aren’t burning things down. They’re just building something that doesn’t need middlemen the way it used to.

And that’s the Cannes subtext no one really wanted to say out loud: the game isn’t ending. It’s just moving without you.

Some agencies—PMG, MiQ—are adapting. Automating the mundane. Selling strategy. Proving their value in the new stack. Others? Still scheduling brainstorms.

The scary part isn’t what AI can do. It’s who’s already using it while you’re still writing LinkedIn posts about “harnessing human creativity.” The real action at Cannes wasn’t on the beach. It was in the villas. And if you weren’t invited?

You’re already behind.