1 Criteo got a buyout offer — PE just ran due diligence on the ad stack
Vista Equity Partners and Quinti Capital made an unsolicited joint bid for Criteo at more than a 50% premium to its recent share price. The stock jumped 21.4% to $23.17 on Monday, putting the French adtech and commerce media firm at roughly a $1.16 billion market value. The board will review the proposal in the coming weeks — engage, seek rival bids, or stay independent.
The signal matters more than the deal. After years of on-and-off sale explorations, professional money looked hard at retargeting-turned-retail-media cash flows and decided they're worth a control premium. When PE starts writing reference checks for adtech, it means the sector's AI story has matured from promise to auditable revenue. Expect more take-private interest in undervalued mid-cap adtech this half.
2 Reddit set AI to catch AI — 23 million spam views blocked daily
In a Monday blog post, Reddit said its new automated systems now block around 23 million spam views every day before they ever reach users, while also speeding up content moderation decisions across the platform.
Reddit's comment sections are prime real estate in the AI-search era — they feed answer engines and buying decisions. That's exactly why they're being flooded with machine-written spam, and why Reddit is defending them with machines of its own. For marketers, the era of cheap "authentic-looking" seeding on Reddit is closing fast; the only durable play left is genuinely useful participation.
3 TikTok banned AI voices from live shopping — proof-of-human is now policy
TikTok updated its Live rules to prohibit non-real-time verbal interaction in shopping livestreams: no AI-generated voices, no audio recordings, no radio-style background audio. Hosts are expected to communicate with viewers in real time, verbally or through sign language.
This is the first major platform rule that treats a live human as a commerce compliance requirement. TikTok knows its shopping flywheel runs on parasocial trust, and a warehouse of looping AI voice streams would burn it. If you're building livestream commerce operations, headcount for real hosts just became non-negotiable — and a differentiator you can market.
4 Washington got the first look at GPT-5.6
OpenAI delayed the full public launch of GPT-5.6 — the Sol, Terra, and Luna lineup — after the US government requested early access under a new executive-order framework that gives Washington up to 30 days with "covered frontier models" before broad release. Access starts with a small group of vetted partners. Sam Altman: extensive safety testing "is not a bad idea. I just don't like the idea of the government picking the customers."
Marketing teams now inherit a new dependency: the frontier-model release calendar has a government review window in it. If your creative tooling, agents, or martech roadmap assumes immediate access to the newest model, build in the lag — and expect enterprise vendors to market "day-one access" as a premium feature.
5 OpenAI hired Apple's consumer marketing muscle
OpenAI appointed former Apple executive Mili Kapoor as Consumer Marketing Lead, bringing nearly two decades of consumer marketing, brand strategy, and product marketing experience across global brands.
Read this alongside OpenAI's ad-product buildout: the company isn't just selling models to developers anymore, it's building a consumer brand with iPhone-style discipline. When the maker of ChatGPT starts marketing itself like a consumer electronics giant, the competition for consumer attention — and eventually for ad dollars inside that attention — gets a lot more serious.
6 Havas created a global Chief Strategy Officer role
Havas appointed Raphaël de Andréis to the newly created role of global Chief Strategy Officer, tasked with driving client-centric growth and AI-led integration across the group. He keeps his seat as CEO of Western Europe.
Every holding company is mid-reorg around AI, but the interesting part is the job description: someone at the top now explicitly owns the AI throughline across creative, media, and data. Agencies are moving from "AI tools in every department" to "one accountable executive for the AI operating model" — a structure clients should start demanding from every partner.
7 Stitch Fix put the customer inside the ad
Stitch Fix expanded Vision, its GenAI style visualization platform, with "See it on me": browse any look in Freestyle, tap once, and get an on-demand AI image of yourself wearing it — generated from a selfie and a full-length photo, instantly shoppable, and shareable to social.
This is personalization crossing from recommendation to embodiment. The product image — ecommerce's oldest static asset — just became personal, generated, and on-demand. Watch conversion data here closely: if seeing yourself in the product measurably lifts purchase rates, every apparel and beauty brand will need a likeness pipeline, and "upload a selfie" becomes the new email capture.
8 AI ad spend is compounding 26% a year — scrutiny included
Research and Markets pegs the AI-in-advertising market at $14.12 billion in 2026, up from $11.2 billion in 2025, on its way to $36.3 billion by 2030 as brands keep increasing spend on AI-powered advertising tools.
Notice what's happening in this issue alone: buyouts, government reviews, platform bans on synthetic content — and the spend forecast still compounds at roughly 26% a year. The scrutiny isn't slowing the money; it's professionalizing it. Budget owners aren't asking "should we use AI" anymore, they're asking "which AI spend survives an audit." Position your stack accordingly.
9 Creators became AI search's trust layer
AI answer engines don't return ten blue links — they synthesize one answer and cite a small set of sources they trust. Increasingly, those sources are creators: first-person reviews and demos are exactly the material answer engines reach for. And YouTube has now overtaken Reddit as one of the strongest social signals feeding AI search authority.
Creator marketing just quietly became an AI visibility strategy, not a social tactic. A creator's honest video review can earn your brand a ChatGPT citation that your own polished product page never will. Brief your creator partners for specificity — named use cases, real numbers, on-camera demos — because that's what the machines cite.
10 US livestream commerce hit $120 billion — the human checkout is scaling
US livestream sales reached about $120 billion last year, with livestream buyers up more than 21% year over year — against China's roughly $1.2 trillion benchmark. Search interest in TikTok-specific analytics platforms like Countik and Kalodata has more than doubled in two years as operators professionalize.
Put items 3 and 10 together and the picture is clean: the fastest-growing checkout in America is a verified human talking in real time. There's 10x headroom against the China benchmark, the platforms are writing human-first rules for it, and the analytics layer is maturing. If your 2027 plan doesn't have a live commerce line item, you're planning for the last cycle.