The week that began on May 22 carried the full weight of what was announced the week before it. Google's May 2026 core update, which started rolling out on May 21 at 08:40 Pacific time, was being felt across the SEO community by the Memorial Day weekend. Search Engine Roundtable reported on May 25 that many practitioners saw significant movement, though not universally. The update followed the March 2026 cycle by approximately seven weeks - that rollout had completed on April 8 after 12 days and produced volatility in which 79.5% of URLs in top-three positions shifted, according to SE Ranking data. Whether the May update will prove more disruptive than March is not yet clear. The rollout window extends up to two weeks and no end date has been confirmed.

Complicating the picture for publishers, the same week Google confirmed a Search Console logging error that wiped all Discover performance data for May 21, the day the core update launched. Clicks and impressions in the Discover performance report for that date were zeroed out. It was the third Discover-related bug reported in recent weeks, and it landed precisely when publishers needed clean baseline data to assess ranking shifts. Search Engine Roundtable covered the confirmation on May 27, noting that Google had acknowledged the data logging issue. Alongside the update, Google officially appealed the US search monopoly ruling, a move that had been expected but that formally opened a new legal chapter in what is simultaneously the most consequential antitrust proceeding in search history.

The DMA fine and regulatory pressure

The European dimension of that pressure intensified separately. PPC Land reported on May 26 that the European Commission was preparing what would be a record fine under the Digital Markets Act - a figure described in several reports as a high triple-digit million euro sum. MediaPost covered the same development on May 26, noting the fine stems from a 2024 investigation and that EU regulators believe the company has not fully addressed anticompetitive behavior in search results even after implementing over twenty changes. The maximum exposure under the DMA is 10% of global annual turnover for a first offence, rising to 20% for repeat infringements. A separate DMA case focused on Gemini and rival AI assistants on Android is expected to reach a decision during the summer.

The fine decision rests with Commission President Ursula von der Leyen. Its timing - before the summer recess - reflects a deliberate regulatory calendar. The outcome will carry practical consequences for how Google distributes visibility across shopping, travel, and local services in European search results. A separate GDPR enforcement action landed in the same week. Spain's data protection authority, the AEPD, fined Amadeus IT Group EUR 14.4 million on May 27 for building passenger profiles from millions of booking records without a valid lawful basis. The global travel technology company, whose reservation infrastructure underpins airlines, hotels, and travel agencies worldwide, had processed sensitive itinerary data at scale. The Amadeus case is distinct from advertising enforcement but illustrates how data profiling at commercial scale continues to draw regulatory action across vertical categories, not only digital media.

Publicis, LiveRamp, and the fight over infrastructure

The story that dominated the prior week continued to generate analysis through May 22-28. Publicis Groupe announced on May 17 an agreement to acquire LiveRamp for a total enterprise value of $2.2 billion, with an equity value of $2.546 billion representing a 29.8% premium to LiveRamp's closing price on May 15. LiveRamp reported Q4 revenue of $206 million, up 9% year-over-year, in the same announcement, with subscription net retention at 107% and fiscal year 2026 revenue reaching $813 million. Publicis framed the acquisition as a bet on data co-creation for the agentic era - LiveRamp connects over 25,000 publisher domains and 500-plus technology and data partners across 14 markets.

Digiday reported on May 27 that the deal is reshaping the neutrality debate around ad tech infrastructure. The central concern is straightforward: LiveRamp's data collaboration platform has operated as a shared utility across the industry, connecting competing agencies, brands, and platforms. Its absorption into a holding company - and the largest one by ad revenue - raises questions about whether those connections will remain neutral or become strategically weighted. Digiday also noted separately that Omnicom has accelerated its exit from LiveRamp following the announcement, citing the conflict created by a key infrastructure tool being owned by its primary competitor. The transaction is expected to close before year-end 2026, subject to regulatory approvals and LiveRamp shareholder sign-off. LiveRamp will continue to operate as an independent business within Publicis, with existing CEO Scott Howe remaining in place.

What makes the deal structurally significant is not the price alone but what LiveRamp enables in an agentic advertising environment. Identity resolution at scale - connecting fragmented user signals across devices, publishers, and channels - is the foundational layer that makes AI-driven targeting meaningful. Whoever controls that layer influences where automated buying systems route impressions. The adtech industry has spent several years arguing over whether clean rooms, first-party data, and contextual signals could replace third-party cookies. The Publicis-LiveRamp deal reframes that debate around who owns the infrastructure for whatever replaces them.

ChatGPT ads: building fast, measuring slower

OpenAI's advertising business moved faster this week than its measurement infrastructure can comfortably keep pace with. PPC Land reported on May 22 that OpenAI had added daily budgets, geo-targeting, and dynamic call-to-action options to the ChatGPT Ads Manager. These changes arrived roughly three weeks after May 5, when OpenAI opened the self-serve Ads Manager to all US businesses, dropped its minimum spend threshold entirely - it had started at $200,000 to $250,000 when the pilot launched in February, fell to $50,000 in April - and introduced CPC bidding alongside the existing CPM model. The recommended starting bid for CPC is $3 to $5 per click.

Search Engine Roundtable confirmed on May 22 that OpenAI released the Ads Manager features including daily budgets and geo-targeting in the same week Google launched its core update. Geo-targeting is a particularly notable addition: it means advertisers can now serve ads to users in specific countries or regions within ChatGPT, which is relevant to any brand with geographic restrictions or local targeting requirements. Digiday reported on May 21 that OpenAI had also introduced a new visual ad format: a larger image unit with an optional call-to-action button that advertisers can personalize. The format still appears below the ChatGPT response, clearly labelled as sponsored.

The advertising infrastructure is assembling quickly. Digiday noted on May 26 that ad delivery has improved in recent weeks but that early doubts about the platform are not easily erased. The original $250,000 minimum spend requirement created a self-selecting pool of advertisers willing to absorb substantial uncertainty. Now that the threshold is gone, the platform's performance claims face a wider and more demanding audience. A Conversions API that OpenAI published on May 5 is the measurement backbone, but post-click attribution at meaningful scale requires brands to implement pixel tracking individually - a time-intensive process that OpenAI has reportedly been managing case by case rather than through automated onboarding. Digiday reported on May 27 that OpenAI is preparing for a potential IPO as early as September, giving the advertising business both a strategic timeline and an investor audience that will evaluate its revenue traction closely.

The parallel story running underneath the ads business is scale. Similarweb data cited by Search Engine Roundtable on May 27 showed that since OpenAI began surfacing more prominent links in ChatGPT answers, referral traffic to publishers is up 150%, pageviews per visit are up 24%, and time on site is up 11%. That is a different kind of value proposition from traditional search advertising: not just impressions or clicks but users who arrive with higher engagement depth. Whether that translates into conversion rates that justify CPC bids at $3 to $5 is the central open question.

Google's AI Mode hits a billion users - and adds ads

The week's numbers on search scale were striking. Google reported that AI Mode had crossed one billion monthly active users globally as of May 2026, data published by Shivani Mohan, Vice President of Data Science and UXR at Google Search, alongside the Google I/O 2026 announcements on May 19. Query volumes in AI Mode have more than doubled every quarter since the feature's US launch in May 2025. The average AI Mode query is now triple the length of a traditional search query, a behavioral shift with direct implications for keyword targeting and ad relevance scoring.

Into that surface, ads are being placed. PPC Land reported on May 25 that Google has introduced sponsored results into AI Mode responses. The format shows ads embedded within AI-generated answers rather than as discrete sidebar or top-of-page units. This follows Google extending sponsored ads into the free listing grid in the Shopping tab on April 20 - that made AI Mode the third distinct surface where paid results have entered what had been organic-only environments. At one billion users, design decisions about ad placement in AI Mode affect a base comparable to traditional web search.

Google on May 27 expanded its Preferred Sources feature into AI Overviews and AI Mode, allowing users to designate sites whose links receive visible badges inside AI responses. The announcement, detailed by Google Product Manager Duncan Osborn, also introduced prominent topic carousels and "Highly Cited" labels. The timing matters: it follows a concentrated sequence of changes - the outbound link update on May 6, the AI Mode behavioral data report on May 19, and the Google I/O announcements on the same day. PPC Land noted that the trajectory across May 2026 represents the most concentrated period of structural change to Google's AI search surfaces since AI Mode launched. Whether carousels and "Highly Cited" labels meaningfully alter referral traffic for publishers remains an empirical question - the previous link outbound update on May 6 was accompanied by claims of improved click-through, but no external verification has been published.

Programmatic infrastructure: RTB at 20 and agentic DSPs

AdExchanger published an in-depth AdExplainer on May 21 tracing how the RTB auction has changed in the roughly 20 years since the first automated digital ad buy. The piece describes a system that has grown dramatically in complexity: publishers generating bid requests, SSPs packaging and routing them, DSPs evaluating them in milliseconds, and generative AI now beginning to reshape each of those steps. The article is notable not for breaking news but for its compression of structural change: what was once a simple two-party transaction now involves multiple intermediaries operating at sub-second latency across display, video, CTV, audio, and retail media simultaneously.

That complexity is precisely what Madhive targeted when it announced on May 27 the expansion of its Maverick AI platform to include agentic capabilities. Madhive, a DSP focused on local television advertising, added an MCP server to Maverick that allows large language models to interact with the platform's data and decision layer across 50,000 daily campaigns. The local TV market has historically operated with higher manual overhead than national programmatic: geo-specific inventory, diverse station ownership structures, and fragmented measurement. MCP server access is a technical architectural choice that allows LLM-based tools - including external agent systems - to query and manipulate campaign parameters through a standardized interface rather than a bespoke API integration.

PubMatic disclosed during its Q1 2026 earnings call on May 8 that its AgenticOS platform had run more than 30 fully autonomous end-to-end agentic campaigns, in addition to over 1,000 direct publisher deals. Agentic deal activity drove 80% year-over-year growth in PubMatic's emerging revenue category, which now represents 14% of total revenue. The company's overall Q1 revenue reached $62.6 million, up 13% year-over-year, with CTV growing 18% and mobile app revenue up 25%. The earnings context frames the AgenticOS numbers: agentic campaigns are still a small absolute share of the SSP's business but are growing faster than every other category it reports.

Piracy, sports ratings, and measurement distortion

AdExchanger reported on May 26 on a new Adalytics study documenting how illicit sports livestreaming services - including operations identified as WatchSports and StreamSports99 - are distributing unauthorized streams of high-profile events to hundreds of thousands of devices. The measurement consequence is direct: those viewing sessions do not appear in Nielsen or Nielsen-equivalent measurement systems. Advertisers buying against sports ratings are therefore bidding on audiences that are, in part, watching through unmonitored channels. The rating numbers that determine sports inventory prices are incomplete by a structurally unmeasurable margin. Adalytics is an ad tech auditing firm, and its reports carry the caveat that they are produced with methodology that not all parties in the ecosystem have accepted as standard. The study nonetheless surfaces a specific and underreported distortion in one of advertising's most premium categories.

AI adoption rate: 65% of Americans weekly

A quantitative anchor for the week's advertising developments came from Edison Research and SSRS, whose data showed that 65% of US adults - approximately 175.5 million people - used AI weekly as of May 2026, up from 52% in February. That is a gain of 36 million weekly users in three months. The report, covered by PPC Land on May 27, identified Gemini as the platform with the fastest growth rate among AI tools. ChatGPT remains the largest platform by absolute users, but the Gemini growth rate is relevant to any advertiser evaluating which AI surface commands the most commercially active audience. The trajectory also provides context for the AI Mode billion-user figure: Google's AI products are growing into a mass consumer behavior, not a specialist one.

Comscore data cited by PPC Land on May 26 showed Anthropic's Claude growing 130% month-over-month in March 2026 to reach 2.66 million US desktop users, while ChatGPT held its dominant lead. The Comscore figures are desktop-only and therefore undercounting mobile usage, which skews younger. The directional trend - a widening field of competitive AI platforms each growing quickly - is consistent with the Edison/SSRS consumer data showing broad adoption across tools rather than consolidation around a single platform.

YouTube, DV360 API, and platform mechanics

YouTube launched a prompt-driven custom feed on the Home tab for signed-in US users on May 27, allowing viewers to type a text prompt describing what they want to watch and receive a refreshed content feed. The feature is functionally a consumer-facing application of the same retrieval and recommendation logic that underlies AI Mode in search: natural language input producing a curated output. Its advertising implications are indirect but real - custom feeds change which content categories and channels viewers encounter, which in turn affects contextual adjacency for ads served alongside that content.

On the same day, PPC Land reported that Google will add full Demand Gen support to the Display and Video 360 API starting June 10, 2026, completing a rollout that had begun in closed beta. Demand Gen - Google's cross-surface campaign type spanning YouTube, Gmail, and Discover - has been available in the DV360 interface but not fully accessible through the API. Starting June 10, Demand Gen line items, ad groups, and individual ads will all be manageable programmatically through the DV360 API. For agencies and enterprise advertisers that manage DV360 campaigns through automated systems or third-party tools, this closes a gap that has required manual workarounds for any Demand Gen activity at scale.

YouTube also announced at Google I/O on May 19 the introduction of personal AI avatars and Gemini Omni remixing for Shorts, with SynthID watermarks and granular creator opt-out controls. The opt-out mechanism is a direct response to creator concern about likeness usage. SynthID, DeepMind's provenance watermarking system, embeds a signal in AI-generated video and audio that persists through compression and re-encoding. Whether the watermark is reliably detectable across distribution channels - particularly on third-party platforms - is a technical question that remains open.

Podcast ad spend and the audio economy

Magellan AI data published on May 27 by PPC Land showed global podcast advertising reaching $408 million in December 2025, a seasonal peak. Beyond the headline figure, the geographic breakdown is the more structurally interesting data point: Germany added 259% more brands to podcast advertising year-over-year in Q1 2026, with financial services leading growth in markets outside the US. iHeartMedia separately disclosed to investors in May that it expects its own programmatic radio revenue to reach approximately $200 million in 2026, up roughly 50% from the $135 million reported in 2025. The audio programmatic expansion is happening in both directions: premium podcast inventory attracting brand-side spending and broadcast radio inventory entering automated buying systems at higher rates than in prior years.

A case with structural implications for CTV advertising is heading toward trial. PPC Land reported on May 27 that a California jury will decide in August 2026 whether Vizio must release the Linux-based source code for its SmartCast platform. SmartCast is the operating system underpinning Vizio's smart TVs and the software layer that determines how advertising inventory is allocated, measured, and delivered on those devices. The case turns on whether SmartCast qualifies as a derivative work under the GPL open-source license. If the court rules that it does, Vizio would be required to publish the source code - a disclosure that would expose the technical architecture of one of the major US CTV ad platforms. The advertising implications range from competitive intelligence risk to potential changes in how Vizio's ACR data handling is documented and audited. The trial is scheduled for August 2026 in California.

Timeline

Share this article
The link has been copied!