A pair of closely related data points shaped the marketing conversation on June 9 and June 10, 2026. One projected a $68 billion market. The other said that the most-discussed platform in that market would capture roughly a tenth of what its own company has told investors to expect. Together, they forced a reckoning with what "AI advertising" actually means, and where the money flows in a period when everyone in the industry is being asked to bet before the numbers are in.

The $68 billion forecast that deflates the chatbot narrative

The dominant story of the day came from EMARKETER. On June 4, the research firm published its US AI Advertising Forecast 2026, projecting that total US spending in and around AI platforms will more than double from $32.03 billion this year to $68.25 billion by 2030. The headline number is large. The structure beneath it is pointed.

More than 80% of that spending, EMARKETER found, flows next to AI content rather than inside AI chatbots. Traditional keyword-based search ads placed alongside Google AI Overviews constitute the biggest category. They dwarf what the firm calls AI conversational search and, further still, the AI chatbot segment that includes ChatGPT, Microsoft Copilot, and Gemini. The report was authored by Nate Elliott, EMARKETER's principal analyst for AI in marketing and commerce, who joined the firm in August 2025 after nearly nine years running his own consultancy and earlier roles at Morning Consult and with clients including LinkedIn, the IAB, Hootsuite, and Digiday.

Elliott posted a summary on LinkedIn on June 9 that pulled no punches: "OpenAI thinks ChatGPT will collect about $50 billion in US ad revenues in 2030. We think their entire addressable market will be 1/10th that." Three structural headwinds underpin the chatbot forecast. First, chatbots cannot carry the same advertising density as search engines. To reach even the figures EMARKETER projects for the chatbot segment by 2030, platforms would need paid placements in nearly every commercial response. Second, CPMs are already falling. ChatGPT launched its advertising pilot at a $60 CPM, and the forecast projects that figure will compress to roughly $15 by 2030. April 2026 data documented CPMs dropping to $25 within weeks of the pilot's February launch, and the compression has not stopped. Third, chatbots will not dominate consumer AI usage; they are one surface among many. The majority of consumer AI interaction, and therefore the majority of advertising opportunity, lives in search-adjacent formats.

The forecast covers three distinct ad types: AI search-adjacent advertising, the largest category; AI conversational search advertising, covering ads inside search engine-based chatbot experiences such as Google AI Mode; and AI chatbot advertising, the smallest segment, referring to ads inside standalone large language model platforms. The chart accompanying the report shows the AI chatbot slice remaining visibly thin across all five years of the projection. By 2030, even under optimistic assumptions, the US chatbot advertising market reaches just over $5 billion. That figure is the ceiling, not the floor.

This matters beyond the OpenAI debate because it reframes what the industry is actually competing for. Barclays had earlier projected ChatGPT ad revenue at $102 billion globally by 2030. Google Search and YouTube ads alone generated $261 billion in 2025. OpenAI has neither a search index nor a video library at comparable scale, and it is trying to build its advertising business from a standing start. The gap between aspiration and addressable market is the central tension the industry is now pricing in.

The AdExchanger coverage published June 9 captured the practitioner mood with precision. Marketers and media buyers are gradually getting more comfortable with AI ad campaigns, but "gradually" is the operative word. Brainlabs' head of programmatic Ben Kahan characterised the spend as "putting money into a black box." A Digiday+ Research survey published June 10, drawing on more than 100 marketing professionals polled in Q1 2026, found that more than four in ten marketers report using AI in social media (49%) or retail media (42%) campaigns, but skepticism remains high for AI-native ad buying. Trust is the friction point.

The Criteo data gives that friction point a specific shape. On June 5, AdExchanger reported that Criteo dropped its ChatGPT minimum investment from $50,000 to $10,000 and announced it would match every dollar clients spend through the platform. Criteo had become OpenAI's first ad tech partner in March 2026. Cutting the entry minimum by 80% within three months is a demand-generation move, not a reflection of inbound pressure. The matching program is a subsidy. The incentives arrive at exactly the moment when buyer interest exists but committed spend has not followed.

MediaPost's coverage from June 9 noted that financial investments in performance advertising are rising as AI-driven marketing methods expand, but that the pressure from IPO-bound platforms is creating a distorted picture of what the actual addressable market for AI-native advertising looks like. Brands are being asked to test platforms whose economics are still being engineered in public. The EMARKETER forecast provides the clearest available calibration of what the realistic ceiling looks like.

The adjacent-content dominance in the forecast also has a direct implication for the Semrush research on LinkedIn citations published on the same day. If 80% of AI advertising through 2030 flows next to AI-generated search content rather than inside chatbots, then the discipline of getting content cited in AI search responses is financially material in a way that chatbot optimisation is not. Semrush's analysis of 89,000 LinkedIn URLs cited by ChatGPT Search, Google AI Mode, and Perplexity, drawn from 325,000 prompts submitted between January and February 2026, found that LinkedIn ranked second in AI citation frequency at 11%, just below Reddit at 11.29% and above Wikipedia at 9.53%. Perplexity cites Company Pages 59% of the time. ChatGPT Search and Google AI Mode favour individual creator posts 59% of the time. Articles account for 50% to 66% of LinkedIn citations across all three platforms.

The asymmetry between platforms matters for B2B content strategy. A brand that publishes only on its Company Page will reach Perplexity but will be largely absent from the citation pools used by ChatGPT and Google AI Mode. Employee content fills the gap those two platforms prefer, but only if the individuals are producing consistent, well-structured posts that surface for specific commercial queries. The playbook Semrush published on June 9 is one of the first dataset-backed guides to AI search visibility that operates at this level of technical specificity.

Elkjop's NOK 20 million GDPR fine and what it exposes about Customer Match

While the AI advertising debate played out across forecasting desks, regulators in Oslo were publishing a different kind of document. On June 9, Norway's Datatilsynet confirmed that it had fined Elkjop Nordic AS and Elkjop Norge AS a combined NOK 20 million for GDPR violations affecting more than six million loyalty club members across the Nordic region. The decision, dated June 1, 2026, closed an investigation that began with an on-site inspection in June 2022, nearly four years earlier.

The case is not principally about a data breach. There was no hack. The violations were structural and deliberate: the retailer's loyalty club consent mechanism was bundled, non-specific, and not freely given. Customers who joined the club had to accept every form of data processing, covering profiling, personalisation, newsletter distribution, analytics, and SMS marketing, in a single all-or-nothing step. Datatilsynet found this violated the GDPR's requirement that consent be granular and separable for distinct purposes. The Elkjop compliance team's own internal description of the design, captured in the ruling, was "all or nothing" and "package." The company had also internally identified in February 2022 that a supervisory authority could find the consent invalid, and had kept the mechanism in place anyway.

At store checkouts, the quality of the disclosure that customers received varied depending on which individual employee was serving them, creating a structural risk that the information was inconsistently delivered and frequently inadequate. The authority noted this explicitly as an aggravating factor.

The second violation is more technically instructive for practitioners working with customer data platforms and advertising tools. Elkjop planned to implement Customer Match using data collected under the loyalty club consent, treating that upstream consent as sufficient legal basis for feeding customer contact details into advertising platform audience matching. The Norwegian authority rejected this on purpose-limitation grounds under Article 6(4) of the GDPR. Customers who joined the club to receive discounts could not have reasonably anticipated their contact details being shared with advertising platforms for audience matching and retargeting. No compatibility assessment had been carried out. No legitimate interest analysis had been documented. The company assumed the purposes were equivalent and proceeded.

The authority also rejected Elkjop's secondary argument that legitimate interest under Article 6(1)(f) could authorise the Customer Match use. The fairness principle in Article 5(1)(a) requires that data not be processed in ways that are unexpected or misleading to data subjects. Receiving discounts and having personal data shared with third-party advertising platforms for audience matching are materially different experiences from a customer's perspective.

The third violation covered inadequate offline conversion documentation, meaning Elkjop could not demonstrate that the conversions attributed to digital campaigns had been documented with sufficient precision to satisfy GDPR requirements. The fourth covered the inclusion of children. Children as young as 15 were eligible for the club. No age verification mechanism existed. Gaming products on the site attracted younger visitors, and the authority concluded that children were enrolled in the profiling systematically.

Customer Match is now undergoing significant technical restructuring at the platform level. Google required all uploads to migrate to its Data Manager API by April 1, 2026. Google added IP address support to Customer Match in the Data Manager API on May 28, 2026, with match-rate improvements expected from Q3 2026, though DV360 remains excluded from that expansion. The timing of the Elkjop ruling, published as these technical changes are rolling out, places the legal requirements and the infrastructure changes in proximity that any advertiser using Customer Match in Europe should be aware of. Retailers across the Nordic region, and more broadly across the EU, using Customer Match on the basis of loyalty program consent face exactly the same exposure the Elkjop case describes.

In a separate but related GDPR development published by PPC Land on June 8, a Dusseldorf court referred GDPR Article 26(2) to the Court of Justice of the European Union, asking whether joint controller terms must appear publicly onlinerather than remaining confidential between the parties to a data processing arrangement. The referral carries major implications for ad tech platforms operating as joint controllers with publishers, where data processing terms are currently kept private between parties rather than disclosed to users. If the CJEU rules that such terms must be publicly accessible, it could require platforms to restructure contracts and public-facing documentation across the entire European market.

TikTok Shop earns a dedicated budget line in agency RFPs

On June 9, AdExchanger reported that TikTok Shop has moved from novelty to named channel in formal agency briefs. Brands with $30 million or more in annual revenue saw 97% year-over-year growth on TikTok Shop and a nearly 80% increase in transaction volume, figures that explain why agencies are receiving briefs that name TikTok Shop specifically alongside Amazon and Walmart. Acadia, which manages Amazon and Walmart strategies for mid-market and enterprise brands, saw three RFPs in the past six months that specifically named TikTok Shop as a required channel. Jared Belsky, Acadia's CEO, described it as "the start of a signal."

The shift is notable for what it means for budget fragmentation. TikTok Shop earning a dedicated line in media briefs represents a structural change in how retail media allocations are made, placing the platform alongside the two dominant commerce media operators rather than in a catch-all social media bucket. The CPM comparison drives part of the logic: TikTok Shop offers lower CPMs than other social news feeds, combined with a younger demographic that Amazon and Walmart underindex against.

The platform's own governance apparatus has been expanding at pace in parallel, and PPC Land published extensive documentation of that expansion this week. TikTok Shop's "Deals for You Days 2026" runs June 17 through July 2, and it introduces a live scoring mechanism capable of removing creators from campaigns mid-event based on daily Promotion Performance Scores. A separate Creator Enforcement Policy now includes frozen commissions, e-commerce bans, and a 90-day repeat-violation rule that strips monetisation access. The Creator Score system, rating creators from 0 to 5 daily, gates product reach and commission tiers. An account health score is replacing the existing violation points system in July 2026.

The content compliance layer is equally detailed. Rules banning AI-generated voices, still frames exceeding 50% of screen, and product page screenshots from promotional LIVEs are now binding on all creators. Expiration date rules published June 5 cover food, beauty, supplements, and medical devices, with specific shelf-life compliance thresholds. Gambling policy rules from May ban raffles, lucky spins, and most forms of card break selling. Fan Groups transform creator DMs into shoppable communities, capping each group at 300 members, while Countdown Bidding enables real-time LIVE auctions with strict eligibility thresholds and binding bids.

The shoppable video guide published by TikTok Shop Academy on May 20, 2026, covering four content archetypes and a 3-to-6-second hook rule, and the Creator Intelligence Series publishing playbooks for 18 creators, round out a governance and creative infrastructure that is more developed than most traditional media buying processes at major agencies.

For agencies writing those TikTok-named RFPs, the operational overhead is substantially higher than for a conventional social media campaign. The platform that earns the dedicated budget line is not the one that existed twelve months ago.

Google search ranking volatility, the SEO documentation update, and AI Mode

Barry Schwartz at Search Engine Roundtable documented on June 8 that Google search ranking volatility continued well beyond the completion of the May 2026 core update on June 2. SEO community chatter spiked from Friday June 5 through the weekend, even as most third-party tracking tools showed relatively calm readings. That disconnect between practitioner observations and tool readings is a recurring feature of core update cycles, where the effects visible to individual site owners often lag or diverge from aggregated volatility signals.

The timing overlapped with a separate documentation change. On June 5, Google revised its Search Central documentation, releasing both an updated "Do you need an SEO?" guide and a new companion document on evaluating third-party SEO tools, services, and advice. The companion document warns specifically against guarantees of ranking improvements, advice tied to unverified tools, and claims related to AEO (Answer Engine Optimisation) and GEO (Generative Engine Optimisation) that Google has not validated. The guidance also adds a section specifically addressing how SEOs should approach optimisation for generative AI. PPC Land reported the update on June 8, noting that the documentation flags risks around AI-generated optimisation recommendations and unverified third-party claims.

On the same day, Google confirmed that Google Chrome will not set AI Mode as the default search experience, walking back a Chrome Canary experimental flag that had briefly redirected address-bar queries to AI Mode. Rajan Patel, a Google VP, characterised the flag as an error. The clarification matters for advertisers tracking how AI Mode will affect traditional search ad inventory volumes, since the mode's default-off status limits how quickly it displaces conventional results pages where the majority of search advertising currently runs.

Google also launched Search profiles in the US, giving publishers and creators a dedicated page on Search and Discover to aggregate articles, videos, and social posts, with audience analytics attached. The feature requires 100,000 followers to access, though Search Engine Roundtable noted that researchers found active profiles that appeared to predate those requirements, raising questions about the beta rollout process. A new Insights section testing inside Search profiles, where publishers can see how searchers interact with their content on Google, was also documented.

Google Merchant Center added a further reporting layer, introducing AI performance insights covering AI Mode, AI Overviews, and the Gemini app, with share of voice and funnel data across five markets. For retailers attempting to understand their visibility inside AI-generated product responses, this is the first native measurement tool that addresses the channel directly. Google Analytics linking Google Business Profile data directly to surface seven GBP metrics, including calls, bookings, directions, website clicks, and messages, inside GA4 reporting, arrived on the same day. For local advertisers, the ability to correlate offline and online customer interactions in a single view without manual data exports is a meaningful operational simplification.

The same week saw GA4 Measurement Protocol enter maintenance mode, with Google confirming no future enhancements are planned for the protocol. The Data Manager API is now the designated forward path for server-side data ingestion. Practitioners relying on the Measurement Protocol for custom event tracking need to evaluate the migration timeline, because the divergence between the two paths will only widen.

Bing added a toggle allowing users to disable AI Copilot search responses with a single click via a browser extension for Chrome and Edge, confirmed by Search Engine Roundtable on June 8. The toggle, which can also be triggered by adding "-ai" to the end of a query, reflects a deliberate choice to let users control AI response visibility. The positioning is particularly interesting in the context of Google's clarification about AI Mode defaults: both major search engines are, for now, treating AI-generated responses as an optional layer rather than a mandatory default. The implications for advertisers monitoring AI-adjacent search ad impression volumes are direct.

World Cup infrastructure, Fire TV, and connected-TV distribution

The FIFA World Cup 2026 opens June 11, and the advertising and media distribution machinery assembled around it reflects the connected-TV and retail media ecosystem in its current state. Amazon's Fire TV launched a dedicated World Cup hub on June 9, integrating FOX One for coverage of all 104 matches across nine countries, with free matches on Tubi and Alexa+ voice commands for match navigation. Bauer Media Audio became an official Fire TV Channels launch partner on the same day, bringing KISS, Hits Radio, and Irish brands to CTV users without subscription requirements. The pairing of sports rights and audio distribution in a single platform launch illustrates how Fire TV is positioning itself as a comprehensive media hub rather than a premium video-only destination.

FOX Sports and ReachTV will air all 104 matches live across 80-plus US airports, reaching 51 million monthly travellers, from June 11 through July 19. On the search side, Google activated live scores, agentic ticket booking, Street View stadium imagery, and Gemini-generated match briefings across Search, Maps, Waze, and the Gemini app. The agentic ticket booking feature is particularly notable because it operationalises the kind of AI-assisted transaction completion that Google has been describing as a future capability, using a live event to test real conversion flows.

Adweek reported that US brands are still struggling to reach Hispanic audiences around the tournament, noting that the Hispanic population in the US has nearly tripled in the 30 years since America last hosted the competition. Approximately 45 million people in the US speak Spanish as a primary language, but significant brand investment is going into campaigns that do not reflect that audience's cultural context or language. The piece cites linguists and multicultural marketing specialists who argue that translation alone is insufficient and that genuine cultural resonance requires different creative strategy from the outset.

The out-of-home picture around the World Cup spans traditional formats and newer digital surfaces. Shakira and Burna Boy's FIFA World Cup 2026 anthem "Dai Dai" debuted at YouTube Global number one with 44 million views on June 4, ending "NO BATIDAO"'s 31-week chart presence, with Michael Jackson dominating in 10 territories. The chart performance is relevant to media buyers because it illustrates the scale of organic distribution that major tournament anthems command, and the kind of cultural context against which paid placements on YouTube will run through July.

C-suite changes, Trade Desk turnover, and the industry talent market

Adweek reported exclusively on June 7 that The Trade Desk's chief revenue officer Anders Mortensen exited after just seven months in the role. No official reason was given. Mortensen joined during a period of significant operational pressure for the DSP, which had been navigating Q4 2025 revenue shortfalls and an extended period of reduced advertiser confidence following widely reported data quality and platform integration issues. The CRO exit is the latest in a series of senior departures at the company. AdExchanger's June 9 roundup flagged that executive turnover continues at The Trade Desk as one of three headline items for the day.

The turnover at the DSP reflects a broader pattern of C-suite restructuring around AI capability, documented in the same AdExchanger piece. Communications leaders are reshaping the C-suite across multiple holding companies, with AI operational literacy becoming a formal evaluation criterion in senior hiring rather than an implicit expectation. StatSocial launched Digital Twins on June 9, a product that allows brands to simulate audience research using AI-generated profiles built from behavioural data. Shepherd, an audience strategy consultancy, reported using it to test news products for niche audiences before committing to editorial investment. The product reflects a shift in research methodology: AI-generated simulations run against real behavioural data as a pre-spend validation step, replacing or preceding survey-based audience research.

The agency market is also consolidating around retail media capability. Adweek reported on June 5 that agencies including Podean, Harvest Group, and EVOQ Group are actively acquiring Amazon and Walmart specialists, driven by client demand for dedicated commerce media execution. The acquisitions follow a pattern where holding companies and independents alike are racing to build proprietary retail media capability rather than relying on platform-side managed services. The same period saw GoPuff expand its retail media offering through a GrowthLoop partnership for more specific audience targeting.

Apple and Siri AI: the implications for app marketing

At WWDC 2026, Apple unveiled Siri AI, built on Google Gemini models, alongside new App Store Creative Assets, ad coordination tools, and changes to subscription management. The EU launch of Siri AI faces a delay, with Apple citing compliance work under the Digital Markets Act. For app marketers, the Creative Assets feature changes how assets can be staged and tested in App Store search ads, and the ad coordination updates affect how Apple Search Ads interoperate with organic App Store visibility.

MediaPost coverage from June 5 noted that Apple was seeking to convince advertisers of its advanced agentic capabilities just ahead of the WWDC event, framing the Gemini-based Siri AI as a meaningful step toward the kind of agent-driven consumer interaction that could eventually mediate app discovery and in-app commerce. The Apple App Store generated $151 billion in advertising revenue in 2025, with AI-featured apps growing four times faster than the App Store average and 850 million users browsing weekly. The scale of that platform context makes the Siri AI launch commercially significant for app advertisers regardless of the EU delay.

The broader AI infrastructure investment picture received parallel coverage. Pinterest committed $4 billion to AWS through 2031, using Trainium and Graviton chips to train visual search models powering over 600 million monthly users. The deal positions Pinterest's AI-powered discovery product as a cloud-native capability rather than an on-premise build, and it provides Amazon Web Services with a major public anchor tenant for its AI infrastructure offerings. Cloudflare added spend limits to AI Gateway on June 5, allowing firms to cap AI costs in dollars by team, user, or model, with fallback routing when budgets run out. The spend management feature addresses one of the most practical operational concerns around AI adoption: unexpected cost overruns from token consumption that were not anticipated in media or technology budgets.

The Spend, covering the financial context around ad tech, highlighted on June 6 that token economics are reshaping how companies pay for AI, and that the bills are nothing like what anyone expected. The article documented how enterprise AI cost structures are diverging sharply from initial projections, with token usage at scale producing cost curves that were not modeled in the buying decisions made six to twelve months earlier. That dynamic is directly relevant to the EMARKETER forecast: chatbot advertising CPM compression is partly an advertiser-side story, but it is also a platform-side economics story, where the cost of generating each AI response competes directly with the revenue that response can carry.

Privacy infrastructure: bots, tracking, and audience data

Two data infrastructure stories from the past 48 hours carry direct implications for measurement and audience planning. The first concerns web traffic composition. Cloudflare data published on June 5 documented that bots now represent 57.4% of all web traffic, with AI training crawlers alone accounting for 50.6%. The finding confirms what had been reported anecdotally across publisher communities: AI brand mentions generated by crawlers are proving commercially worthless for publishers whose human traffic and referral revenue have declined. The measurement gap, where AI systems mention publishers without sending users to their sites, is now documented at scale.

The second concerns tracking visibility. isblocked.fyi launched as a tool that aggregates 22 major adblocker filter lists to reveal whether tracking domains are silently blocked, with free daily lookup and email alerts. For advertisers relying on pixel-based conversion tracking, the tool provides a diagnostic layer that was previously unavailable without manually testing against individual filter lists. The tool is particularly relevant in the context of the GA4 Measurement Protocol entering maintenance mode: as server-side tracking via the Data Manager API becomes the forward path, the exposure to adblocker-level filtering increases for those who have not yet migrated.

Microsoft Advertising's September 2 UTM auto-tagging update will separate Search, Audience, Shopping, and Performance Max campaign types in analytics, fixing a longstanding attribution gap that had blended distinct campaign performance into a single traffic source in GA4 and other analytics platforms. The change, confirmed by Search Engine Roundtable on June 5, requires no action from advertisers but will alter how campaign performance reads in historical and comparative reporting from September onward. Anyone with automated reporting built on current UTM structures will need to update downstream views.

Microsoft Advertising also lifted the global ban on crypto exchange Audience Ads in May 2026, ending a four-year restriction that had confined crypto exchange brands to search placements across the Audience Network. The policy change opens programmatic display inventory to a category that had been excluded since 2022.

Also noted

  • June 9, 2026: Bots now represent 57.4% of all web traffic by Cloudflare's measurement, with AI training crawlers alone at 50.6%, and AI brand mentions are proving commercially worthless for publishers losing referral traffic as the human-to-bot traffic ratio inverts. PPC Land
  • June 8, 2026: Barb's UK TV campaign analytics platform hit 600 users since its January launch, with Amazon Prime Video integration now under way, consolidating television measurement across linear and streaming in a single planning environment for the first time in the UK market. PPC Land
  • June 7, 2026: The Trade Desk CRO Anders Mortensen exited after just seven months in the role, adding to an extended pattern of executive turnover at the DSP as it navigates performance pressures and efforts to rebuild advertiser confidence. Adweek
  • June 8, 2026: Google upgraded NotebookLM on June 8 with Gemini 3.5, a secure cloud computer, more than 100 software skills, and 11 new downloadable output formats for paid tiers, expanding the tool from a research note-taking product to a more general knowledge work platform. PPC Land
  • June 6, 2026: Liftoff Mobile priced its IPO at $23 per share on June 3, raising $437 million and debuting on Nasdaq under the ticker LFTO above $30, making it the first ad tech IPO since MNTN listed in 2025, and providing an external market valuation reference for the mobile app advertising sector. PPC Land