Roku’s $19bn sale talks reshape CTV’s ad ownership question
Roku explores a full sale at $19bn. LinkedIn’s B2B creator marketplace, Google’s Data Manager API v1.7, and the agentic ad stack wave arrive simultaneously.
There is a sentence that appears in almost every Roku press release, earnings letter, and investor call: the company reaches more than 100 million streaming households. It has been deployed so often it has become background noise. On June 12, 2026, it became the precise reason Roku’s stock closed 22% higher at $143.66, as Bloomberg News reported what Reuters confirmed through six independent sources: Roku is exploring strategic options including a full sale of the company.
The story moves fast. No announcement came from Roku itself. No acquirer was named. No timeline was given. What emerged is a picture of a company in early-stage exploration with at least one unnamed U.S. media company, while also examining the narrower option of a PIPE transaction, a private investment in public equity that would bring in capital without a full ownership transfer. PPC Land reported the full detail of the disclosure on Saturday morning, noting that Roku did not respond to requests for comment. The talks may not result in any transaction, and the range of options on the table suggests the company is consulting rather than closing.
What has changed is the financial foundation beneath those talks. Roku ended Q1 2026 with $2.38 billion in cash and zero long-term debt, a combination that gives any negotiation unusual structural weight on the seller’s side. Platform revenue reached $1.13 billion in the quarter, up from a year earlier, with advertising accounting for $613 million of that figure, a 27% year-on-year increase. PPC Land’s Q1 2026 earnings coverage placed ad spending routed through DSPs up 40% year on year, a rate of growth that sharply exceeds what the broader programmatic market has reported. Adjusted EBITDA for the quarter was $148 million, up 165% year on year. Total net revenue came in at $1.25 billion. The company is no longer the speculative growth story it was in 2021; it is a platform generating real margins.
Roku’s profitability milestone, first established in Q4 2025 when free cash flow hit a record $484 million, preceded this moment by several months. So did a structural disclosure that matters greatly to any M&A analysis: in April 2026, as PPC Land detailed, the company separated advertising and subscriptions revenue for the first time, breaking up the previously combined platform revenue line. Subscriptions contributed $519 million in Q1 2026, up 30% year on year. That clean separation gives any acquirer a clearer picture of two distinct revenue engines with different growth rates and different competitive dynamics.
The question now being asked by every analyst who covers streaming media is: who buys this, and why? The what is fairly clear. Roku is the largest streaming platform in the United States by household penetration. Its operating system runs on Roku-branded devices and on third-party televisions built by OEM partners. That OS position gives Roku ACR data, meaning automatic content recognition data, across a vast installed base regardless of which streaming service a viewer is using. That data asset is what a media buyer or tech platform would actually be acquiring, alongside the advertising infrastructure, the content distribution agreements, and the direct billing relationships Roku holds with millions of households paying for Premium Subscriptions.
A media company acquiring Roku would gain a distribution layer it does not currently control. A tech platform acquiring it would gain a TV operating system at scale. Streaming analyst Dan Rayburn, citing the LinkedIn post he published on June 12, noted that sale rumors surface for Roku periodically, adding that what is different here is the specificity of the Reuters sourcing network.
LinkedIn’s B2B creator infrastructure goes live
While Roku’s potential exit dominated Friday’s conversation, a different kind of structural shift arrived at LinkedIn earlier in the week with considerably less drama. On June 10, 2026, LinkedIn announced two products: Creator Marketplace and BrandWorks, both built directly into Campaign Manager. Together they represent LinkedIn’s first purpose-built infrastructure for connecting B2B advertisers with professional creators at scale inside its own platform.
Digiday, which published its own coverage on June 10, framed the launch as LinkedIn’s attempt to own B2B creator discovery before other platforms encroach on its professional network positioning. The piece noted that B2B creator discovery has stayed fragmented because LinkedIn never had an API plugged into creator marketing software, a gap this launch is explicitly designed to close.
AdExchanger’s June 11 daily roundup added a forward-looking layer, citing Business Insider reporting that LinkedIn is planning a broader slate of influencer tools for its 2027 fiscal year, which begins in July 2026. These include a dealmaking marketplace, a subscription feature, and a direct-purchase mechanism for creators to sell advisory sessions. The implication is that this week’s launches are phase one of a longer build.
Creator Marketplace is a discovery and partnership tool that sits inside Campaign Manager, LinkedIn’s advertising interface. Marketers search for creators by topic and content expertise, evaluate their audience composition and past content performance, and identify organic posts that already mention the brand. Those posts can then be amplified through Thought Leader Ads, a LinkedIn format that promotes organic content from any LinkedIn member into targeted paid feeds. The mechanic separates the creation phase from the distribution phase: a brand can find a creator who has already talked about a relevant product and put money behind that existing post rather than commissioning new content from scratch.
BrandWorks is positioned differently. Rather than a self-service discovery tool, it is a managed service where LinkedIn’s own creative team develops the campaign strategy, works through the creative approach, and handles creator coordination on behalf of the brand. The target is B2B organizations that want professional creator marketing but do not have the internal capacity to run it. LinkedIn’s own 2026 Global B2B Marketing Outlook, drawn from YouGov research conducted between January 14 and February 5, 2026 among 1,299 B2B marketers across the USA, UK, France, Germany, and India, found that 77% of B2B buyers need to trust and know a brand before engaging with it. That figure is the commercial argument for why creator marketing on a professional network carries different weight than creator marketing on a consumer platform.
The creator side of the product is opt-in. Creators who choose to participate can curate a portfolio of selected content pieces for brands to review, set a preferred contact email, include management contacts, and evaluate partnership requests on their own terms. Nothing is automatic. That architecture reflects a lesson other platforms have learned the hard way: creator marketplaces fail if creators feel their information is being harvested without consent.
The timing connects to a broader milestone that PPC Land reported on June 12. LinkedIn Brazil reached 100 million members on June 12, placing it third globally behind only the United States and India. To mark the milestone, LinkedIn opened 100 LinkedIn Learning courses for free for 100 days. The geography matters for B2B advertising: Brazil is a large and growing business market where professional social media engagement is high, and 100 million opted-in members on a professional network represents significant scale for B2B campaigns.
The same week, PPC Land also covered LinkedIn’s own labor market data, which shows EU hiring 26% below 2019 levels while AI-related roles across Europe have surpassed 351,000 since 2023. LinkedIn’s analysis attributes the hiring decline to macroeconomic pressure rather than AI displacement, a distinction that matters for how B2B brands read the professional audience they are trying to reach on the platform.
Google’s Data Manager API v1.7 closes the GMP signal gap
The most technically significant development of the week may also be the least visible to the broader marketing press. On June 12, PPC Land reported on Google Data Manager API version 1.7, published on May 28, 2026, which extends offline conversion event ingestion to the full Google Marketing Platform stack for the first time.
Until v1.7, the Data Manager API could route offline conversion events to Google Ads and Google Analytics. With this release, it adds Campaign Manager 360, Search Ads 360, and Display and Video 360. The mechanism uses the existing IngestEventsRequest method: the same request structure now routes to GMP destinations through the API’s fan-out logic, which can deliver a single request to multiple downstream platforms simultaneously. Developers do not need a separate endpoint. A new FLOODLIGHT_CONFIG account type has been added to the AccountType enum, giving advertisers a dedicated destination type for Floodlight-based measurement that accurately reflects how CM360, SA360, and DV360 share data in production environments.
Search Engine Land, which covered the same release, noted that Google is positioning the Data Manager API as the central hub for conversion and audience data, giving advertisers a unified way to manage measurement and Customer Match across its ad platforms. The practical implication is that organizations running campaigns across the full GMP stack can now send offline conversion signals into every part of that stack through a single API integration rather than managing separate ingestion workflows for each platform.
The IP address addition is the second major change in v1.7. A new composite_data field on the AudienceMember resource accepts a CompositeData object, which allows an advertiser to bundle multiple identifier types in a single audience member record. One of those identifier types is IP address. Google states in its release notes that, beginning in Q3 2026, including IP addresses with corresponding observation timestamps will improve Customer Match rates, increasing audience reach and match accuracy. The IP address signals are ingested as hashed inputs and used to strengthen the probabilistic matching that underpins Customer Match rather than being stored as raw personally identifiable data.
Four new fields also appear in the AdIdentifiers resource: dclid, impression_id, match_id, and encrypted_user_ids. The dclid field accepts a DoubleClick Click ID, which is the identifier that Campaign Manager 360 appends to non-YouTube click events across the Google Marketing Platform. Its absence from earlier API versions had created an awkward gap: without a dedicated dclid field, there was no clean way to associate an offline conversion event with a display or video click tracked through CM360. The impression_id field extends that logic to view-through attribution, allowing an offline transaction to be matched to an ad impression rather than requiring a click event.
A conversion_count field added to the Event resource enables tracking of conversion quantities, which is relevant for retail implementations where a single purchase event involves multiple units or for lead-generation programs where multiple interactions are tracked against a single campaign exposure.
This v1.7 update arrives at a specific moment in the Data Manager API’s lifecycle. Earlier in 2026, Google began pushing developers to migrate offline conversion imports away from the Google Ads API and toward the Data Manager API. Search Engine Roundtable reported in March that Customer Match data uploads using the Google Ads API were set to fail after April 1, 2026 for accounts that had not migrated. The v1.7 release, arriving after that migration deadline, extends the Data Manager API’s footprint considerably, making it harder for GMP-level operators to justify maintaining separate ingestion workflows.
The same data infrastructure thread connects to the Google Analytics Source Group field that PPC Land covered on June 12, 2026. Google Analytics launched Source Group on June 11, standardizing messy source values from Facebook, TikTok, and AI assistants into clean cross-channel reporting dimensions. The API expansion and the analytics standardization are parallel moves in the same direction: getting cleaner, more complete signal data into Google’s measurement infrastructure.
The agentic layer: four launches in one day
June 11, 2026 produced one of the more concentrated sets of simultaneous ad technology product launches this year. PPC Land’s synthesis framed the day’s announcements as a collective contest over which companies will control the infrastructure layer through which AI agents transact advertising, rather than treating each launch in isolation.
Mediaocean released NIVO AI, an advertising intelligence layer built on top of the Innovid ad serving platform that Mediaocean acquired for $500 million in early 2025. The combined stack processes more than $200 billion in annualized advertising spend and its Prisma media management platform is used by more than 100,000 people. NIVO is built from twelve specialized agents organized into four functional categories. Two creative agents cover asset generation and predictive performance scoring. Four delivery agents handle campaign trafficking, QA validation, decisioning, and taxonomy standardization. Two measurement agents answer performance questions through natural language queries and surface creative insights. Two additional agents address workflow optimization and custom functions. The claimed result from pilot testing is a 90% reduction in campaign setup time, though the company has not published the methodology behind that figure.
Magnite launched Orchestration, a coordination layer that sits between buyer-side AI agents and premium publisher supply, with dentsu and DIRECTV Advertising as the first beta partners. The product is positioned differently from NIVO: where Mediaocean is building intelligence for the buy side, Magnite is building the routing infrastructure that lets autonomous buyer agents interact with publisher inventory without requiring human mediation of each transaction. The distinction matters because ad tech’s conventional structure requires human-managed IO and trafficking steps that automated agents cannot currently complete without custom workarounds.
Teads launched EngageOS, a publisher feed operating system that collapses editorial and advertising decisioning into a single engine, with Magnite as the launch partner for programmatic demand. Teads frames EngageOS as a response to the specific threat facing open-web publishers: AI traffic is eroding organic reach, reducing the page views that underpin advertising revenue. By merging the content relevance layer and the ad auction layer, EngageOS is designed to let publishers monetize engagement signals that the current IAB-standard auction framework cannot capture.
Walmart Connect announced its partnership with Google Display and Video 360 to activate shopper audiences on YouTube with closed-loop retail sales measurement. The mechanism gives YouTube advertisers access to Walmart’s first-party purchase data for audience targeting while providing measurement that traces ad exposures to confirmed retail transactions. Walmart’s data asset includes purchase history across one of the largest retail networks in the United States, which gives audience segments built on that data a degree of behavioral specificity that standard third-party segments cannot match.
Sitting above all four announcements is Similarweb’s June 2026 AI traffic data, which PPC Land covered on June 11. ChatGPT’s share of global AI platform traffic fell to 52.7%, down from 76.4% twelve months earlier, a 23.7-percentage-point decline in one year. Gemini captured 27.3% of AI traffic. Claude grew from 1.6% to 8.9%, roughly tripling its share in twelve months. The fragmentation of AI platform traffic is directly relevant to the advertising infrastructure race: NIVO, Magnite Orchestration, Teads EngageOS, and Walmart Connect on DV360 are each positioning for a market where advertising spend is increasingly likely to flow through AI interfaces rather than traditional browsers and apps.
MiQ’s Sigma AI platform expansion, announced on June 11 to mark the platform’s one-year anniversary, added a Planning Agent, a Total Measurement module, and expanded the underlying data infrastructure to 2.5 petabytes of daily data, with activation capabilities across 16 environments. MiQ is a programmatic managed services company rather than a DSP or SSP, and Sigma’s expansion illustrates a parallel track: rather than building new buying or selling infrastructure, some operators are building AI intelligence layers on top of existing programmatic pipelines. Ogury’s SONA platform, announced the same day, takes a similar approach, using agentic AI and what it calls Persona Intelligence to convert a campaign brief into a structured, persona-based media strategy deployable across CTV, desktop, and mobile without requiring manual planning steps.
The net effect of these launches is a picture of the programmatic stack being rebuilt from both ends simultaneously. On the buy side, NIVO and SONA are introducing AI agents into campaign setup and planning. On the sell side, Magnite Orchestration and Teads EngageOS are building new infrastructure for how supply gets matched and allocated. In the middle, Walmart’s retail data integration into DV360 is an example of first-party data asserting itself as a structurally superior targeting input, an outcome that disadvantages the third-party data market further.
Viant contributed a separate data transparency move this week, launching SupplyIQ, a free publisher-facing portal that shows exactly how Viant’s DSP scores, bids on, and values their CTV and programmatic inventory. The product is notable because it inverts the normal opacity of a DSP’s bidder logic, giving publishers visibility into demand signals that have historically been invisible from the supply side.
Also noted
June 12, 2026: Spotify redefined podcast plays as 30-second listens, aligning with the Audio Measurement Partnership’s new standard, and launched five new creator analytics tools including Audience Segments and Episode Trends. Spotify rewrites podcast plays and launches 5 new creator analytics tools
June 11, 2026: CheckedUp earned MRC accreditation for TV Video Ad Plays and Explorer Waiting Room TV Engagement, the first specialty waiting room TV to receive this certification and a milestone for pharmaceutical point-of-care advertising measurement. CheckedUp wins MRC accreditation
June 11, 2026: Meta blocked agentic AI company Manus from its internal systems and ordered staff to migrate off the platform as China’s NDRC moved to unwind the $2 billion acquisition deal that had brought Manus into Meta’s AI stack. Meta severs Manus data access as Beijing forces $2bn deal apart
June 12, 2026: Google expanded PayPal Hyperwallet payout options to 21 Latin American territories for AdSense, AdMob, and Ad Manager publishers, adding PayPal and cash pickup as payment methods. Google brings PayPal Hyperwallet payouts to 21 LATAM territories
June 11, 2026: Reddit rolled out video replies in comments for all users across iOS, Android, and desktop, enabling video uploads in eligible SFW public communities and adding a new engagement surface that has direct implications for social advertising context. Reddit launches video in comments