OpenAI confirmed plans to check commercials inside ChatGPT throughout the week of January 20, marking a shift from CEO Sam Altman’s earlier characterization of adverts as a “final resort” for the platform’s enterprise mannequin. The announcement, made on a Friday earlier than a vacation weekend, left advertisers with restricted details about implementation particulars, pricing buildings, or timeline specifics.
Michelle Merklin, VP of search at impartial efficiency company Tinuiti, instructed AdExchanger the corporate has “extra questions than solutions at this level.” OpenAI representatives knowledgeable the company there’s presently no wait checklist for advertisers, however the firm has not offered direct entry to its newly shaped promoting group. Merklin famous OpenAI has been “very noncommittal about all the things” relating to the adverts rollout.
Within days, The Info reported OpenAI established a baseline CPM of roughly $60 for ChatGPT promoting placements. The pricing rivals stock prices for the Las Vegas Sphere and NFL sport broadcasts. Meta’s promoting CPMs sometimes vary from $10 to $20, whereas YouTube campaigns typically attain the excessive finish round $60. OpenAI’s pricing construction stands out as a result of the platform lacks attribution techniques or pixel networks to trace conversions that advertisers use to measure marketing campaign effectiveness.
The price-per-impression mannequin introduced by OpenAI prices advertisers primarily based on impressions or views quite than clicks. This cost construction goals to push advertisers towards significant natural engagement over direct hyperlink efficiency. Nevertheless, early advertiser requirements prompt preliminary marketing campaign spending would stay under $1 million throughout the launch section.
OpenAI CFO Sarah Friar defended the timing throughout a Davos look on January 22, stating “early is a bizarre phrase” when discussing promoting fashions. Friar emphasised that profitable advert fashions require scale, including “when you will have 800 million weekly energetic customers, you are actually far past most of the corporations who began in that mannequin.” Demis Hassabis, CEO of Alphabet-owned DeepMind, questioned whether or not OpenAI moved too shortly, telling journalist Alex Heath it was “attention-grabbing they’ve gone for that so early” and suggesting “perhaps they really feel they should make extra income.”
The contextual focusing on capabilities distinguish ChatGPT from conventional promoting platforms. Yang Han, CTO of DSP StackAdapt, defined the platform can goal customers at exact moments primarily based on dialog content material. ChatGPT maintains reminiscence of person interactions, doubtlessly understanding when somebody final ran a race and associated queries they’ve made. This permits product suggestions tailor-made to particular person shopper wants past fundamental demographic focusing on.
The promoting implementation faces technical challenges. Turning into absolutely contextually related requires a broad advertiser base akin to Meta’s tens of millions of energetic advertisers. Till ChatGPT achieves related scale, the platform dangers exhibiting mediocre commercials that would alienate customers. Han characterised this as “a little bit of a chicken-and-egg drawback” the place restricted advertiser participation might deter new advertisers from becoming a member of.
Tyler Holoubek, assistant director of funding at impartial company KSM Media, noticed various shopper reactions to ChatGPT promoting alternatives. Some purchasers approached testing with enthusiasm, in search of early learnings. Others in regulated industries together with prescribed drugs and monetary providers confirmed better hesitation concerning the untested channel.
The promoting improvement arrives as OpenAI explores new merchandise distinct from conventional social codecs. An organization spokesperson clarified that separate AI techniques will consider whether or not conversations have industrial intent earlier than surfacing related commercials in ChatGPT responses. OpenAI maintains that adverts will probably seem solely after customers present clear curiosity or buy intent, guaranteeing relevance and sustaining person belief.
Meta announced on January 21 the worldwide growth of promoting on Threads to all customers, finishing a testing program that started in January 2025. The rollout, beginning the week of January 26, extends promoting availability to all markets the place Threads operates, reaching the platform’s greater than 400 million month-to-month energetic customers.
The growth represents the ultimate section of Meta’s monetization technique for the text-based social platform. Meta launched Threads in July 2023 with out promoting capabilities, prioritizing person progress earlier than introducing industrial parts. The corporate initiated testing with restricted advertisers on January 24, 2025, when the platform had accrued 300 million month-to-month energetic customers. Testing expanded to all eligible advertisers on April 23, 2025, although supply remained restricted to pick markets.
Ads on Threads make the most of Meta’s AI-powered promoting system, providing the identical personalization stage customers expertise on Fb and Instagram. Out there codecs embody picture adverts, video adverts, carousel adverts, Benefit+ catalog adverts, and app adverts. Technical specs assist 4:5 facet ratio rendering, with belongings taller than 4:5 cropped and vertically centered for format consistency.
The Threads feed placement prompts by default for brand spanking new campaigns utilizing Benefit+ or Guide Placements in Meta Adverts Supervisor. Advertisers in search of to exclude Threads from their campaigns should manually choose out by placement settings. Meta emphasised that advert supply will initially stay low as the corporate reaches world person availability by a gradual rollout throughout the coming months.
Technical capabilities developed considerably all through 2025. The platform initially supported solely single picture and video commercials when testing started in January. Meta launched carousel advert assist on October 6, 2025, adopted by Benefit+ catalog adverts on October 28. The catalog performance presently helps pictures and picture carousels, whereas slideshow, video, and entry playing cards stay unsupported. Side ratio specs expanded on October 28 to accommodate the 4:5 rendering format.
The timing for the growth demonstrates Meta’s measured strategy to platform monetization. Meta said throughout the preliminary testing section that it might monitor the restricted take a look at carefully earlier than broader implementation. The corporate emphasised that commercials assist folks uncover companies whereas enabling advertisers to increase present Meta campaigns to the dialog platform with out bespoke artistic or extra resourcing necessities.
Microsoft Prebid caching termination threatens publishers
Microsoft introduced it’s going to stop caching Prebid video files on April 30, shortly after the corporate shuts down the Xandr DSP. The change impacts publishers who depend on Prebid and Google Advert Supervisor for video commercial supply, doubtlessly creating speedy and expensive disruptions.
When advertisers win video advert impressions in Prebid auctions and publishers use Google Advert Supervisor, Google runs a secondary public sale evaluating the Prebid bid towards bids from its personal demand stack. If the Prebid bid wins, GAM requires the video artistic to be served. In the course of the transient second earlier than GAM makes the ultimate public sale choice, an middleman should host the video advert file from Prebid. Microsoft Promoting has offered this storage service since buying Xandr in 2021, following AppNexus and AT&T’s Xandr.
Patrick McCann, SVP of analysis at Raptive and Prebid board member, estimated month-to-month caching prices between $50,000 and $250,000. These figures mirror present utilization ranges, however prices might attain seven figures month-to-month inside a couple of years as Prebid adoption grows and video advert merchandise develop. Round 4 in 5 US websites that enable advert serving use GAM, which means if 60 % of Prebid video advert protection in that provide chain is affected, an unlimited variety of precise adverts might fail to fireside on pages come February.
Microsoft already prolonged the deadline from January 31 to April 30, offering extra transition time. Nevertheless, McCann famous that advertisers and publishers typically look forward to absolute deadlines, doubtlessly leaving many unprepared. Publishers who fail to implement different options will see video adverts end in VAST errors, which means no video will seem.
Technical workarounds exist by a number of SSPs and video advert networks. Magnite, PubMatic, Mediavine, and JW Participant already supply Prebid video caching as a paid choice for purchasers. Publishers should consider these alternate options and implement options earlier than the April 30 deadline to take care of video commercial supply.
The caching termination aligns with Microsoft’s broader strategic shifts in promoting know-how. The corporate introduced in Could 2025 it might discontinue Microsoft Make investments (previously Xandr) with an efficient date of February 28, 2026. Company Vice President Kya Sainsbury-Carter cited incompatibility between conventional demand-side platform fashions and Microsoft’s imaginative and prescient for conversational, personalised, and agentic promoting futures.
Prebid launches publisher-side AI agent framework
Prebid announced on January 29 it’s taking possession of code developed utilizing Advert Context Protocol that may energy publisher-side AI brokers. The group that established open-source requirements for header bidding software program now offers publishers with instruments to construct autonomous AI brokers for promoting advert stock.
The providing, named the Prebid Gross sales Agent, is obtainable to all publishers no matter Prebid membership standing. Publishers can obtain the software program from GitHub and supply suggestions on its improvement. Rising frameworks like AdCP are establishing groundwork for AI-to-AI transactions, however anticipating publishers to construct and preserve their very own autonomous brokers represents a stretch with out standardized instruments.
The event arrives because the promoting trade races towards a future the place AI brokers negotiate programmatic offers autonomously. Prebid’s involvement ensures publishers have entry to standardized instruments quite than requiring particular person improvement efforts for agentic capabilities. The open-source strategy mirrors Prebid’s profitable header bidding standardization, which diminished writer reliance on proprietary options.
AdCP framework faces implementation skepticism
Dr. Aaron Andalman, Chief Science Officer at Cognitiv, printed analysis on January 26 questioning whether or not Advert Context Protocol delivers significant enhancements past workflow standardization. Since its launch a number of months in the past, AdCP has been positioned as a foundational step towards agentic promoting the place AI techniques plan, execute, and optimize media buys with out human intervention.
AdCP builds on Mannequin Context Protocol, an open customary launched by Anthropic and broadly adopted throughout the trade. MCP addresses limitations of huge language fashions in taking dependable actions inside software program. Andalman argues that promoting’s agentic future represents a compelling imaginative and prescient counting on capabilities that don’t exist but. AdCP might finally make shopping for simpler, but it surely won’t make shopping for higher by itself.
The protocol standardizes how AI techniques work together with promoting platforms and instruments. Nevertheless, standardized workflows don’t mechanically translate to improved media outcomes. Advertisers feeling strain to implement AdCP ought to consider whether or not the framework addresses precise enterprise wants or represents untimely adoption of incomplete know-how.
Meta’s newest Andromeda replace, the fastest and most advanced iteration of Meta’s ad-retrieval system, has modified the rhythm of Meta Adverts supply. Monica Shukla, VP Biddable COE at an unnamed company, wrote on January 28 that supply cycles are sooner, artistic will get picked up and exhausted with exceptional pace, and optimization feels extra dynamic than ever.
The elevated velocity has launched confusion amongst advertisers. Many assume that as a result of Andromeda adapts shortly, they’ll provide large volumes of artistic, depend on automation, and let the system deal with optimization. Others count on the algorithm will floor insights, repair sign points, or compensate for funds constraints. Shukla emphasised that Andromeda is highly effective however removed from omniscient, counting on sign power, strategic readability, and inventive asset high quality.
One persistent misunderstanding stems from the idea that importing 20 or extra creatives per advert set is universally useful. When artistic provide dramatically outweighs accessible funds, studying slows, supply fragments, and the system struggles to tell apart sign from noise. Many smaller or rising manufacturers fall into this lure, assuming they need to hold tempo with world advertisers’ artistic velocity regardless of inadequate budgets to assist significant studying throughout quite a few belongings.
A simpler strategy aligns artistic quantity with funds actuality and marketing campaign targets. Higher-funnel campaigns can assist a broader vary of ideas, however solely at budgets that enable every asset to realize traction. Manufacturers attaining the best carry from Andromeda have adjusted not by outsourcing judgment to the algorithm however by refining their artistic technique. Match artistic quantity to funds and marketing campaign targets quite than defaulting to most artistic uploads.
Federal immigration enforcement seeks advert tech surveillance
US Immigration and Customs Enforcement published a request on January 23 at 1:00 PM Japanese Commonplace Time in search of info from promoting know-how corporations about location information and surveillance capabilities. The request for info, posted on SAM.gov, asks corporations to exhibit how their platforms monitor inhabitants actions by cell functions, promoting platforms, and information dealer networks.
ICE seeks operational capabilities for location monitoring, geofencing, and behavioral evaluation instruments utilized in industrial promoting. The company specified it really works with “growing volumes of legal, civil, and regulatory, administrative documentation from quite a few inner and exterior sources.” Firms should reply by February 2, 2026 at 10:00 AM Japanese Commonplace Time, with the request turning into inactive February 3.
The submitting states it represents market analysis quite than a right away procurement, emphasizing “this isn’t a solicitation announcement” and “doesn’t commit the Authorities to contract for any provide or service.” Nevertheless, the request indicators federal regulation enforcement analyzing industrial promoting know-how for investigative deployment. Firms submit responses by a Microsoft Type submission portal with written documentation restricted to 10 pages.
ICE indicated it’s going to choose a number of respondents to current reside demonstrations of operational capabilities, platforms, and information providers. The submitting doesn’t specify demonstration places. The industrial surveillance instruments ICE seeks function globally by cell functions, promoting platforms, and information dealer networks assembling complete inhabitants monitoring capabilities.
The timing coincided with November 2024 investigative reporting exposing how information brokers promote location coordinates monitoring US army and intelligence personnel actions, demonstrating industrial surveillance capabilities related to authorities pursuits. California authorized rules for an information dealer deletion platform in November 2025, taking impact January 2026, whereas Texas filed a lawsuit towards Hisense over good TV surveillance affecting 1.27 million residents in November 2024.
The request raises questions on promoting know-how’s position in authorities surveillance packages. Location information collected for promoting focusing on functions will be repurposed for monitoring particular people or teams. The promoting trade has confronted growing scrutiny over information assortment practices and potential misuse of shopper info.
Timeline
- January 21, 2026: Meta announces world growth of promoting on Threads to all customers beginning the week of January 26, finishing yearlong testing program
- January 22, 2026: OpenAI confirms plans to check commercials inside ChatGPT, revealing restricted particulars about implementation timeline or pricing
- January 22, 2026: OpenAI CFO Sarah Friar defends promoting timing at Davos, citing 800 million weekly energetic customers as justification for scale
- January 23, 2026: ICE publishes request for info on SAM.gov in search of promoting know-how location monitoring and surveillance capabilities
- January 26, 2026: Microsoft confirms it’s going to cease caching Prebid video recordsdata on April 30, threatening writer video advert supply
- January 26, 2026: Meta begins rolling out commercials to all Threads customers globally throughout all markets
- January 26, 2026: Dr. Aaron Andalman publishes evaluation questioning Advert Context Protocol’s means to enhance media outcomes
- January 27, 2026: The Information reports OpenAI set baseline CPM of roughly $60 for ChatGPT promoting placements
- January 28, 2026: Monica Shukla writes about Meta’s Andromeda replace accelerating artistic exhaustion and supply cycles
- January 29, 2026: Prebid announces it’s taking possession of Advert Context Protocol code to energy publisher-side AI brokers
- January 29, 2026: AdExchanger reports on advertiser hesitation towards ChatGPT’s untested channel missing attribution techniques
- February 2, 2026: Response deadline for corporations to submit info to ICE about promoting know-how surveillance capabilities
Share this text
