1 Magnite Q1 2026: CTV crosses 51% of revenue for the first time
Magnite reported Q1 2026 yesterday. Contribution ex-TAC: $160.9M (+10% YoY); CTV ex-TAC: $82.3M (+30%); CTV is now 51% of total contribution ex-TAC, up from 43% a year ago. Adjusted EBITDA grew 16% to $42.9M (27% margin). Net income swung from a $9.6M loss to $4.4M profit. EPS $0.13 beat the $0.11 consensus. Guidance for the year was raised.
The Magnite story used to be "the SSP that hopes CTV grows." It just became "the SSP whose business is mostly CTV." That's a structural shift — for Magnite, for every CTV holdco's ad pitch, and for the buy-side tools that need to upgrade from display-era reporting. Pair this with the Amazon–Roku reach play (#8) and you get the same throughline from the supply side: open-web display is the side hustle now.
2 Snap kills its $400M Perplexity deal — and warns Iran is costing $20–25M a month
Inside yesterday's Q1 print, Snap disclosed two material updates. First, the $400M Perplexity integration "amicably ended" in Q1 — the chatbot inside Snapchat was tested but never broadly rolled out, and Snap's full-year guidance now "assumes no contribution" from the partnership (vs. a previously expected ~$324M in 2026). Second, the Middle East conflict is costing $20–25M per month in ad revenue, with annualized exposure that could exceed $200M if it continues. Snap also confirmed a 16% workforce cut and is doubling down on AR Specs.
The Perplexity collapse is the first big AI distribution deal of this cycle to die in public, and it should reset how every platform writes its next AI partnership. The lesson isn't "AI search isn't real" — it's that the integration shape (where users see the AI, who controls the surface, who keeps the data) is the actual product. If Snap and Perplexity couldn't agree on how to roll it out, expect Meta, TikTok, and the news publishers to be much sharper about that contract before signing one.
3 Rakuten launches Mirai — affiliate marketing's first conversational AI agent
Rakuten Advertising shipped Mirai yesterday, branding it "affiliate marketing's first advanced AI optimization agent." Advertisers can now build and manage affiliate offers in natural conversation: Mirai analyzes the business objective, suggests commission structures, and autonomously generates the offer code in real time, turning multi-step affiliate ops into a chat. Rakuten says additional capabilities — partner identification, recruitment, performance optimization, and reporting — will roll in over the next few months.
Affiliate has historically been the most manual lever in performance marketing — offers, codes, payouts, all configured by hand. Mirai is the first agent product that targets the workflow itself, not just the reporting on top of it. Expect Rakuten's competitors (impact.com, Awin, CJ, Partnerize) to ship their own conversational agents within the quarter — and watch Shopify, which sits next to all of them, to decide whether it builds a native one or partners.
4 Meta builds "Hatch" agent + an in-feed AI shopping tool for Instagram — Q4 launch
Per The Information yesterday, Meta is building Hatch, a consumer AI agent that can complete tasks autonomously — a direct response to OpenAI's Operator and Anthropic's Claude. Today's Hatch runs on Anthropic's Claude Opus 4.6 and Sonnet 4.6, but Meta plans to migrate it to its own Muse Spark model before launch. In parallel, Meta is shipping an AI shopping assistant inside Instagram — product info, AI Q&A, and checkout, all without leaving Reels or feed. Internal target: Q4 2026.
Two big tells here. First, Meta is willing to ship on Anthropic's stack publicly — an admission that the LLM race is now "use the best model, ship the product, swap later." Second, the in-feed shopping tool is a direct shot at TikTok Shop, whose 2026 US GMV forecast hit $23.4B (+48% YoY). If Meta can make Reels checkout one tap, the next decade of social commerce gets re-decided in the next two quarters.
5 PubMatic Q1 prelim beats — but the CGO and Americas CRO both leave the same week
PubMatic preliminary Q1: revenue ~$62.4M (vs. $58–60M guided); adjusted EBITDA ~$2.5M (vs. -$0.5M to $1M guided). CEO Rajeev Goel: "Our AI solutions are seeing the fastest adoption of any product launch in our history." But on the same release, the company announced Chief Growth Officer Paulina Klimenko (15-year tenure) is stepping down for health reasons and Americas CRO Kyle Dozeman is leaving to pursue an entrepreneurship opportunity. PubMatic retained Heidrick & Struggles to find a global CRO and consolidate revenue functions under that role. Full Q1 release lands today.
Read this as a leadership reorg disguised as departures. The "global CRO" framing tells you PubMatic is collapsing the buy-side and publisher orgs into one go-to-market — which is also what AdCP and the IAB Tech Lab framework (#6) are pushing the rest of the industry toward. The fastest-adopted products in PubMatic's history (Activate, the curated-marketplace stack) are agentic. Expect the new CRO hire to be an AI-native exec.
6 AdCP vs IAB Tech Lab: agentic ad standards face off in Palm Springs this week
The Digiday Programmatic Marketing Summit kicked off in Palm Springs May 6–8, and the headline isn't a vendor announcement — it's a standards war. Ad Context Protocol (AdCP), founded by Yahoo, Optable, PubMatic, Scope3, Swivel, and Triton Digital, is built on Anthropic's Model Context Protocol and aims to standardize agent-to-agent ad transactions. IAB Tech Lab launched a competing Agentic RTB Framework (ARTF) and is "agentifying" existing programmatic standards with new SDKs.
This is the OpenRTB moment for AI agents. One side is open-source-first, MCP-native, governed by a working group. The other is the incumbent body that wrote the rules of the last 10 years of programmatic. Whoever wins decides whether your agent talks to my publisher's agent through a vendor-neutral protocol or through an IAB-defined wrapper on top of OpenRTB. For most marketers, the practical signal is simple: your DSP is going to plumb both, and the one that gets a dominant share of agent traffic in 12–18 months sets the next decade of ad ops.
7 TikTok + Vistar Media: vertical TikTok ads land on horizontal billboards at scale
TikTok expanded its Out of Phone program with Vistar Media on Tuesday, letting brands "reformat and reimagine" their vertical TikTok ads for digital out-of-home. The challenge is structural — TikTok ads are vertical-first, billboards are horizontal — so Vistar's Creative Studio handles the creative refit (static screens, side panels, IRL-context placements) and serves the work across DOOH inventory at scale. The two have partnered since late 2024; this expansion turns it into a default OOH path for any brand running TikTok creative.
This is TikTok solving a real problem most CMOs haven't solved: cross-screen creative mismatch. With OOH ad spend climbing while social CPMs flatten, brands want to stretch one creative budget across screens. The unspoken upside for TikTok: every billboard running a TikTok-native ad is a recruitment tool for the platform's brand-safe story among CFOs who still don't trust short-form social.
8 Amazon + Roku: 80M-household CTV partnership goes live with first results
The Amazon–Roku CTV interoperability deal is now live, with logged-in reach across ~80M US streaming households (~80% of total). Early test results: advertisers reached 40% more unique viewers on the same budget, and frequency dropped by ~30%. The deal is the largest scale partnership ever between the #1 retail-media-meets-CTV stack (Amazon) and the largest CTV-OS by US households (Roku).
The Achilles heel of CTV has always been fragmentation — you reach a household 14 times across 6 apps and call it 14 unique viewers. Amazon and Roku just chose to fix that problem on the supply side, sharing logged-in graphs across both. Expect the rest of the CTV ecosystem (Disney, NBCU, Netflix, Tubi, Pluto) to feel the pressure to plug into Amazon's DSP — which, combined with Magnite's Q1 (#1), is why "open-web display" is no longer where the smart media-buying spend goes.
9 Disney is moving Hulu's ad targeting onto Disney+ — full app fold-in by year-end
Per Digiday, Disney plans to extend Hulu's ad-targeting capabilities into the Disney+ ad-supported tier as part of the full Hulu fold-in by end of 2026. Hulu has historically had the more advertiser-friendly stack (audience graphs, dynamic ad insertion, third-party measurement); Disney+'s ad tier launched leaner. The combined ad-supported app's annual revenue is projected past $5B. The Nintendo Switch Hulu app already shut down in February as the first migration step.
Disney's Upfront on May 12 is going to spend most of its stage on this. The combined Disney+/Hulu app is essentially Disney's answer to the Roku–Amazon and YouTube Brandcast scale plays — one targeting layer, one measurement story, one ad-supported audience that rivals YouTube's logged-in reach in the US. The thing to watch: whether Hulu's third-party measurement vendors (iSpot, VideoAmp) keep their seat at the table once the apps merge, or whether Disney consolidates measurement under its own AI-on-TV pitch from CES.
10 WARC's Future of Measurement 2026: AI is moving measurement upstream
WARC released The Future of Measurement 2026 this week. Three big calls. (1) Outcomes are replacing attribution: post-hoc reporting on click paths is less and less relevant when the platform decides who sees what. (2) AI is moving measurement upstream — from a reporting function to a decision-making system that powers dynamic planning and live optimization. (3) Creative intelligence is now its own measurement category — AI scoring creative before launch, and tying creative variables to commercial outcomes at scale.
This is the fastest-changing layer in marketing right now and the one most teams are slowest to adopt. The barriers WARC names are real: data quality, budget, and creative-media silos that block predictive modeling. But the direction is clear — in two years, the marketers who can't answer "what will this creative do before we ship it?" will be operating on lagging data while their competitors are running ahead with creative intelligence baked into planning.
The throughline
Streaming got the spend. Agents got the stack. Magnite's CTV crossing 51% and the Amazon–Roku partnership are the supply-side proof that CTV is now where the dollars sit. Snap killing the Perplexity deal, Rakuten shipping Mirai, Meta wiring up Hatch, PubMatic reorging around a global CRO, and AdCP vs IAB Tech Lab fighting for the agentic ad standard — all five say the same thing about the demand side: the buying happens through agents, the question is whose.
The ad stack of 2024 was: human in DSP → auction → CTV/social/open web. The ad stack of late 2026 looks like: agent in DSP → agent-to-agent protocol → CTV-first inventory with AI measurement. WARC is telling us measurement is moving upstream into that flow. TikTok is telling us creative has to ride across screens to keep up. Disney is telling us the next bundle is ad-targeting + scale. The marketers winning the second half of 2026 will be the ones who pick their agent, pick their CTV partner, and pick their measurement story before the rest of the org even realizes the stack changed.