Warner Bros. Discovery Picks AWS for Agentic AI Ad Stack

Warner Bros. Discovery is rebuilding its advertising technology on Amazon Web Services, replacing siloed internal workflows with autonomous AI agents that plan, place, and optimize campaigns across its U.S. linear and digital channels. The company named AWS as its preferred cloud provider for the next-generation ad stack, with unified media planning set to roll out in Q3 2026 and composable order management, pricing, and stewardship following in Q4. Read on its own, the release looks like a cloud migration. Read against the rest of the WBD release, it is a structural reset of how ads get bought inside one of the largest media companies in the world.

Warner Bros. Discovery, Inc. trades on Nasdaq as WBD and runs household-name channels including HBO, CNN, Discovery Channel, and TNT. AWS is its preferred cloud provider and supplies the foundation models, agent runtime, and storage layer underneath WBD’s new platform. The two companies positioned the work as an “agentic, AI-native decisioning” stack that handles planning, forecasting, optimization, and closed-loop measurement on its own. The phrasing makes clear that the agents, not human planners, are now the default path through the funnel.

What WBD Actually Announced

The June release covers three things at once: a public designation of AWS as WBD’s preferred cloud provider, a new architecture for the company’s advertising technology stack, and a public roadmap through the rest of 2026. WBD framed the move as a rebuild from the ground up of internal workflows that have, until now, run as separate stacks across linear and digital.

The press release describes the platform as an “agentic AI-driven approach” to how premium inventory is “planned, activated, optimized, and monetized.” Inside it sit autonomous AI agents that WBD says continuously self-optimize, learning from the outcome of each campaign to improve the next one. Buyers get to target either specific brands or audience segments across linear and digital, with WBD providing inventory allocation recommendations. The release does not detail pricing tiers or advertiser commitments, and the platform’s commercial terms remain undisclosed.

WBD described the work as a re-imagining of its own infrastructure, not a partnership resold to outside media owners. The positioning turns the announcement into a corporate IT project with ad-tech implications, rather than a third-party product WBD is wrapping its name around. The release also does not state which competitors’ stacks WBD considered or rejected, leaving the AWS choice presented as a settled fact rather than a contested decision.

The Agentic Layer That Does the Work

WBD’s new advertising technology stack runs on six named AWS services, each with a defined job inside the platform. Amazon Bedrock AgentCore is the runtime that lets WBD build, connect, and optimize the AI agents themselves, while Amazon Bedrock hosts the foundation models that those agents call on inside a closed WBD instance. Amazon SageMaker trains custom machine learning models for WBD’s exclusive use, kept inside WBD’s own data-protection and segmentation controls.

Underneath, Amazon Simple Storage Service (Amazon S3) holds the data lake in Apache Iceberg format, Amazon Elastic Container Service (Amazon ECS) powers application hosting, and Amazon Quick gives WBD’s ad sales teams a natural-language assistant to query the data and surface recommendations. Together, the six services replace what was, by WBD’s own description, a long-standing set of traditional internal workflows that had to be re-keyed and reconciled across business units. The data lives in S3, the models train in SageMaker, and the agents act through Bedrock AgentCore, all inside WBD’s tenancy. None of the layers is a third-party optimization product WBD is licensing.

The architectural choice has implications that go beyond the press release. WBD’s agents run inside WBD’s own data boundary, training on first-party advertiser and audience data with the company’s segmentation and security controls applied. The company is, in effect, becoming its own model operator through AWS, with custom training in SageMaker and custom deployment through Bedrock AgentCore. That move lines up with how AWS has been pitching Bedrock AgentCore to media buyers as the production layer for transactional AI agents. The architecture gives WBD a faster iteration loop on its own data, while leaving AWS as the underlying substrate for every model call and every storage transaction.

That substrate role is also the most consequential dependency in the deal. AWS handles the agent runtime, the foundation models, the storage, the hosting, and the internal sales assistant, so a future re-platforming of any one of those layers would now require rebuilding parts of WBD’s ad stack against a different cloud provider, not just swapping a vendor.

• Amazon Bedrock AgentCore: runtime to build, connect, and optimize WBD’s AI agents.
• Amazon Bedrock: foundation models in WBD’s closed instance.
• Amazon SageMaker: custom ML training under WBD’s own data controls.
• Amazon Simple Storage Service (Amazon S3): data lake in Apache Iceberg format.
• Amazon Elastic Container Service (Amazon ECS): application hosting.
• Amazon Quick: natural-language AI assistant for WBD’s ad sales teams.

The Rollout Calendar Through 2026

WBD started the rollout earlier this year, framing 2026 as the deployment year for the new stack. Per the release, the company began shipping three things in 2026: agentic automation for direct response and commercial workflows, advanced audience forecasting, and enhanced measurement and attribution. Those pieces are the foundation the company is building on for the two larger milestones ahead. The first half of 2026 has therefore been about plumbing; the second half is about the buyer-facing surface.

Q3 brings unified media planning, the moment when advertisers can plan and buy WBD’s linear and digital inventory through a single workflow. Q4 follows with a phased rollout of composable order management, pricing, and stewardship, the systems that determine how a campaign is priced, ordered, and reconciled once it runs. The company has not named specific advertiser launch partners or revenue targets tied to either phase.

1. Earlier in 2026: agentic automation for direct response and commercial workflows.
2. Earlier in 2026: advanced audience forecasting.
3. Earlier in 2026: enhanced measurement and attribution.
4. Q3: unified media planning across linear and digital.
5. Q4: phased rollout of composable order management, pricing, and stewardship.

Unifying Linear and Digital in One Buy

The most consequential claim in the release is that linear and digital inventory will run through the same planning and optimization workflow. For decades, the two sides of TV advertising have lived in separate operations, sold by separate teams, planned against separate metrics, and measured against separate KPIs. The WBD release is built around the premise that AI agents can unify linear and digital ad buying while keeping each channel’s character.

Dr. Nage Sethu, SVP of Technology for Converged Advertising and Linear Systems at Warner Bros. Discovery, framed the move as the “next frontier of advertising,” with convergence bringing linear and digital together on a single platform while each channel keeps its own character. His quote also names the constraint: the unified layer is, in his words, “measurable and optimizable at cloud scale with agentic, AI-native decisioning.” That phrasing makes the unification conditional on the cloud stack running beneath it, and AWS is, by WBD’s own account, that stack. The release leans on the same phrasing across its marketing copy, signalling a unified design language for the platform’s public rollout.

The implications for the buy side are concrete. An advertiser running both a national TV campaign and a companion digital video flight could, in theory, see one set of forecasts, one inventory recommendation, and one set of optimization rules applied across both. The plan replaces what WBD calls “previously siloed business units” with continuous automated campaign optimization that adjusts in flight. WBD has not yet shared metrics on lift, error rates, or pacing improvements against the legacy stack. Buy-side performance claims therefore rest on the platform’s architecture rather than disclosed benchmarks.

Building with AWS has been critical to streamlining the buyer’s experience across linear and digital, powering critical layers of our data, forecasting, and next-generation agentic advertising stack.

The quote comes from Dr. Nage Sethu, SVP of Technology for Converged Advertising and Linear Systems at Warner Bros. Discovery, in the company’s June release on its agentic AI advertising platform. WBD has not yet published an advertiser roster for the Q3 rollout.

The Wider Race for Agentic Ad Tech

WBD is not alone in this move. DecodeTV reported the announcement landed shortly after Fox launched its own agentic AI advertising platform, putting two of the largest U.S. broadcast owners through similar overhauls within weeks of each other. AWS itself is using Cannes Lions 2026, the advertising industry’s annual showcase on the French Riviera running June 22 to 26, to make the agentic ad pitch broadly across the industry. The festival agenda, laid out in Cannes Lions 2026 agenda for agentic advertising, includes a Tuesday, June 23 panel titled “How AI Agents Are Rewriting the Rules of Advertising,” with AWS general manager Samira Panah Bakhtiar and WBD’s Sethu on stage at Amazon Port.

Both networks are leaning on the same set of cloud-based AI tools to collapse the separation between linear and digital ad buying. For WBD, the cost of being later than Fox on the same approach is structural: an advertiser who can buy Fox’s agentic inventory with one workflow will not want a second workflow for WBD’s. The bigger picture, visible from the Cannes Lions agenda, is that premium video advertising is converging on a single agentic blueprint run out of the same cloud provider.

AWS has been pushing Bedrock AgentCore to media buyers as the production layer for transactional AI agents, with the Cannes Lions programming positioning Bedrock, AgentCore, and custom silicon as the building blocks for agentic advertising, hyper-personalized content, real-time advertising, and privacy-safe monetization. For media owners, that positioning turns AWS from a back-end cloud provider into a strategic advertising partner with visibility into campaign data. For independent ad-tech vendors, the same positioning narrows the field of where they can sit in a premium video campaign. For advertisers, the practical question is whether their existing measurement and verification vendors plug into the new stack, or whether the agentic layer becomes a closed loop. WBD’s release does not answer that question publicly.

The Fox parallel also sets the competitive clock. Two of the largest U.S. broadcast owners now have publicly disclosed agentic AI ad platforms, both built around cloud-native AI tooling. The shape of the U.S. premium video ad market over the next four quarters now looks like a race to define the agentic standard, with the platform that wins it setting the workflow every buyer ends up using.

Where This Leaves Buyers and Sellers

The upside for advertisers is straightforward. One planning surface for linear and digital inventory, one forecasting layer, one set of measurement rules, and one optimization loop that adjusts in flight. For brands that already buy across both sides of WBD’s business, the friction of running two parallel campaigns against two separate teams should compress. The release explicitly names interactive ad formats and more customized ad content as additional buyer-facing wins. Faster budget allocation across WBD’s portfolio of brands is the third buyer-facing claim attached to the rollout.

The cost lands in three places at once. AWS becomes the substrate for a significant slice of how U.S. premium video advertising is bought, deepening the cloud provider’s grip on the media industry’s most strategic workflow. Traditional ad-buying workflows inside WBD’s own operations face displacement as the agents take over planning, optimization, and order management, while advertisers running campaigns that depend on third-party data, custom measurement partners, or non-AWS verification tools will need to confirm those still plug into the new composable layer; WBD has not publicly staffed the transition or named compatible vendors.

For agencies and independent ad-tech vendors, the platform is a statement of where WBD believes the value sits, anchored in the media owner’s first-party data and the cloud provider’s agents. Buy-side tools that have historically sat between the two are now positioned outside that core. Bakhtiar framed the deal in the same release as the start of a deeper, longer partnership between the two companies. The two-quarter gap to unified media planning is now the calendar against which that framing will be measured.

AWS is proud to deepen our relationship with Warner Bros. Discovery at such a transformative moment in media and advertising. By combining WBD’s iconic content portfolio and rich audience signals with AWS’s cloud and agentic AI capabilities, we’re enabling a new era of intelligent, automated advertising that delivers better outcomes for brands and viewers alike.

The quote comes from Samira Panah Bakhtiar, general manager of Media, Entertainment, Games & Sports at Amazon Web Services, in the same WBD release on its agentic AI advertising platform. WBD has not yet published an advertiser roster for the Q3 rollout.

Frequently Asked Questions

When does unified media planning go live at Warner Bros. Discovery?

Unified media planning is set to roll out in Q3 2026, per the company’s June announcement. Earlier in 2026, WBD had already begun deploying agentic automation for direct response and commercial workflows, advanced audience forecasting, and enhanced measurement and attribution. Q4 2026 follows with a phased rollout of composable order management, pricing, and stewardship.

What does the agentic AI platform actually do for advertisers?

The platform runs autonomous AI agents that handle planning, dynamic forecasting, real-time optimization, and closed-loop measurement across WBD’s U.S. linear and digital channels. Advertisers can target specific brands or audience segments. WBD’s system provides inventory allocation recommendations and continuous self-optimization across campaigns. The release stresses interoperability with the wider ad-tech stack but does not name compatible vendors.

Which AWS services power the platform?

The stack rests on Amazon Bedrock AgentCore for the agent runtime, Amazon Bedrock for foundation models, Amazon SageMaker for custom machine learning, Amazon Simple Storage Service (Amazon S3) in Apache Iceberg format for the data lake, Amazon Elastic Container Service (Amazon ECS) for application hosting, and Amazon Quick as an internal natural-language AI assistant for ad sales. The release named the six AWS services but did not list any third-party measurement or verification vendors that plug into the new platform.

How does this fit with Fox’s agentic AI push?

DecodeTV reported the WBD announcement arrived shortly after Fox launched its own agentic AI advertising platform, putting two of the largest U.S. broadcast owners through similar overhauls within weeks of each other. AWS itself is using Cannes Lions 2026 to make the agentic ad pitch broadly, with a Tuesday, June 23 panel featuring WBD’s Sethu and AWS’s Bakhtiar. The wider read is that premium video advertising is converging on a single agentic blueprint run out of the same cloud provider.