The May 1, 2026 episode of ExchangeWire's MadTech Podcast brought together three stories that, taken together, sketch a picture of programmatic infrastructure in transition. The hosts - Lindsay Rowntree, ExchangeWire COO, Rachel Smith, ExchangeWire CEO, and head of events Grainne Reid - covered Bedrock Platform's containerized DSP deployment inside Index Exchange's Index Cloud, Meta's reported moves toward connected television advertising, and News UK's launch of a synthetic audience planning tool called Times ExplorAItion. Each story touches a different layer of the ad tech stack, but all three point toward the same underlying pressure: the cost and efficiency constraints that have historically favored the largest platforms are being challenged from multiple directions at once.
The containerized DSP: what Index Cloud actually does
The technical centerpiece of the episode was Bedrock Platform becoming the world's first demand-side platform to run its bidder inside an exchange rather than on external public cloud infrastructure. PPC Land covered the April 21, 2026 launch in detail, including the engineering mechanics that underpin the model. The core shift is architectural. Under traditional programmatic buying, when a consumer loads a page and an ad opportunity fires, the bid request travels across the open internet to public cloud servers - Amazon Web Services, Google Cloud Platform, or Microsoft Azure - where the DSP runs its decisioning logic before sending a response back to the exchange. Every network hop adds latency. Every millisecond of latency degrades signal quality. And the compute costs of evaluating bid requests at scale force DSPs to throttle their intake, accepting only a fraction of available impressions rather than the full bid stream.
That throttling constraint is measured in queries per second, or QPS. According to Damian Naglak, head of engineering at Bedrock Platform, "Throttling existed because processing each bid request - egress, cross-network latency, serialization - carried a real per-request cost." The natural response for any DSP has been to filter: evaluate only the share of the bid stream most likely to produce a winning bid, and discard the rest. The problem is that discarding impressions creates a hard ceiling on what AI-driven optimization models can do. If a DSP sees only 20% to 30% of available impressions, its models are trained and operated on a partial view of the market.
Rachel Smith, ExchangeWire CEO, described the significance of the announcement on the podcast. "It's been a while since we've seen some kind of technical innovation at this level that stands to have such a far-reaching impact on the infrastructure layer of the ad tech industry," she said. According to Smith, the move shifts bidding "from a standard DSP infrastructure into being within Index's SSP cloud infrastructure" in a containerized manner that "makes it cheaper, more efficient, it reduces latency, and it means then that that containerized bidder can work on very specific outcomes for buyers in a scaled environment."
Index Exchange CEO Andrew Casale has characterized the problem as a "cloud tax" - operational costs from external infrastructure dependencies that consume resources and constrain innovation. The container execution window runs in under five milliseconds, according to Index Exchange. Each container is cryptographically signed by the partner and verified by Index Exchange, with access tightly permissioned. DSPs receive decisions, not the intelligence behind them. What runs inside each container is controlled entirely by the partner - a design intended to preserve competitive confidentiality while enabling co-location.
Lindsay Rowntree pointed to live streaming as the environment where the latency problem is most acute. At a Video Week panel just weeks before the launch, Index Exchange's head of UK buyer development John Tigg had flagged latency as the single biggest barrier to monetizing live streams. "The QPS, the queries per second load is so great that you just can't, certainly for smaller DSPs, get access to everything you want to be able to properly analyze all the inventory coming your way," Rowntree explained on the podcast. Smaller independent DSPs, she argued, have been systematically disadvantaged because they lack the infrastructure budget to process the full bid stream at the speed large walled-garden DSPs can. Containerization changes that calculus directly.
Shane Shevlin, co-founder and CEO of Bedrock Platform, put the argument plainly: "Scale shouldn't be limited to the largest DSPs." For buyers, the practical implication is that a containerized DSP gains exposure to a higher volume of impressions. That broader view of the market enables better-trained optimization models, which should in turn produce better campaign outcomes - even if the infrastructure change itself remains invisible to the advertiser. Rowntree noted that what advertisers will care about is "how do I make my campaign better," rather than the plumbing behind it.
The announcement has also prompted a wider conversation about whether this signals a structural shift in the DSP-SSP relationship. Smith acknowledged the framing circulating in trade media - that this represents a win for SSPs at the expense of DSPs - but pushed back on it. Financial analyst Brian Wieser of Madison and Wall drew attention to the reductive nature of that framing, she noted. The more constructive interpretation is that containerization enables independent DSPs to compete more effectively for open internet spend, which benefits publishers with unsold inventory as well as buyers seeking alternatives to walled-garden environments.
The Agentic RTB Framework released by IAB Tech Lab in November 2025 for public comment is directly relevant here. Index Cloud's live container deployments directly informed the framework's design, which establishes a standardized way for containers to plug into exchange infrastructure consistently and securely across the ecosystem. Partners currently building on Index Cloud alongside Bedrock include Chalice AI, inPowered AI, and Nano Interactive. Zillow became the first brand advertiser to test containerized RTB with Index Exchange and Chalice AI in August 2025, a pilot that demonstrated how the architecture can surface site-level signals - ad ratio per page, refresh rates, organic video length - that traditional bid requests do not carry.
Meta and CTV: the identity graph meets FAST channels
The second story discussed on the May 1 episode concerned Meta's reported conversations with supply-side platforms, TV manufacturers, and publishers about accessing connected television inventory at scale. According to multiple anonymous sources cited on the podcast, Meta has been in talks with SSPs including Magnite and Comcast, as well as OEM partners, about how to extend its advertising offerings into CTV.
Neither Rowntree nor Smith expressed surprise. Meta's core business model faces structural pressure from multiple directions. User growth on Facebook is challenged - "why on earth would new users be signing up to Facebook as a platform?" Smith asked on the podcast. Instagram provides a younger audience, but neither host sees sustained user growth as a given. Ad loads are at a maximum on existing surfaces. The Metaverse investment delivered limited returns. CTV, with its vast and largely underfilled inventory, represents a logical growth avenue for a company with Meta's advertising infrastructure.
The strategic logic centers on Meta's identity graph - its database of logged-in users with known demographic and behavioral attributes accumulated across Facebook and Instagram. Rowntree argued that overlaying that graph onto CTV inventory, particularly FAST channel supply, is a natural next step. FAST - Free Ad-Supported Streaming Television - has expanded dramatically in recent years. FAST channels have been one of the fastest-growing inventory categories in CTV, with the global FAST market estimated at $9.4 billion in 2024 and projected compound annual growth of 23% through 2030. Yet much of that inventory remains unfilled.
Rowntree described a firsthand encounter with this gap during a weekend stay at an Airbnb where she watched Samsung's FAST channel lineup. "I was just amazed at all the different options," she said on the podcast. "What amazed even more was the advertising options, most of which had not been taken." Ad breaks would contain a single advertisement, or sometimes just dead air with a placeholder message. "The scale of opportunity there is absolutely vast for, if Meta wanted to tap into that," she observed.
The partnership angle with OEM manufacturers such as Samsung is particularly significant. Smart TV manufacturers collect automatic content recognition (ACR) data - identifying what content is displayed on screen by matching pixels or audio signatures against known programming. This generates exposure-based signals tied to specific content and moments, without relying on cookie-based or login-based identifiers. "They have the ACR data. They can see exactly who is watching what and kind of what time they're leaving it," Rowntree noted. Overlaying Meta's identity graph onto ACR data from a TV manufacturer would create a combined dataset of unusual richness - known social identities matched with passive viewing behavior at the household level.
Smith raised the challenge that CTV requires content ownership or deep content partnerships that Meta currently lacks. "The people that are really gonna win in the CTV world will be, I believe, content owners," she argued. Meta does not own streaming content - unlike Netflix, Disney, or Amazon. Any CTV play would depend heavily on partnerships with publishers, SSPs, and OEMs rather than on owned inventory. Rowntree noted that Meta has previously acquired sports rights, including UEFA Champions League and La Liga broadcasts in Latin American and Indian markets, though those efforts predated the current CTV landscape. Whether Meta would consider acquiring content assets at this stage remains speculative.
CTV ad spend is projected to grow nearly 14% in 2026, and more than half of respondents in industry research expect over 60% of CTV buys to occur programmatically in 2026. CTV's share of programmatic media budgets has doubled from 14% in 2023 to 28% in 2025. The channel's growth has sharpened scrutiny over supply quality, measurement standards, and the cost of transacting through multiple intermediaries. A Meta entry at scale would add a major new demand source to an ecosystem already undergoing significant consolidation, and would test whether its social advertising infrastructure can translate to the lean-back viewing context of television.
Times ExplorAItion: the publisher data play
The third story in the May 1 episode was News UK's introduction of a synthetic audience planning tool named Times ExplorAItion - with the AI in exploration capitalised. According to the podcast, the tool combines subscriber behavior, reader panels, engagement statistics, and third-party industry data to let advertisers plan and simulate campaign performance before execution.
The Times reported 659,000 digital subscribers as of the podcast recording date, up 7% in the prior year. That subscriber base provides a foundation of logged-in, intentional readers - not merely casual browsers behind a login wall, but paying subscribers whose behavior can be tracked, analyzed, and modeled with a level of precision not available from anonymous inventory. "The Times has a really good, they have good numbers. They have a really high level of digital subscribers. They've got a good subscriber base. It is a very, very premium audience," said Rowntree on the podcast.
The tool's value lies in its position in the advertising workflow. Rather than engaging with media planners only at the point of buying, News UK uses Times ExplorAItion to enter the conversation at the planning stage. Smith described this as enabling "some kind of simulation of how your campaign would perform within that audience." That pre-campaign modeling capability allows News UK to demonstrate the likely value of its audience before a single pound of budget is committed - a more powerful sales argument than post-campaign reporting.
The synthetic audience component is notable. By constructing simulated audiences from multiple data inputs rather than relying on raw segments alone, the tool can model scenarios that the existing subscriber base alone might not fully represent. This is particularly relevant for advertisers trying to understand how a campaign targeting a niche demographic would perform at the scale The Times can offer.
Smith framed the development within a broader trend of publishers asserting more control over the advertising supply chain. She referenced Nodals AI, whose platform focuses on campaign execution following the kind of simulation and planning that tools like Times ExplorAItion enable. "You can see all of these tools are starting to bleed into one another and give publishers a real edge, um, in terms of control over, firstly, their most important asset, which is their consumer data, and secondly, over that supply chain," she said on the podcast.
Rowntree extended the argument to smaller publishers, suggesting that tools of this kind create an opening for publisher data pooling. If a smaller publisher cannot replicate The Times's subscriber scale independently, collaborating with other publishers to combine first-party datasets could allow a similar simulation-based pitch to advertisers. The Guardian's growth trajectory was cited as evidence that premium news brands are building subscriber bases capable of generating genuinely valuable data. According to the podcast, The Guardian has reached 1.3 million paying supporters globally - a figure that speaks to the financial viability of the subscription-and-advertising hybrid model that now underpins most premium publishers' strategies.
For the marketing community, the three stories discussed in the May 1 episode share a common thread: the transfer of infrastructure leverage away from intermediaries and toward either independent platforms with better cost economics, or publishers with richer first-party data. The programmatic margin debate has been sharpening throughout 2025 and into 2026, with research from the Incorporated Society of British Advertisers finding that publishers receive only 36 cents of every media dollar spent through demand-side platforms. Containerized bidding, CTV expansion, and publisher-led planning tools each address a different slice of the same inefficiency problem - and the pace at which these structural changes are arriving suggests the pressure on intermediaries will continue to intensify.
Timeline
- August 2025: Zillow becomes the first brand advertiser to test containerized RTB with Chalice and Index Exchange, demonstrating supply-side signals available through the containerized architecture
- September 12, 2025: Index Exchange announces duration-based reporting for streaming TV measurement, replacing impression-only metrics with time-aware measurement
- September 16, 2025: Index Exchange becomes first SSP to integrate Gracenote contextual intelligence for streaming TV advertising
- November 2025: IAB Tech Lab releases Agentic RTB Framework v1.0 for public comment; Index Cloud's live container deployments directly informed its design
- February 12, 2026: Index Exchange embeds AI attention signals into its SSP through a partnership with xpln.ai, moving attention metrics from post-campaign reporting into pre-bid optimization
- April 14, 2026: Mediavine brings 18,000 publishers' audience data into Index Marketplaces, making Index Exchange the first SSP to enable buyers to activate Mediavine's curated segments
- April 21, 2026: Bedrock Platform and Index Exchange bring the first containerized DSP bidder live inside Index Cloud infrastructure, placing Bedrock's decisioning logic inside the exchange rather than on external public cloud servers
- May 1, 2026: ExchangeWire's MadTech Podcast episode covering Index Cloud containerized DSPs, Meta's reported CTV talks with Magnite, Comcast, and TV manufacturers, and News UK's launch of Times ExplorAItion synthetic audience planning tool
Summary
Who: Bedrock Platform, Index Exchange, Meta, News UK (The Times), and ExchangeWire COO Lindsay Rowntree and CEO Rachel Smith, with additional context from Andrew Casale (Index Exchange CEO), Shane Shevlin (Bedrock CEO), and Damian Naglak (Bedrock head of engineering).
What: ExchangeWire's MadTech Podcast on May 1, 2026 examined three structural developments in ad tech: the April 21 launch of the world's first containerized DSP deployment inside Index Exchange's Index Cloud infrastructure (Bedrock Platform); reported Meta conversations with Magnite, Comcast, TV manufacturers, and publishers about entering CTV advertising; and the introduction of News UK's Times ExplorAItion tool, which uses subscriber behavior, reader panels, and third-party data to simulate campaign performance for advertisers.
When: The podcast episode aired May 1, 2026. The Bedrock-Index Exchange containerized deployment went live April 21, 2026. The Meta CTV talks were described as ongoing, based on anonymous sources. The Times ExplorAItion tool was introduced as a new product from News UK at the time of the episode.
Where: Index Cloud infrastructure and Bedrock Platform are London-based; Index Exchange is headquartered in New York. Meta operates globally. News UK publishes The Times and The Sunday Times in the United Kingdom.
Why: Containerized DSP infrastructure addresses the QPS throttling and cloud cost constraints that have historically limited independent DSPs relative to large walled-garden platforms. Meta is seeking new advertising revenue streams as user growth on its core social platforms slows and ad loads reach their ceiling. News UK is using premium subscriber data to enter the planning stage of the advertising workflow - moving beyond inventory selling toward audience-value demonstration - as premium publishers face increasing competition for direct advertiser relationships.