AI at the Edge: Chalice and OpenX Rewire the Programmatic Video Grid
- 20% reduction in cost per high-value action for Hyundai campaign
- 67% decrease in CPMs compared to traditional programmatic methods
- AI models operate in microseconds for real-time decisioning
Experts would likely conclude that this partnership represents a significant architectural shift in programmatic video advertising, enabling more efficient, privacy-compliant, and outcome-driven campaign optimization.
AI at the Edge: Chalice and OpenX Rewire the Programmatic Video Grid
NEW YORK, NY – June 16, 2026 – In a move that signals a fundamental rewiring of the digital advertising grid, Chalice AI and OpenX today announced a partnership to embed custom artificial intelligence directly into the supply side of video advertising. The collaboration allows advertisers to run their own proprietary AI models inside OpenX’s supply-side platform (SSP), making real-time decisions on video ad impressions before they even go to auction. It’s a technical shift that moves intelligence from the demand-side periphery to the supply-side core, promising a future of video advertising that is more efficient, accountable, and resilient to the industry’s privacy headwinds.
At the heart of the announcement is the integration of Chalice AI’s decisioning platform with OpenXBuild™, a software suite launched in January 2026 designed to give marketers programmable access to supply. This isn't just another bolt-on optimization tool; it represents a deeper architectural change in how programmatic value is created, shifting from post-bid analysis to pre-bid intelligence.
The Supply-Side Revolution: From Static Curation to Dynamic Decisioning
For years, programmatic video has been a balancing act. Advertisers sought premium, brand-safe environments but often had to sacrifice the granular, outcome-based optimization available in other channels. Curation helped, bundling quality inventory into manageable deals, but these structures remained largely static. Optimization was a task left to demand-side platforms (DSPs), which decide whether to bid on an impression from a distance, adding latency and operating with incomplete information.
This partnership flips the model. By using a technology known as 'Custom Decisioning' or containerization, Chalice AI packages an advertiser’s custom AI model into a secure, cloud-native container that runs directly within OpenX’s infrastructure. This container acts as an always-on bidding agent, evaluating every available video impression in real time against the advertiser's specific business goals—be it a sale, a sign-up, or a store visit.
“Video has become the centerpiece of digital advertising, but it has not always been optimized with the same level of precision as other channels,” said Adam Heimlich, CEO of Chalice AI. “By bringing AI into curated video supply, advertisers can move beyond buying access to premium environments and start optimizing for outcomes within them.”
This is the “unbundling” of programmatic that visionaries have predicted. Instead of relying on the black-box algorithms of a DSP, advertisers can now deploy their own bespoke logic at the most critical juncture: the supply source. The result is impression-level decisioning that happens in microseconds, without the data loss or latency inherent in the traditional bidstream. As OpenX CTO Joel Meyer puts it, “This is what the future of supply-side decisioning looks like. Advertisers aren’t just accessing premium video, they’re actively optimizing within it, in real time, using their own data and models.”
The Hyundai Case: Proving the ROI of Edge Intelligence
Abstract technical shifts only matter if they generate tangible value. Early results from this partnership provide compelling evidence. In a campaign for Hyundai, the integrated solution delivered a 20% reduction in cost per high-value action and a staggering 67% decrease in CPMs compared to the “next-best alternative,” which represents traditional programmatic buying methods.
The key is the definition of “high-value action.” Chalice’s models are not trained on proxy metrics like viewability or clicks. Instead, they are trained on an advertiser’s own first-party data to predict the likelihood of a specific impression leading to a measurable business outcome. For an automotive brand like Hyundai, this could be a user configuring a car online or scheduling a test drive. The AI agent, operating inside OpenX’s environment, continuously values each impression based on this prediction, ensuring that ad spend is directed only toward inventory that drives real business growth.
This moves the industry beyond a conversation about media quality and toward a model of media accountability. The dramatic drop in CPMs for Hyundai wasn't achieved by buying cheaper, lower-quality inventory. It was achieved by surgically identifying and paying the right price for the most valuable impressions within premium video environments, effectively filtering out the waste before a bid is ever placed. This level of efficiency is simply not possible when optimization logic sits dozens of milliseconds away on the demand side.
A Resilient Grid for a Post-Cookie World
Perhaps the most significant long-term implication of this partnership is its answer to the industry’s biggest existential threat: signal loss. As third-party cookies disappear and privacy regulations tighten, the ability to target and measure effectively is eroding. The Chalice-OpenX model offers a durable, privacy-forward alternative.
Because the advertiser’s AI model operates within the secure confines of OpenXBuild, it can leverage first-party data and make decisions without that sensitive information ever leaving the controlled environment or being exposed to multiple intermediaries in the bidstream. This minimizes data leakage and enhances privacy compliance. The system doesn't rely on tracking individual users across the web; it relies on predictive intelligence applied directly to the supply path.
This approach is part of a broader industry trend toward supply path optimization and advertiser-controlled AI. Chalice AI has already pioneered a similar containerized solution with SSP Equativ, and other performance-focused advertisers are seeing success with custom decisioning. The technology transforms the SSP from a simple pipe for inventory into an intelligent, programmable decisioning layer, creating a more secure and efficient power grid for advertising value to flow.
While barriers to adoption—such as the complexity of developing custom AI and the need for a cultural shift in media buying—remain, the direction of travel is clear. The era of blunt, proxy-based targeting is ending. In its place, a new model is emerging, one where advertisers bring their own intelligence to the supply source, fusing the goals of brand-building and performance marketing into a single, highly efficient workflow. This partnership isn't just an integration; it's a blueprint for the future of accountable video advertising.
📝 This article is still being updated
Are you a relevant expert who could contribute your opinion or insights to this article? We'd love to hear from you. We will give you full credit for your contribution.
Contribute Your Expertise →