The Trade Desk’s AI-powered buying platform, Kokai, now processes over 13 million ad impressions per second, applying thousands of signals to bid decisions in milliseconds. The company says the bulk of client spend has already migrated to the system, and full adoption is expected by year-end. But even as Kokai modernizes programmatic advertising, a larger shift in how users find information—driven by AI search summaries that slash referral traffic to publishers—could erode the very open-web inventory that The Trade Desk depends on.
Launched in June 2023, Kokai is the company’s most ambitious AI bet. It builds on earlier machine learning work (Koa) but distributes deep learning across the entire buying stack: impression scoring, predictive clearing, dynamic audience segmentation, and budget pacing. The Trade Desk frames it as a “copilot”—advertisers set goals and guardrails, while the AI optimizes toward those targets in real time. Early case studies cited by management show steep improvements in reach and acquisition metrics for clients like Samsung and Cashrewards, and investor communications consistently highlight accelerating spend among Kokai adopters.
How Kokai works: AI as a co-pilot, not a replacement
At its core, Kokai ingests a massive firehose of data—the company says it’s the equivalent of more than 13 million ad impressions every second—and scores each opportunity across thousands of signals. The system then decides whether to bid and how much, all in the time it takes a web page to load. That technical foundation powers a suite of tools inside the Kokai user interface: Koa Audiences for dynamic audience creation, the TV Quality Index for connected TV measurement, the Retail Sales Index for retail media, and more.
The philosophy is overtly human-in-the-loop. “It’s not about replacing the marketer; it’s about giving them superpowers,” CEO Jeff Green has said on earnings calls. The platform lets advertisers set budgets, frequency caps, and brand safety rules, then the AI pursues those objectives, continuously learning from campaign performance. This mirrors an industry-wide shift toward outcome-based buying, but The Trade Desk argues Kokai is uniquely tuned to the open internet and to privacy-first identity solutions like Unified ID 2.0.
Adoption and early performance
By early 2025, roughly two-thirds of clients had moved to Kokai, and the company expects near-complete migration by year-end. In earnings transcripts, executives stress that Kokai users spend more and grow faster than those on legacy interfaces—a pattern that has held for several quarters. While vendor-supplied case studies always warrant scrutiny, the consistency of the message across multiple quarters and independent analyst checks lends some credibility.
For example, Samsung’s European team reportedly used Kokai to boost reach by double-digit percentages in a recent campaign, while Australian cashback platform Cashrewards saw a significant lift in customer acquisition efficiency. These improvements are attributed to Kokai’s ability to continuously optimize across a fragmented supply chain, something manual campaign management struggles to replicate at scale.
The dark side: AI search is starving the open web of traffic
The growth narrative, however, is shadowed by an existential risk. AI-powered search—Google’s AI Overviews, Microsoft Copilot, Perplexity, and others—is dramatically reducing the number of users who click through to publisher websites. Instead of seeing ten blue links, users get a synthesized answer right on the search results page. That convenience comes at a cost: Pew Research Center found that sessions triggering AI overviews generated link clicks only about 8% of the time, versus roughly 15% when no overview appeared. That’s a near-halving of click propensity. Other analytics firms report year-over-year traffic declines for major news outlets of 30% to 40%.
For The Trade Desk, fewer clicks mean fewer ad-supported impressions on the open web. As supply shrinks, the remaining quality inventory could become more expensive, squeezing advertiser margins and reducing the reach that independent DSPs can offer. This isn’t a distant threat—it’s already measurable.
Making matters worse, the AI platforms are beginning to insert ads directly into these AI-generated answers. Google confirmed it is expanding ads inside AI Overviews and its new “AI Mode,” making existing Search, Shopping, and Performance Max campaigns eligible for placements within AI responses. Microsoft has integrated ads into Copilot and is developing conversational ad formats. When AI answers start carrying sponsored messages, advertiser budgets will naturally follow, further concentrating demand inside a few walled gardens and away from the open programmatic ecosystem.
The Trade Desk’s countermoves
Facing a structural headwind, the company is executing a multi-pronged defense.
Premium, curated supply. Initiatives like OpenPath and the “Sellers and Publishers 500+” list aim to aggregate brand-safe, high-quality placements that advertisers will still pay a premium for, regardless of search referral flows. This curated approach both protects margins and gives The Trade Desk a differentiated offering to pitch against commoditized inventory.
CTV and retail media push. Connected TV and retail media are less reliant on organic search, making them natural hedges. Indsutry forecasts call for continued double-digit growth in CTV ad spending, and retail media networks are expected to add tens of billions in global spend. The Trade Desk already holds a strong position in programmatic CTV, and management is leaning in.
Generative AI creative tools. Kokai now integrates with generative AI creative vendors like Rembrand, Spaceback, and Bunny Studio. Rembrand, in particular, enables virtual product placements and spatially aware brand insertions in video content, letting The Trade Desk capture a slice of the creative value chain—not just media buying. If these tools scale, they could lock clients more tightly into the Kokai ecosystem and expand the company’s total addressable market.
Strengths and opportunities
- Platform modernization at scale. Kokai is a genuine engineering achievement, applying distributed deep learning to a complex, real-time problem. The rapid adoption and consistent performance anecdotes suggest it’s more than vaporware.
- Curated supply and transparency. Products like OpenPath and the Sellers & Publishers 500+ give advertisers concrete ways to find premium, measurable inventory—crucial if open-web scale softens.
- Channel diversification. CTV and retail media are the fastest-growing digital ad channels, and The Trade Desk is well positioned in both.
- Early move into creative. By aggregating generative AI creative partners, the company could become a one-stop shop for ad creation and delivery, deepening client stickiness.
Risks and unanswered questions
- Inventory compression and pricing pressure. If AI search permanently reduces publisher traffic, the remaining open-web impressions could become more expensive and less scalable, hurting campaign economics and driving budgets toward platforms that control their own inventory.
- Platform monetization of AI experiences. Google’s and Microsoft’s built-in advantages—they control both the AI discovery layer and the ad inventory—are structural advantages The Trade Desk cannot replicate. Ads inside AI answers could siphon significant demand away from open-programmatic buys.
- Client friction with forced automation. Not all advertisers embrace full automation. Anecdotal reports from agencies cite workflow disruptions and trader-efficiency concerns during Kokai migrations. If a meaningful subset of high-spend clients resists, it could slow monetization and create churn risk.
- Measurement reliability in a fragmented landscape. As ad spending fragments across closed AI experiences and proprietary retail media networks, building a cross-platform “truth set” grows harder. The Trade Desk’s measurement tools (TV Quality Index, Retail Sales Index) are a start, but advertisers will demand auditable, consistent metrics—failing to deliver could tilt budgets toward vertically integrated stacks.
- Regulatory and legal uncertainty. Antitrust actions against major AI platforms, content licensing disputes, and evolving rules around LLM training data could reshape the economics of AI search—for better or worse—over the medium term. The outcome is highly unpredictable.
What it means for advertisers and investors
For advertisers: Test Kokai on controlled budgets before committing core spend. Use holdout experiments to validate vendor performance claims. Simultaneously reassess channel mix, prioritizing CTV, retail media, and curated open web where measurement and inventory quality align with goals. Build a creative workflow that can leverage generative AI tools without sacrificing brand safety.
For investors: Monitor three leading indicators on a quarterly basis:
- The percentage of client spend running through Kokai and whether retention and expansion metrics improve for Kokai adopters.
- Trends in open-web supply and CPMs for premium inventory—is compression materializing, or is premium inventory holding up?
- The pace at which AI search features monetize with ads and whether those placements shift advertiser demand away from open-web buys.
The bull case: Kokai drives measurable performance gains, client stickiness increases, and The Trade Desk captures market share in CTV and retail media, offsetting any open-web compression. The bear case: a sustained migration of discovery to integrated AI search experiences concentrates budgets inside platform owners and permanently reduces the addressable inventory pool for independent DSPs.
The road ahead
Trade Desk faces a classic tech-industry fork. Kokai is both a defensive necessity and an offensive bet—it modernizes the product, embeds AI deep into the stack, and creates new levers for measurement and creative orchestration. Those are meaningful strengths. Yet the external environment is hostile in ways product upgrades can’t fully control.
The company’s playbook—invest deeply in AI to improve intrinsic value, while aggressively expanding into channels less exposed to AI search cannibalization—is strategically coherent. But execution risk is high, competition is intense, and the regulatory environment adds further uncertainty.
The coming 12–24 months will tell if Kokai is a growth engine or an insurance policy against a shrinking open web. Advertisers and investors alike should watch the data, not the narrative, to determine which scenario wins out.