Podcast Summary

Podcast: AdTechGod Pod
Host: AdTechGod
Episode: The Refresh News: August 18 Walmart–Trade Desk Shift, Google’s AI Ad Fraud Push, and Upfronts 2025 Numbers
Date: August 18, 2025
Guest/Speaker: Kate (Marketecture/Refresh Segment Host)

Episode Overview

This episode delivers a brisk, insightful roundup of major developments in the advertising technology world for mid-August 2025. The main topics include the evolving partnership between Walmart and The Trade Desk, Google’s deployment of advanced AI against ad fraud, and a deep dive into the freshly released Upfronts 2025 TV and streaming ad spend data.

Key Discussion Points & Insights

1. Walmart and The Trade Desk: Changing Relationship Status

(Start – 04:45)

Partnership Background: Walmart and The Trade Desk have had an exclusive, high-profile partnership since 2021, recently renewed for four more years. However, Walmart is signaling interest in “exploring its options,” causing industry speculation (01:45).
Strategic Tension:
- Trade Desk’s access to Walmart’s retail data is “a critical status symbol and differentiator” in a landscape where independent DSPs increasingly look alike in features and data.
- Despite the renewed partnership, “Walmart may be looking for greater negotiating power or expanded revenue opportunities… or they may be building out their own in-house buying platform.” (04:10)
Industry Pressure:
- The Trade Desk faces existential competition from Amazon, Google, and Meta, which boast vertical integration and direct access to “highly detailed first-party customer data.”
- The host notes, “Advertisers can't resist the siren song of the all-powerful performance first capabilities…[of] Amazon, DV360, and Meta, thanks to their vertically integrated platforms.” (02:55)
Trade Desk’s Differentiation Efforts:
- Features like the AI-powered Kokai, OpenPath, and acquisition of metadata provider Sinceran are mentioned, though “opinion waffles on both the utility and functionality of Kokai.”
- OpenPath has triggered pushback from publishers who feel “pushed out of the market.”
Notable Quote:
- “For any DSP, losing a differentiator like this just adds more weight onto an increasingly heavy load.” – Kate (04:37)

2. Google’s AI Approach to Fighting Ad Fraud

(04:46 – 07:25)

Behind-the-Scenes Efforts: Google has “been quietly testing out machine learning and AI-based tools to eliminate or mitigate invalid traffic activity in web and app environments over the last year and a half.” (04:52)
Technology Details:
- Google is using “multimodal large language models designed to browse the web and apps the way a human would” (05:18), including the ability to “understand actual functionality and navigational components,” record accidental clicks, and screenshot poor ad quality.
- These models send suspicious activity for human review before enforcing policy.
Core Result: “One of these efforts led to a 40% reduction in mobile invalid traffic.” (06:08)
AI’s Evolving Role:
- “As humans, we cannot physically keep up with or address the sheer volume of issues… and although we're already using machine learning based AI techniques… we've also been waiting for these advanced AI opportunities like large language models to be put into play.” (06:32)
Impact: These advances represent “a lot of promise for the ad industry to be able to create human-like reviews at a scale that we as humans just physically cannot.”
Notable Quote:
- “Large language models are able to understand context and engage on a level that is more human than ever. They improve on traditional machine learning thanks to their reasoning capabilities.” (06:49)

3. Upfronts 2025: TV vs Streaming Ad Spend

(07:26 – 09:40)

TV Ad Commitments Trend:
- “For the third year in a row, primetime TV ad commitments have declined.”
- 2025 Figures:
  - Broadcast primetime: Down 2.5% to ~$9.1B (from $9.34B).
  - Cable: Down 4.3% to $8.68B (from $9.1B).
  - Streaming: Up 18% to $13.2B (from $11.2B).
  - Total: $31B for 2025 Upfronts vs $29.5B in 2024.
Market Drivers:
- Advertisers “continue to invest in more niche audiences… across a multitude of platforms” instead of chasing “monoculture moments” with primetime TV.
- Shift reflects broader societal move towards “subcultures or microcultures.”
Economic/Negotiation Dynamics:
- Despite early 2025 tariff uncertainties, ad spend exceeded expectations; some networks “cut rates to secure more ad dollars,” but not as drastically as 2024.
- CPM discussion:
  - Broadcast: $43.50 (down 4.1%)
  - Cable: $19.35 (down 6.1%)
  - Streaming: Down 7.6% (attributable to increased supply, not advertiser resistance)
Premium Content & Streaming Growth:
- Notable shift of “primetime content, particularly live sports, into streaming environments”—e.g., Tubi streaming the Super Bowl.
- Major players now “making [premium content] ultra accessible for advertisers.”
Notable Quote:
- “It’s no surprise that advertisers want to invest more in streaming as audiences flock to it, but they’re also after the advanced capabilities that tech-powered streaming platforms present: nuanced audience targeting, more sophisticated measurement opportunities, and the growing ability to transact… programmatically.” (09:16)

Memorable Moments

Taylor Swift Reference as Opener: Sets a light, pop-culture-infused tone (“At risk of alienating our listeners by making yet another Taylor Swift reference…this has been a huge week for Swifties.” – 01:04)
Metaphor for Walmart–Trade Desk News:
- “If people were still using Facebook to broadcast their relationship status to the world, Walmart and the Trade Desk’s would have just been updated from exclusive to open…” (01:17)
Industry Realism:
- “[Advertisers] actually invest in platforms that can show direct correlation to business outcomes… regardless of whether those results should actually be taken at face value.” (03:44)
Optimistic Closer:
- “That’s all we have time for this week. Thanks for joining us for the refresh and we’ll catch you next week.” (09:40)

Important Segment Timestamps

[01:03] – Intro to main topics (Walmart/Trade Desk, Google’s AI, Upfronts news)
[01:17 – 04:37] – Walmart/Trade Desk partnership analysis and industry context
[04:46 – 07:25] – Google’s AI use in ad fraud prevention
[07:26 – 09:40] – Upfronts 2025 results and insights into TV vs. streaming spend

Takeaways

The Trade Desk’s exclusive relationship with Walmart could be waning, highlighting broader challenges for independent DSPs facing the might of vertically integrated Big Tech players.
Google’s use of advanced AI and large language models for ad fraud defense is making a notable impact, demonstrating the necessity of AI to keep pace with malicious actors.
Linear TV ad spend continues to erode, with streaming commanding a growing share thanks to audience behavior shifts and the programmatic, data-driven capabilities platforms can offer.
The episode’s tone remains accessible, wry, and closely attuned to industry realities—balancing numbers, strategic context, and a dash of cultural commentary.

Transcript

A (0:00)

Exclusive inventory, fragmented tech, high minimums. Programmatic is built for the few, but advertisers need scale without the complexity. Adlib sits above it all, orchestrating activation, optimizing 24, 7, 365 and giving you back 40% of the time spent juggling multiple DSPs and media partners. One login every platform. No trade offs go to getadlib.com G-E-T a d l I b.com support for this podcast and the following message comes from America's Navy the Navy offers new graduates hands on training and experience in careers like computer science, aviation and medicine, plus education and sign on bonuses. Parents help your grads start their career today@navy.com.

B (1:03)

At risk of alienating our listeners by making yet another Taylor Swift reference, just know this has been a huge week for Swifties. I'm Kate with marketecture and this is the Refresh, your weekly download on what went down in advertising. Today is Monday, August 18th, and this week on the Refresh, we're covering Walmart and the Trade Desk's new relationship status, Google's use of large language models to fight ad fraud, and a nice recap of Final Upfront's numbers from Variety. So let's get into it. If people were still using Facebook to broadcast their relationship status to the world, Walmart and the Trade Desks would have just been updated from exclusive to open, leaving industry speculators and skeptics to wonder, is this Walmart's way of saying, you're losing me? Walmart has partnered with the Trade desk since early 2021, and the two recently renegotiated their partnership last year to extend for another four years. Despite this re upped commitment, it sounds like Walmart may want to start exploring its options in its Q2 earnings. Retail data, particularly from Walmart, was cited as a key revenue driver for the Trade Desk, while CTV accounted for 40% of their client spend. And despite having a positive Q2 earnings overall, investors were left feeling skittish, wondering if the programmatic powerhouse can grow at the rate that is expected of them, particularly in the face of increasing challenges from big tech players like Amazon, who's also focused on retail Media and CTV, as well as AI's ongoing erosion of the open Web as we know it. Independent DSPs like the Trade Desk are increasingly starting to appear homogeneous, at least when it comes to the inventory and data targeting sources they offer, as well as basic platform features or functionality. So beside being a key revenue source for them, exclusivity with a partner like Walmart is a critical status symbol and differentiator for the Trade Desk. They've also differentiated themselves through the rollout of their AI powered buying platform Kokai, the launch of openpath, and their acquisition of Sinceran, a metadata provider they're using to identify and transact on media quality. But as Ad Age recently reported, opinion waffles on both the utility and functionality of Cokai. While openpath is still generating pushback from publishers who feel like they're being pushed out of the market, some of the Trade Desk's problems may be self created, but many of them are reflections of broader challenges the ad tech industry is facing. At the end of the day, advertisers can't resist the siren song of the all powerful performance first capabilities provided by the likes of Amazon DV360 and Meta, thanks to to their vertically integrated platforms, many of which are also integrating AI in flashier forms. These platforms also have a superpower that the Trade desk and other DSPs like them just don't direct access to owned inventory and more importantly, highly detailed first party customer data. Despite these giants receiving ongoing criticism for their media quality and overall integrity, just look at any recent analytics report, they're irresistible to advertisers. The Trade Desk isn't just fighting a David versus Goliath battle, and I'm talking comparatively here. It's not entirely fair to say other DSPs are competing against the Googles and Amazons of the world the way that the Trade Desk is They're fighting a battle of will against advertisers who proclaim they want to invest wisely into media and platforms that prioritize the integrity of the broader industry and the integrity or reputation of their brand as a result, while actually investing in platforms that can show direct correlation to business outcomes which are then easily and confidently shared in QBRs and boardrooms alike, regardless of whether those results should actually be taken at face value. All in all, Walmart may be looking for greater negotiating power or expanded revenue opportunities when it comes to who they share their retail data with, or they may be building out their own in house buying platform. I'm skeptical of this, at least in the near term, but they certainly have the resources to do it for any dsp. Losing a differentiator like this just adds more weight onto an increasingly heavy load. Moving over to Google, where we have to give credit where credit is due, they've been quietly testing out machine learning and AI based tools to eliminate or mitigate invalid traffic activity in web and app environments over the last year and a half. The big G told Kendra Barnett over at Adweek that the efforts have led to meaningful reductions in IVT. Since late 2023, Google's traffic quality division has been working with teams at Google research and Google DeepMind to deploy multimodal large language models that are designed to browse the web and apps the way a human would. These models aren't just running pre programmed algorithms for analysis, they're using advances in large language models to understand the actual functionality and navigational components of an app or website. These tools are also designed to monitor and capture activity on these properties. They're recording things like accidental clicks and screenshotting elements of poor ad quality, whether that's a disruptive experience or deceptive practices, say ads that are difficult for the user to see or engage with, but still garner impressions. This activity is then sent to a human for review to decide how to enforce policy appropriately. A key data finding is that one of these efforts led to a 40% reduction in mobile invalid traffic Ad fraud has been one of ad tech's longest running battles, and we've been using AI to identify malicious schemes and undesirable activity for years now. But as the landscape has exponentially expanded, particularly within open web and mobile inventory, so has ad fraud and issues of brand safety. As humans, we cannot physically keep up with or address the sheer volume of issues that are proliferating in these spaces. And although we're already using machine learning based AI techniques to combat malicious activity in the ad ecosystem, we've also been waiting for these advanced AI opportunities like large language models to be put into play. Large language models are able to understand context and engage on a level that is more human than ever. They improve on traditional machine learning thanks to their reasoning capabilities which allow them to understand and interpret language and meaning rather than solely working within predefined rules. Google's research is demonstrating that using these AI capabilities to identify and resolve issues of ad fraud carries a lot of promise for the ad industry to be able to create human like reviews at a scale that we as humans just physically cannot. Rounding out the refresh today is a little wrap up of upfronts. Variety reported that for the third year in a row, primetime TV ad commitments have declined. Advertisers continue to invest in more niche audiences who are watching across a multitude of platforms and programming rather than betting most of their money on monoculture cultural moments that come with primetime viewing. One thing I've been thinking about a lot lately is society shift towards subcultures or microcultures. This plays out perhaps most prominently in TV and video environments. So while it's no surprise to see Upfront's investments follow this trend line, it is affirming to see the numbers that back it up. Let's talk about those numbers. Dollars committed to Broadcast Primetime fell 2.5% to about $9.1 billion, while the previous year netted $9.34 billion. On the cable side, spend commitments fell 4.3% to $8.68 billion, down from $9.1 billion the year prior. In sharp comparison, streaming saw Investments rise almost 18% to $13.2 billion, a $2 billion improvement on the $11.2 billion that were committed during the 2024 upfronts. Total ad commitments rose nearly the same amount, up to $31 billion from 29.5 billion in 2024. At the start of the upfront season back in May, we were in the thick of tariff uncertainty and it was widely speculated that advertiser spend would be stifled. However, the increased commitments being reported now seem to suggest that advertisers weren't as dissuaded as anticipated, and in some cases networks did cut rates to secure more ad dollars. But these cuts weren't as significant as they were last year. CPMs for broadcast landed at $43.50, a 4.1% decline, and $19.35 for cable, which was a 6.1% decline. On the streaming side, the average CPM for a 32nd ad fell by 7.6%. However, streaming CPM declines are more likely stemming from an imbalance of supply and demand in the overall market thanks to an influx of inventory rather than the need to negotiate with advertisers to entice them to buy ad space. It's no surprise that advertisers want to invest more in streaming as audiences flock to it, but they're also after the advanced capabilities that tech powered streaming platforms present, nuanced audience targeting, more sophisticated measurement opportunities, and the growing ability to transact on that inventory programmatically. On top of that, major platforms have been focused on achieving and growing profitability by increasing ad loads, upping prices to encourage shifts to ad supported tiers, while hopefully reducing churn and cracking down on password sharing. When it comes to programming itself, we see networks and platforms make significant moves throughout 2024 and into 2025 to bring primetime content, particularly live sports, into streaming environments. Tubi streaming, the super bowl is the biggest example that comes to mind, but if you've been a consistent follower of the refresh, you know that all of the major players have been bringing premium primetime content into their streaming platforms and more importantly, making that ultra accessible for advertisers. That's all we have time for this week. Thanks for joining us for the refresh and we'll catch you next week.