Newsletter · · Ashutosh Agarwal

Walled Gardens Win as The Trade Desk Stumbles - Weekly Online Ads Podcast Recap - Week of August 30, 2026

Weekly Online Ads Podcast Recap for the week of August 30, 2026. Podcast synthesis on The Trade Desk's growth collapse and the value trap debate, Meta's 28 percent ad growth against a free cash flow of 784 million dollars, retail media going full funnel and off platform, ad tech's golden age of M&A, and ChatGPT advertising becoming a real line item.

Weekly Online Ads Podcast Recap

Week of August 30, 2026: Walled Gardens Win as The Trade Desk Stumbles

Much of this issue centers on the Q2 2026 earnings season (Meta, Microsoft, Amazon, The Trade Desk) plus the Cannes Lions 2026 conversations.

Executive Summary

  • The walled gardens are winning, and the open internet's champion is stumbling. The single loudest story is The Trade Desk (TTD): revenue growth collapsed from 12% to 3% to a guided ~12% decline, the stock fell ~20% on the print and sits ~90% off its high, and hosts on Motley Fool Hidden Gems Investing branded it a "value trap" even at a 6.6x forward P/E. Meanwhile Meta, Alphabet and Amazon keep compounding ad revenue at double digits off far bigger bases.
  • Meta's ad machine is healthier than ever, but the market is punishing it anyway. Meta ad revenue grew 28% to $60.8 billion, yet free cash flow cratered to $784 million from $8.6 billion a year earlier as AI capex (guided to $130 to $145 billion) swallows the profits. On The Rundown and Squawk on the Street, the debate is whether advertising alone can ever justify the spend. Jim Cramer flatly said the ad model "doesn't work" for monetizing compute.
  • Retail media is being re-cast as "the future of all media." At Cannes Lions, consultant Andrew Lipsman and Instacart's Adam Silverblatt argued retail media is now full-funnel and moving off-platform onto The Trade Desk, Roku, Pinterest and TikTok, with a McCormick case study showing a 93% incremental audience and 20%+ lower CPMs (The CPG Guys).
  • Ad tech is in a self-described "golden age" of M&A. LiveRamp (to private equity), DoubleVerify (to Nielsen), and Walmart's purchase of CTV firm Vibe dominated Mobile Dev Memo. The through-line: everyone is buying data, identity and measurement, not inventory.
  • Google's ad-tech monopoly may survive the antitrust remedy. On Marketecture, the take on Google's new "BuyerDirect" product was blunt: the DOJ asked Google to spin out the ad exchange but not the ad server, and BuyerDirect "blows that out of the water" by rebuilding the money-maker inside the server the company gets to keep.
  • Social video has passed CTV, and AI-made "creative" plus AI-made fraud are both scaling. The IAB pegs the US video ad market at $82 billion, with social video ad spend now ahead of connected TV (Next in Media). Meta says 8 million advertisers now use its generative-AI ad tools, while The Marketing Architects warns AI "did not invent junk inventory… it just really lowered the cost of making it."
  • ChatGPT ads are real and growing, just measured differently. OpenAI is on track to beat a $1 billion ad-revenue target this year (eMarketer sees ~$10B by 2030), is running mostly display dollars (not search), launching Criteo-style product-carousel ads, and partnering with AppsFlyer on measurement.
  • Tickers most discussed: TTD (bearish), META (bull ad business / bear cash flow), GOOGL/GOOG (antitrust + AI search), AMZN (~$70B ad run-rate, +26%), APP/AppLovin ("great product, momentum stock"), RDDT (+64%), plus MGNI, PUBM, ZETA, CRTO, DV, PINS, ROKU and WMT/Walmart Connect.

(a) Dominant Themes

1. "The walled garden is winning", the open internet's toughest quarter

The most-discussed narrative, by a wide margin, was the divergence between the big closed platforms (Meta, Alphabet, Amazon) and the open-internet ad-tech ecosystem led by The Trade Desk. On Motley Fool Hidden Gems Investing (Aug 7), the hosts walked through The Trade Desk's earnings collapse in plain numbers. As John put it: "You look at the last three quarters, we had 12% growth in the first quarter, only 3% growth in this quarter and forecasting potentially a 12% drop in revenue in the upcoming quarter. This is supposed to be a huge growth business."

Co-host Lou tied it directly to the platform shift: "I think the walled garden is winning… the internet is getting more closed off. The trade desk has always said we have a solution for that. We're not seeing it." He noted that CEO Jeff Green pointedly did not attack the walled gardens on the call this time, "very wise on his part because you look at those numbers, all of those growing by double digits at much higher revenue bases."

The shared point of view across commentators: the money is consolidating into a handful of large, logged-in, first-party-data platforms, and independent middlemen have to prove they still have a reason to exist.

2. Meta's "great ad business, terrible cash flow" paradox

Every markets-focused podcast this month circled the same Meta contradiction. On The Rundown (Jul 30), host Zaid Admani (Public.com) laid it out: "The core business is doing fine. It's doing great, actually. Revenues jumped 28% to $60.8 billion." The problem is what happens after the ad dollars come in. Earnings of $6.18 per share missed the $7.22 expected (dragged by a $2.4 billion legal charge and $1.2 billion of severance), and free cash flow fell to "just $784 million… which is down from the $8.6 billion they did a year ago." In plain terms: essentially all the cash Facebook and Instagram ads throw off is now being poured into AI data centers and chips, with capex guided up to $130 to $145 billion for the year.

The core worry, in Admani's words: "investors want to see an entirely new AI revenue stream, and Zuck hasn't been able to show investors what that will be… this is also the same guy who spent $80 billion on the Metaverse with not much to show for it."

3. Retail media 3.0, "the future of all media"

At Cannes Lions 2026, The CPG Guys (Aug 22) hosted independent consultant Andrew Lipsman and Instacart's Adam Silverblatt for the clearest articulation of where retail media is heading. Lipsman's framing: retail media "isn't just bottom funnel. It is full funnel," and "I've said that the future of retail media is the future of all media… because it's going to be an intersection with every single major pool of ad spend." He argued a "fourth wave", performance TV, is now "splintering off" from retail media.

Silverblatt detailed Instacart's off-platform push, letting brands activate first-party shopper data on outside platforms "like the Trade Desk, Roku, Pinterest, TikTok." The proof point was a McCormick Cinco de Mayo campaign run on The Trade Desk (with Flywheel): of the taco-seasoning-buyer audience Instacart supplied, "93% of an audience that they wanted to use… could not have been reached with the other data," and the brand cut CPMs "by over 20%" while lifting new-to-brand buyers. A second example, Deep Indian Foods, "doubled their ROAS and… reduce[d] CPM costs" using in-flight optimization.

4. Ad tech's "golden age" of M&A, buying data, not inventory

On Mobile Dev Memo (Aug 12), host Eric Seufert and guest Maor Sadra (CEO of INCRMNTAL) framed a wave of deals, LiveRamp (to private equity), DoubleVerify (to Nielsen), and Walmart's acquisition of CTV firm Vibe, as a strategic shift. Earlier M&A cycles were about buying inventory and reach; this one is about data, identity, measurement and decisioning. As Sadra put it: "Right now, I think we're in a… golden age again in [ad]tech." On Walmart/Vibe specifically: "Walmart has such a huge footprint of user-level data… you're kind of building a competitor to Amazon almost overnight."

5. AI is now inside the ad, for creative, for measurement, and for fraud

AI ran through every episode. Meta now has "8 million advertisers using their Gen.AI production tools," per Seufert, a scale that, he argued, makes standalone AI-creative startups hard to justify. On the buy side, The Marketing Architects (Aug 4) warned the same tools are supercharging junk inventory: "AI did not invent junk inventory… It just really lowered the cost of making it." And on CTV, AdTechGod Pod (Aug 19) showed AI enabling scene-level contextual targeting that simply didn't exist before.

6. ChatGPT and answer-engine advertising goes from theory to line item

Across Marketecture Ep. 185 (Aug 7) and Mobile Dev Memo, ChatGPT advertising graduated from "will they, won't they" to a measurable, if young, channel: an ~$1 billion revenue target this year (eMarketer models ~$10B by 2030), Criteo-style product-carousel ads heading into Q4, an AppsFlyer measurement partnership, and Amazon's strategic investment. Forrester's Nikhil Lai noted the surprising nuance: the dollars flowing in are "mostly display dollars… rather than search dollars. Paid search remains fully funded."

(b) Active Debates

Debate 1: Is The Trade Desk a value or a value trap?

  • Value-trap side (the prevailing view): On Motley Fool Hidden Gems, both hosts landed on "trap." Lou: "Fool me once, shame on you. Fool me twice, shame on me. At this point, I am not going to see value here until I see some articulation from them that they know how to solve the mess they're in… it's too risky to try and catch this fallen knife." John agreed, "Value trap for now," warning the 6.6x forward P/E is deceptive because a guided revenue decline means "that is actually going to go up here in the coming weeks." Both flagged a leadership problem: Jeff Green "needs to move upstairs… his voice is no longer resonating on Wall Street."
  • Execution-vs-secular question: The unresolved sub-debate is why it's failing. John leaned toward a fixable execution/leadership problem (peers Magnite, PubMatic and Zeta Global are growing double digits in the same market), whereas Lou leaned structural, "the internet is getting more closed off," and openly questioned whether an AI-agent, bot-driven web even responds to advertising the way the open programmatic model assumes.

Debate 2: Can advertising ever pay for the AI build-out? (Meta)

  • "The ad model doesn't work for this" side: On Squawk on the Street (Jul 30), Jim Cramer was scathing: "Do you want to spend $4.62 billion in reality labs and get $431 million in revs back? The cumulative losses there are $80 billion, David. I just think you don't get to monetize with advertising. It just doesn't work." He mocked Zuckerberg's musing about selling excess compute as Hamlet-like, "To spend or to lease, that is the question," and contrasted it with Microsoft, which "own[s] a distribution system, which AI can be sold through."
  • "It's an applied-AI story and it's on fire" side: On Marketecture Ep. 184, the counter-view was that Meta is precisely the AI winner people underrate: "continue to see them outpacing in terms of revenue growth and… being able to turn the dials on ad impressions and ad prices. That's the AI story to me… It's an applied AI story." Zuckerberg's own framing (played on Squawk): "it would be foolish to basically just sell all of the compute and take a short-term profit… when you have the opportunity to build intelligence on top of it."
  • The tell in the tape: Both sides agreed the 28% ad growth is real; they disagree on whether investors should credit AI-driven ad improvement or demand a brand-new revenue line. The stock's ~10% drop suggests the market currently sides with the skeptics.

Debate 3: Social video vs. connected TV, who wins the next dollar?

  • Social is winning (for now): On Next in Media (Aug 18), the IAB's Chris Brutale reported social-video ad spend "overtook CTV last year and it's likely to extend its lead," driven by small and mid-sized advertisers who "find social to be more effective in delivering across the purchase funnel than CTV." His reason: SMBs are "super comfortable with search and social," and the social platforms keep adding self-serve, AI-generated creative and native commerce.
  • CTV is closing the gap: The counter is that CTV is rapidly building the same self-serve, low-floor ($500), performance-oriented tooling (TV Scientific to Pinterest, Vibe to Walmart), and the desire from SMBs to be on the big screen is clearly there, "there's an entire market of millions of advertisers… who primarily just live on search and social who also want to be active in CTV." A separate Next in Media episode (Aug 6) argued YouTube has effectively already closed the gap for many buyers. Kyle Krueger of Deep Focus said among indie and boutique agencies "YouTube is seen as a major player. And there's almost a parity with what you would receive from traditional TV."

Debate 4: Does "media quality" actually matter, or is it optimization theater?

  • Skeptical side (contrarian): The Marketing Architects team argued most "quality" vocabulary, viewability, brand-safety, contextual adjacency, describes the environment, not whether a real human saw the ad. Rob DeMars (Misfits and Machines) went spiciest: "A great ad on the Super Bowl is a great ad next to a cooking show… The programming that you're putting it on doesn't play into it as strongly as the creative itself." Their view: if your incrementality measurement is honest, "spend less time policing the inputs."
  • The non-negotiable floor: But all agreed there is a hard floor: fraud. DeMars: "the floor is fraud… Bot traffic, spoof domains, made for advertising sites. Those are not a quality debate. They're more of just a conversation around theft," and crucially "going direct doesn't protect you." An eMarketer/Verve survey they cited ("Media Buyers Struggle to Define Quality Inventory and It's Costing Them") found buyers can't even agree on what quality means; reach was the single most common answer.
  • The CTV fraud sub-debate: On Next in Media, the IAB's Brutale was more sanguine. Fraud is "the thing that keeps buyers up at night," but he framed it as a transparency problem more than a rampant-theft problem, noting "only slightly more than half said even direct deals… have a huge amount of trust in the provenance of where their ads run."

Debate 5: Will the DOJ's Google ad-tech remedy actually change anything?

  • "Google already routed around it" side: On Marketecture Ep. 184, the analysis of Google's new "BuyerDirect" was pointed. The DOJ's remedy targets the ad exchange (AdX) for divestiture, "but for some reason the DOJ's remedies… did not ask for the ad server to be spun out." BuyerDirect lets buyers book inventory straight through Google Ad Manager, potentially "replac[ing] AdEx as the next moneymaker after losing the antitrust case," leading to the verdict that "the DOJ is shown with this new product to have missed the boat."
  • "The judge will shut it down" side: The counter is that if the DOJ wins, it "would almost certainly ask the judge to disable this," and Google's defense that BuyerDirect is "not an exchange… it's like an order processing system" that still takes a variable fee "doesn't really hold water." The honest conclusion from both hosts: "the next shoe to drop hasn't dropped yet on this."

Debate 6: Is "performance TV" a real new channel or a category error?

  • Bullish (with a caveat): Forrester's Nikhil Lai sees genuine value in performance TV for mid-market and SMB advertisers not previously on CTV.
  • Bearish on CPA-guaranteed TV: But he was explicit: "I am bearish on CPA guarantees on TV… I am very skeptical of the attribution that's done to actually meet those guarantees." His warning to new-to-TV brands treating CTV like "next gen search or next gen social": "if you begin to try to hold TV accountable to a really aggressive [CPA] goal, you are going to lose share of voice… brands don't grow by optimizing for the lowest possible CPM or the lowest possible CAC."

(c) Tickers Mentioned

TTD, The Trade Desk (bearish)

  • Bull case (as discussed): Optically cheap at a 6.6x forward P/E and ~$5 billion enterprise value; the open-internet/CTV market it addresses is still expanding; peers in the same space are growing.
  • Bear case (dominant): Growth collapsed 12% to 3% to guided -12%; stock down ~20% on the print and ~90% off its high; walled gardens winning; leadership and credibility problem; CEO blaming macro is a red flag.
  • Source: "The Trade Desk's Woes & A New AI Donut?", Motley Fool Hidden Gems Investing, Aug 7, 2026.
  • Speakers: Hosts "John" and "Lou" (Motley Fool analysts).
  • Quote: "Fool me once, shame on you. Fool me twice, shame on me. At this point, I am not going to see value here until I see some articulation from them that they know how to solve the mess they're in. Right now, I think it's a trap or it's too risky to try and catch this fallen knife." Also: "I think the walled garden is winning… the internet is getting more closed off. The trade desk has always said we have a solution for that. We're not seeing it."

META, Meta Platforms (bull ad business / bear free cash flow)

  • Bull case: Ad revenue +28% to $60.8B; an "applied AI story", Meta keeps turning "the dials on ad impressions and ad prices"; won its antitrust case; 8M advertisers on its Gen-AI ad tools.
  • Bear case: Free cash flow fell to $784M from $8.6B a year ago; capex guided to $130 to $145B with no clear new AI revenue line; Reality Labs lost $4.62B against $431M of revenue; Cramer says advertising can't monetize the compute build-out.
  • Sources: The Rundown, Jul 30, 2026 (Zaid Admani, Public.com); Squawk on the Street, Jul 30, 2026 (Jim Cramer / David Faber / Carl Quintanilla, CNBC); Marketecture Ep. 184, Jul 31, 2026.
  • Quote (Cramer): "Do you want to spend $4.62 billion in reality labs and get $431 million in revs back? The cumulative losses there are $80 billion, David. I just think you don't get to monetize with advertising. It just doesn't work."
  • Quote (Admani): "Meta generated just $784 million in free cash flow last quarter, which is down from the $8.6 billion they did a year ago."

GOOGL / GOOG, Alphabet (mixed; antitrust overhang + AI search strength)

  • Bull case: Ads still compounding at double digits off a huge base; a fully integrated vertical stack (own TPUs, own Gemini model); Gemini's "AI mode" search is winning power users; new BuyerDirect product may preserve ad-server economics even if AdX is divested; YouTube is "the undisputed number one streaming platform" reaching TV parity for buyers.
  • Bear case: DOJ ad-tech remedy and monopoly overhang; raised capex spooked investors; BuyerDirect could be forced shut if DOJ prevails.
  • Sources: Marketecture Ep. 184, Jul 31, 2026; Squawk on the Street, Jul 30, 2026; "Connect YouTube Ad Buys to CTV Campaigns", Next in Media, Aug 6, 2026 (Andy Davidson, Bright Mountain; Kyle Krueger, Deep Focus).
  • Quote (Marketecture, on BuyerDirect): "for some reason the DOJ's remedies in the antitrust trial did not ask for the ad server to be spun out. They asked for the ad exchange to be spun out… And this announcement of [BuyerDirect] just blows that out of the water."

AMZN, Amazon (bullish on ads)

  • Bull case: Ads on a $70B trailing-12-month run-rate, "solidly number three behind meta and Google"; a nearly $20B quarter, +26% YoY; highest-margin part of the business that "feeds the machine of the core retail business"; Amazon is now the world's largest company ($717B revenue).
  • Bear case: Not really raised. Ads are a bright spot; the debate is elsewhere (cloud and AI capex).
  • Sources: Marketecture Ep. 184, Jul 31, 2026; Marketecture Ep. 185, Aug 7, 2026.
  • Quote: "Amazon reported last week… almost a $20 billion quarter for Amazon ads up 26% year over year stock went crazy."

APP, AppLovin (bullish product, cautious on the stock)

  • Bull case: Product delivers consistent incremental results for advertisers; the standout public winner in a long-dry ad-tech M&A market; actively acquiring (bought CTV firm WURL).
  • Bear case: Valuation disconnect. Forrester's Nikhil Lai: "there is a disassociation between the intrinsic value of the product and the company and the extrinsic value of the stock… it's trading on momentum."
  • Sources: Marketecture Ep. 185, Aug 7, 2026 (Nikhil Lai, Forrester); "WURL's Mikey Garcia…", AdTechGod Pod, Aug 19, 2026; Mobile Dev Memo, Aug 12, 2026.
  • Quote (Lai): "for the most part folks are finding consistency with it… [but] it's trading on momentum."

RDDT, Reddit (bullish, high growth)

  • Bull case: Ad revenue up 64% year over year, cited as one of the fastest growers in the group.
  • Bear case: Not discussed in depth.
  • Source: Marketecture Ep. 185, Aug 7, 2026.
  • Quote: "Reddit ads up 64%."

MGNI, Magnite and PUBM, PubMatic (bullish, as TTD contrast)

  • Bull case: Both reporting double-digit growth in the same programmatic space where The Trade Desk is decelerating, used as evidence TTD's problem is company-specific, not sector-wide.
  • Bear case: Not separately argued.
  • Source: Motley Fool Hidden Gems Investing, Aug 7, 2026.
  • Quote: "there are other companies in the space, AdTech, Magnite and Pubmatic reporting double digit growth."

ZETA, Zeta Global (bullish)

  • Bull case: "Very strong growth rates," named as a more attractive name than TTD in the space; a host said he had been adding to his position.
  • Bear case: Not discussed.
  • Source: Motley Fool Hidden Gems Investing, Aug 7, 2026.
  • Quote: "one that I have been interested in recently and been adding to my position too is Zeta Global."

CRTO, Criteo (neutral / structural)

  • Context: Framed as the template ChatGPT is copying. Its product-carousel / catalog-listing ad format is the model OpenAI is launching for Q4; Criteo's CEO reportedly sees more upper-funnel search behavior in AI-driven results.
  • Source: Marketecture Ep. 185, Aug 7, 2026 (Nikhil Lai, Forrester).
  • Quote: "They're launching product carousel ads… which again is a very Criteo like format."

DV, DoubleVerify (M&A target)

  • Context: Being acquired by Nielsen, cast as a legacy-measurement player buying "tech-first" verification and pre-bid capabilities; part of the "buy data and measurement, not inventory" wave. (LiveRamp to private equity and Walmart to Vibe were discussed alongside.)
  • Source: Mobile Dev Memo, Aug 12, 2026 (Maor Sadra, INCRMNTAL; Eric Seufert); also referenced on Next in Media, Aug 18, 2026.
  • Quote: "DoubleVerify is a tech-first company that does pre-bids, does measurement… super interesting for Nielsen to get their hands on."

PINS, Pinterest (neutral / performance-TV angle)

  • Context: Pinterest's TV Scientific partnership pushes it into performance and connected TV; also named as an off-platform destination for Instacart's retail-media data. The most interesting open question raised: whether CTV exposure lifts response to Pinterest social retargeting.
  • Sources: Marketecture Ep. 185, Aug 7, 2026; The CPG Guys, Aug 22, 2026.
  • Quote: "TV scientific going to Pinterest, and there's more of a vibe going to Walmart."

ROKU, Roku (neutral)

  • Context: Cited as an off-platform activation partner for Instacart's shopper data; separately, Roku's push into AI-generated content channels and ad-supported streaming came up as a strategy note.
  • Sources: The CPG Guys, Aug 22, 2026; Morning Brew Daily, Aug 13, 2026.
  • Quote (CPG Guys): brands can "activate it out on the open web across partners like the Trade Desk, Roku, Pinterest, TikTok."

WMT, Walmart / Walmart Connect (bullish on retail-media ambition)

  • Bull case: Acquisition of CTV firm Vibe lets Walmart apply its user-level purchase data to connected TV, "building a competitor to Amazon almost overnight"; third-party sellers now make up ~two-thirds of Walmart's ad revenue, a long tail to activate with Gen-AI creative.
  • Bear case: Skeptics note "Walmart hasn't quite figured out what to do with Vizio," so Vibe is another attempt in the same vein.
  • Sources: Mobile Dev Memo, Aug 12, 2026; Marketecture Ep. 185, Aug 7, 2026.
  • Quote: "imagine that you can now expand basically that into a CTV space. You're kind of building a competitor to Amazon almost overnight."

MSFT, Microsoft (context and contrast, not an ad name)

  • Context: Repeatedly used as the anti-Meta, the market rewarding Microsoft's AI capex because Azure (+43%, past $100B annualized) and 30M+ Copilot seats give it a clear monetization path, unlike Meta's advertising-only model. Included because the entire Meta ad-monetization debate is framed against it.
  • Sources: The Rundown and Squawk on the Street, Jul 30, 2026.
  • Quote (Cramer): "They own a distribution system, which AI can be sold through. 365 Windows, GitHub, the commercial backlog was $178 billion."

Also referenced (limited detail)

  • LiveRamp, acquired by private equity; described as the dominant CTV identity player despite legacy roots (Mobile Dev Memo, Aug 12).
  • Nielsen, acquirer of DoubleVerify; the legacy-measurement player "playing catch-up" (Mobile Dev Memo, Aug 12).
  • IAS, Integral Ad Science, named as an independent verification vendor brands should run outside their DSP's own reporting (The Marketing Architects, Aug 4).
  • OpenAI / ChatGPT (private), ~$1B ad target this year, Criteo-style carousel ads, AppsFlyer measurement deal, Amazon investment; display dollars leading (Marketecture Ep. 185, Aug 7; Mobile Dev Memo, Aug 12).