Newsletter · · Ashutosh Agarwal

Meta's Ad Engine Accelerated and the Market Punished It Anyway - Weekly Online Ads Podcast Recap - Week of August 3, 2026

A recap of what 37 podcast episodes said about the online advertising industry between July 27 and August 3, 2026, filtered through peak Big Tech earnings week, including Meta's 28 percent ad growth against a collapsing margin, Google's AI Overviews now appearing in 43 percent of searches, and two rich retail-media operator interviews.

Weekly Online Ads Podcast Recap

Week of August 3, 2026: Meta's Ad Engine Accelerated and the Market Punished It Anyway

Coverage window: podcasts published 2026-07-27 through 2026-08-03. This was peak Big Tech earnings week, with Meta, Alphabet, Amazon, Microsoft, Apple, Reddit and others all reporting, so almost every episode this week filtered the online-ads industry through the lens of Q2 2026 results. 37 episodes qualified inside the 7-day window.

Executive Summary

  • The week's single loudest story: Meta's advertising engine is firing on all cylinders, but the market punished the stock anyway. Meta grew revenue 28% to $60.8 billion, faster than any other Big Tech ad platform, yet the stock fell as much as 11% after hours because profit dropped 13%, operating margin collapsed from 43% to 31%, and CEO Mark Zuckerberg would not give investors a clear plan for how his enormous AI spending ($130-145 billion this year) will ever pay off. As DA Davidson's Gil Luria put it on Prof G Markets, the result "was barely passable."
  • Under the hood, AI is visibly making Meta's ads better. On Avory - Markets and Investing, investor Sean Emory detailed that Meta's Advantage+ AI ad suite crossed a $75 billion annual revenue run rate, new Facebook ad models lifted clicks 8.3% and conversions 15.7%, and "every public Instagram [Reel] post and every feed post is now running through an LLM," with more than half of recommended feed content now less than a day old, up from about 25% a year ago. Ad impressions grew 14% and price per ad grew 12%.
  • Google's "AI Overviews" search-disruption thesis got sharper. On The Rundown, host Zaid Admani cited data that Google's AI-generated answers now appear in 43% of searches, up from about 15% a year ago, raising the question of whether people will stop clicking the sponsored links that are Google's bread and butter. For now the ad business is still strong: search ad revenue rose 14% to $81.6 billion. But on The Canadian Investor, the hosts flagged that Alphabet went free-cash-flow negative (roughly -$6 billion) as capex heads toward ~$200 billion, and management refused to answer analysts' repeated questions about returns.
  • Amazon's ad business was a quiet star of the best-received report of the week. On Chit Chat Stocks, the hosts noted Amazon advertising grew 26% year over year and is now an "almost $100 billion business... a high margin one at that." Overall AWS grew 37% (its fastest in 18 quarters); one host asked, only half-joking, whether it was "the greatest quarterly report in the history of capitalism."
  • Retail media had two rich, non-earnings interviews. Walmart Connect International / Sam's Club Connect chief Rich Lehrfeld made the case on The CPG Guys that retail media must move beyond "last-click" sponsored search toward full-funnel, membership-data-driven advertising. Instacart's Chief Commercial Officer Ryan Hamburger explained on Retail Remix how ads revenue grew 16% (fastest in over two years) with 9,000+ brands, and how Instacart now uses its first-party grocery data to power campaigns on Meta, TikTok and Pinterest.
  • Reddit beat on the numbers but stumbled on users. On Schwab Network, Reddit posted EPS of $1.25 (vs. $0.99 expected) and revenue of $804.9 million (vs. $731.8 million), with advertising revenue of $762 million, but daily active uniques grew only ~18% against expectations of 27-30%, and the stock reversed from up ~3% to down ~4%.
  • Generative-AI ad creative was the emerging soft theme. Adobe's enterprise CMO Rachel Thornton described Adobe's new "Brand Intelligence" tool on The Agile Brand, built to stop AI-scaled creative from "morph[ing] over time into something that is basically slop." Canva executives and a Claude Design walkthrough on The Journey with Morgan DeBaun rounded out the picture.

(a) Dominant Themes

1. "Great ad business, terrible stock," Meta becomes the poster child for AI-spending anxiety

The dominant, shared narrative across more than a dozen shows was the strange split screen of Meta's quarter: an advertising machine that is arguably accelerating, wrapped inside a stock the market beat up.

The core numbers, as laid out on The Rundown: revenue jumped 28% to $60.8 billion, but earnings came in at $6.18 per share versus the $7.22 Wall Street expected, dragged down partly by a $2.4 billion legal charge and $1.2 billion of severance. Guidance for the current quarter ($61-64 billion) came in a touch light. Most alarming to commentators, free cash flow was just $784 million, down from $8.6 billion a year earlier, meaning essentially all the cash the core Facebook/Instagram ad business throws off is now being poured into AI data centers and chips.

The shared point of view: the ad business is not the problem, the spending is. As host Zaid Admani summarized, "the core business is doing fine. It's doing great, actually... The problem, though, was the profit and also the guidance." The recurring complaint across shows was that Zuckerberg "didn't offer much clarity on the earnings call on how all this AI CapEx will lead to more revenue," and that improving the existing ad business "isn't good enough for the market" when the spend is this large, because investors "want to see an entirely new AI revenue stream."

On Prof G Markets, DA Davidson's Gil Luria delivered the bluntest verdict, "It was barely passable. They beat by just a little bit and they guided below expectations for next quarter," while still giving credit where due: "Let's be kind to Meta for a second. They just grew 28%. That's almost twice as fast as Google. There's massive share gainers in the advertising market. That's the part investors would love to own. It's all the other stuff that they're less comfortable with."

2. AI is now the visible engine inside the ad platforms, and Meta gave the clearest proof

Beneath the stock-price drama, the most substantive ad-industry content of the week was the granular evidence that AI is directly lifting ad performance. The richest single source was Avory - Markets and Investing, where investor Sean Emory walked through Meta's post-earnings call in detail:

  • Advantage+, Meta's AI-powered ad suite, "crossed... $75 billion of annual run rate revenue," a business bigger than most public companies on its own.
  • Meta disclosed that "every public Instagram [Reel] post and every feed post is now running through an LLM," and that "more than half of the recommended feed content is now less than a day old" (up from roughly 25% a year ago). Fresher, better-ranked content keeps engagement, and therefore ad impressions, higher.
  • New Facebook ad models "increase the clicks by 8.3% [and] conversions by 15.7%." Advertisers getting better returns is exactly why "price per ad is up 12%," alongside impressions up 14%.
  • On the enterprise and AI-agent front: "over one million businesses are now using meta business agents every week," "over nine million small businesses are using at least one AI creative tool," and family-of-apps other revenue hit "$1 billion this quarter, up 73% year over year," driven mostly by WhatsApp paid messaging. A Brazilian rental-car company, Movida, saw WhatsApp AI agents lift bookings 44%, with agents "independently resolving 85% of customer conversations."

Wolf Research's Shweta Kajuria reinforced the point on Bloomberg Tech: Meta's ad business is "growing in the mid to high 20% range... why is that when the overall digital market is growing at, call it, 11% to 12% and they're growing more than double that clip? And it's because of the AI integrations that they're doing that is making the recommendations and targeting a lot better."

3. The AI-Overviews search-disruption thesis is getting harder to ignore

The second-biggest structural theme was Google search. On The Rundown, host Zaid Admani cited SimilarWeb data that "Google's AI-generated answers are now appearing in 43% of searches. This is up significantly from a year ago when they were appearing in about 15% of searches." His plain-English worry captures the whole debate:

"if Google is straight up answering the questions at the top of the page, more and more people will stop clicking the links... this could also hurt Google's own business model because Google's bread and butter is charging websites to be the top sponsored link in search results. If people stop clicking links altogether, do advertisers keep paying Google to be a sponsored link?"

The counterweight: "For now, though, the search business is still doing great. Last week's earnings showed that ad revenue was up 14% to $81.6 billion," and Google says it is beginning to monetize the AI answers, "but we'll have to see if that's as lucrative as the old business model."

4. Retail media grows up, from "last-click" toward full-funnel and first-party data

Two of the most useful non-earnings interviews of the week both landed on the same idea: retail media networks (the ad businesses that retailers like Walmart, Amazon and Instacart run on their own sites and apps) are trying to graduate from simple sponsored-search-and-buy into full-funnel brand building.

On The CPG Guys, Walmart Connect International & Sam's Club Connect head Rich Lehrfeld argued the biggest mistake advertisers make is treating retail media as "this last click purchase. Lower funnel, last click, point of sale." His pitch centers on Sam's Club's membership model as a data advantage: "We know everybody that walks into a club. We know everybody online," giving "100% attribution of all purchases" and a "longitudinal life cycle of a customer" that lets Walmart measure long-term business outcomes, not just short-term ROAS (return on ad spend). His repeated ask to brands: "give us a brief... give us a little and we'll give you a lot."

On Retail Remix, Instacart CCO Ryan Hamburger described a business that just "crossed $1 billion in quarterly revenue" and $10 billion in gross transaction value, with an average order value of $113. The ad piece: "Our ads revenue was up 16%, which is the fastest pace for us in over two years. We now have north of 9,000 brands working with us on our platform." Crucially, Instacart's edge is purchase intent, "we sit at a place of real purchase intent... we know what the customer has historically bought, not just... what they're browsing," and it now uses that first-party grocery data "to now power campaigns on platforms like Meta, TikTok, Pinterest."

5. Generative-AI ad creative: the "anti-slop" arms race

A quieter but recurring theme was the tooling to make ads with AI. On The Agile Brand, Adobe's enterprise CMO Rachel Thornton introduced "Adobe Brand Intelligence," which combines "brand ontology... computer visioning and then a reasoning engine" to ensure that "every asset, whether it's your first asset or your 10,000 asset, they all are on brand" and don't "morph over time into something that is basically slop." Her optimistic framing: the exercise forces brands to finally "quantify what trust means with your brand... you're probably doing the job you should have done 20 years ago."

Canva's leaders struck a similar note on the design-democratization side, describing Canva's shift toward an AI-first, conversational design assistant (Canva AI 2.0). And on The Journey with Morgan DeBaun, the host demonstrated using Claude Design to generate Instagram carousels, animated launch videos and website designs, while recommending brands still invest ~$4,000-5,000 up front in a professional designer to set a brand system that AI can then execute consistently. The common thread: AI slashes the cost of producing creative, but brand guardrails become the scarce, valuable input.

(b) Active Debates

Debate 1: Is Meta's AI capex a masterstroke or a repeat of the Metaverse mistake?

This was the sharpest disagreement of the week.

The bear side was voiced most colorfully by Jim Cramer on Squawk on the Street. He pointed to Reality Labs, "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," and argued the ad model can't carry the compute bill: "I just think you don't get to monetize with advertising. It just doesn't work." He compared Zuckerberg's philosophizing to Hamlet: "To spend or to lease, that is the question... I don't want philosophy. I want profits." The Rundown drove the historical parallel home: "this is also the same guy who spent $80 billion on the Metaverse with not much to show for it."

The bull side came through on Chit Chat Stocks, where the hosts flagged Meta's EV/EBIT down to ~16 and called the stock "dirt cheap," with one host saying he'd flip from "psychologically short this whole year" to long: "ultimately, advertising revenue is going to grow. It's still going to be a high-margin business. And I could see a world where they're generating twice as much operating income in five to seven years as they are today." The bull expectation is that Zuckerberg will eventually force "a period of rationalization on some of the excess spending."

The swing question, framed on Bloomberg Tech, is whether Meta can build a credible enterprise/cloud business at all: it is "late to that cloud game... late to that API game," and faces open-source competition from cheap Chinese models like Kimi K3. "Who those anchor customers are going to be? That's a million dollar question."

Debate 2: Why is the same capex "punished" at Meta and Google but "rewarded" at Amazon and Microsoft?

Several shows wrestled explicitly with why identical AI-spending behavior produces opposite stock reactions. On Schwab Network, the host asked why Amazon's ~$200 billion capex "almost feel[s] like typos" yet is "palatable" while smaller numbers get names "significantly punished." The answer offered: demonstrated growth buys forgiveness. "Alphabet [was] hurt for increased AI spending... Meta hurt for rising expenses and falling free cash flow. Microsoft was rewarded because they've been able to deliver growth." Amazon gets a pass because it can "continue to convince investors that they're able to deliver growth across retail, across advertising, and keep up to speed... in their AWS segment." The diversification of Amazon's high-margin ad business is explicitly part of that cushion.

Debate 3: Is Alphabet's changed business model a red flag or a temporary phase?

On The Canadian Investor, the two hosts split cleanly. Revenue rose 24%, Google Cloud jumped 82%, backlog grew to $514 billion, and operating income rose 30% to $40 billion, but capex guidance has been raised ~11% since Q4 to roughly $200 billion, pushing free cash flow to roughly negative $6 billion, and depreciation is already up 60% year over year.

One host stayed constructive ("I'm still pretty bullish on Alphabet. I own it"), viewing the negative free cash flow as expected and the spending as necessary. The other was scathing about management dodging return questions: "they were asked... 'what kind of return on invested capital do you see?' And Google did not answer... Either they know and it's not good and they're not disclosing it, or they don't know. And either way, I think this is concerning... I wouldn't touch it with a 10-foot pole." The deeper worry both raised: Google's prized asset-light, high-margin model "has completely flipped over on its head," the very thing that made investors love it. A second layer of doubt: if the chips' useful life is really 3-4 years rather than the 6 Google depreciates over, capex would have to reaccelerate even faster.

Debate 4: Reddit, is the digital-ad transition sustainable, or is user growth already stalling?

On Schwab Network, Reddit beat on nearly every financial line, EPS $1.25 vs. $0.99, revenue $804.9 million vs. $731.8 million, advertising revenue $762 million, ARPU up ~15%, and strong forward guidance ($860-870 million next quarter vs. ~$830 million expected), yet the stock flipped from up ~3% to down ~4%. The debate was about the quality of the growth. Bull: "this company is in a bit of a transition... they've moved from this kind of user engagement story into a really strong, profitable digital ad story," with untapped "opportunity among advertising and AI and data." Bear: daily active uniques grew only ~18% "versus expected 27% to 30%," and "these numbers are strong, but they may not be showing enough sustainable growth that the investors are wanting." Reddit's data-licensing-to-AI angle was flagged as the thing investors want more color on.

Debate 5: Retail media, bottom-funnel sales machine or full-funnel brand medium?

Rich Lehrfeld on The CPG Guys explicitly framed the industry tension: brands "traditionally... went from basic reach and frequency" to "measurable things, clicks or short-term ROAS," but "the future is advertisers... wanting to understand more broader business objectives." His frustration with pure ROAS measurement was candid: campaigns get judged on sales that "didn't exist two weeks later, three weeks later. We spent all this money and I got all these sales, but they didn't exist." He conceded the shift is uneven, larger CPGs "are a little ahead," combining brand and shopper-marketing teams, while the middle of the market is still catching up. The unresolved question is whether retailers can actually prove long-term brand value with the same rigor they prove a last-click sale.

(c) Tickers Mentioned

  • META (Meta Platforms). Bull: ad business grew 28% (roughly double Google's rate) and is gaining share; AI is visibly lifting ad performance (Advantage+ at $75B run rate, conversions +15.7%); stock looks cheap at ~16x EV/EBIT. Bear: profit fell 13%, operating margin collapsed 43% to 31%, free cash flow nearly gone ($784M), and Zuckerberg gave no credible plan to monetize $130-145B of AI capex; Reality Labs still bleeding.
  • GOOGL (Alphabet). Bull: massive quarter, revenue +24%, Google Cloud +82%, backlog $514B, operating income +30% to $40B; AI is a tailwind and search ad revenue still grew 14% to $81.6B; capex is "necessary investment." Bear: free cash flow turned negative (~-$6B) as capex heads to ~$200B; management wouldn't answer return-on-capital questions; AI Overviews (now 43% of searches) threaten the click-based ad model; depreciation +60%.
  • AMZN (Amazon). Bull: advertising grew 26% year over year to a ~$100B, high-margin business; AWS +37% (fastest in 18 quarters); AI and chips each at $25B run rate; diversification lets the market forgive the ~$200B capex. Bear: enormous capex and tariff exposure; the same spending scrutiny that hit Meta and Google applies if growth ever falters.
  • RDDT (Reddit). Bull: beat on nearly every financial line, transitioning into a "strong, profitable digital ad story," ARPU +15%, strong forward guidance, untapped data-licensing and AI opportunity. Bear: daily active uniques grew only ~18% vs. expected 27-30%; expectations had run ahead.
  • AAPL (Apple). Bull: 16% revenue growth, iPhone revenue +22% on a strong upgrade cycle (helped by the "Hermes orange" iPhone, big in China); passed NVIDIA to become most valuable company again, the "anti-capex" trade. Bear: stock fell ~8% after hours on high expectations and a lofty ~41x P/E as the "slowest grower" in big tech; chip and component shortages a headwind.
  • CART (Instacart), retail media. Bull: crossed $1B quarterly revenue and $10B GTV; ads revenue +16% (fastest in 2+ years) with 9,000+ brands; sits on true purchase-intent first-party data now powering off-platform campaigns on Meta, TikTok, Pinterest; enterprise "Storefront Pro" lifts retailer sales ~10 points. Bear: none materially raised, this was an interview rather than an earnings-reaction debate.
  • WMT (Walmart), Walmart Connect and Sam's Club Connect. The membership model as a data advantage: near-100% purchase attribution, longitudinal customer lifecycle, and a push to sell brands on full-funnel outcomes rather than last-click ROAS.
  • ADBE (Adobe), generative-AI ad creative. Brand Intelligence as the "anti-slop" guardrail layer for AI-scaled creative, and a bet that brand consistency becomes the scarce input once production cost collapses.
  • MSFT (Microsoft), the ad-spending foil. Rewarded for holding capex flat and showing accelerating cloud growth, the contrast case that defined how the market graded everyone else this week.