Two privacy regimes and one new entrant were each supposed to dilute social-advertising concentration. None did. The IAB and PwC’s full-year 2025 report puts U.S. social advertising at $117.7 billion, up 32.6 percent from 2024 and now 40 percent of the $294.6 billion U.S. digital total. Madison and Wall’s January update places Google, Meta, and Amazon at 55 percent of total U.S. advertising revenue over the trailing four quarters, up from 23 percent in 2016, with the top-10 sellers at 72 percent. eMarketer’s 2026 forecast has Meta overtaking Google in global ad revenue for the first time, $243.46 billion to $239.54 billion, with the Triopoly capturing 62.3 percent of worldwide digital ad spend.
The shape underneath those numbers, and the more contested cut on top of them, is what landed at StreamTV Europe in Lisbon in mid-April: four platforms capturing more than 90 percent of social-ad revenue, with Meta alone taking 70 percent of the pie.
The argument Maria Rua Aguete, global head of media and entertainment at Omdia, advanced last week on LinkedIn is that the trade conversation has the wrong subject. “A lot is said about #YouTube and for good reason,” she wrote. “But we should be talking much more about #Meta. […] Facebook and Instagram together account for around 70% of social media advertising revenues. In a market where just four platforms capture over 90%, the level of concentration is extraordinary.” Aguete’s authority on the read is that the data is her firm’s: she leads the Omdia team that produced the underlying research and presented it on the StreamTV Europe stage that The Hollywood Reporter’s Georg Szalai covered on April 14. She has also been on this concentration thesis publicly since at least 2022, when she put TikTok’s ad-revenue trajectory on a path to overtake YouTube and Meta combined by 2027.
Where the four-platform cut converges with everything else
The 90 percent and 70 percent figures are Omdia segmentation reported by The Hollywood Reporter, not numbers in any open-access Omdia release. (The firm’s April 1 forecast deck projects social video platforms generating roughly $400 billion in total streaming advertising revenue by 2030; the social-ad concentration cut surfaced in the keynote and in Szalai’s reporting.) That tier distinction matters for citation discipline. It matters less for the underlying read, because the corroboration is sitting next to it.
Brian Wieser’s January note at Madison and Wall puts the structural frame on the same gravity. “What increasingly separates the largest sellers from the rest is not scale alone, but the ability to offer advertisers a single, global system that spans multiple formats and points of contact,” Wieser wrote, with the trailing-four-quarters Triopoly share at 55 percent of U.S. ad revenue and the top-10 at 72 percent. eMarketer is in the same neighborhood from a different denominator: Max Willens, the firm’s principal analyst, told eMarketer’s own audience that “in surpassing Google, Meta has essentially had many of its core strategies validated.” Meta Platforms’ (NASDAQ: META) Q1 2026 SEC filing, filed last week, supplies the magnitude: $55.024 billion in advertising revenue in a single quarter, up 33 percent year-over-year, ad impressions up 19 percent and average price per ad up 12 percent. Alphabet (NASDAQ: GOOGL) and Amazon (NASDAQ: AMZN) print Q2 inside the same week in late July.
What the firms diverge on is denominator and frame, not direction. eMarketer’s Triopoly figure is global digital ad spend; IAB’s social total is U.S.-only; Aguete’s Omdia cut is global social, per the deck behind the keynote. Anyone who silently swaps those denominators is publishing a different argument than they think they are. What the data agrees on across denominators is that whatever measurement you pick, U.S. or global, social-only or all-digital, concentration is up sharply on the long arc and tight on the short one.
Where the Hollywood Reporter-reported Omdia cut adds something the others do not is the four-platform call inside social, and Meta’s specific share of it. That is inference layered on Omdia’s segmentation rather than a number Meta has confirmed; the Q1 print validates the absolute scale that makes 70 percent plausible without certifying the segmentation itself. We read it that way.
The number underneath the walled-garden thesis
We published a walled-garden opinion on April 27 arguing that the 2026 wave of full-stack ad-tech consolidations had emptied the procurement meaning of “independent.” That piece worked the buy-side: four 2026 deals, four-of-five layers of the walled-garden stack each, Wise and Vanderhook and Henick on the record describing the structural conflicts they were finishing the construction of. What it did not have was the demand-side number underneath the buy-side rationale. Aguete’s data is that number.
Read the timeline against itself. GDPR landed in 2018 and structurally cut the third-party-cookie infrastructure outside walled gardens; Facebook and Google’s combined share kept rising. Apple’s App Tracking Transparency shipped in April 2021 and cost Meta an estimated $10 billion in 2022 revenue; Meta responded with Andromeda and Advantage+, products only first-party data and compute at scale could afford to build, and crossed a $20 billion Advantage+ run rate by Q4 2024. TikTok’s global ad revenue ran from $3.88 billion in 2021 to roughly $17.2 billion in 2024 and joined the four-platform set rather than splitting it. Three resets, three reinforcements, one cumulative concentration number landing on a Lisbon stage. Our view is that the buy-side stack-build cohort the April 27 opinion criticized is the procurement system reading the same data we are reading and concluding that auditability is what’s left to compete on.
The honest version of the “Meta is bigger than the discourse admits” argument is the boring one: the discourse’s center of gravity is years behind the data. The reasonable disagreement is whether the gap closes through new privacy rules, an Amazon or TikTok takeoff in social-ad share specifically, or a structural break we cannot see yet. We do not see public data supporting any of those paths today.
The first real test is the late-July earnings cycle, when Meta, Google, and Amazon all print Q2 within a week of each other. If Meta’s auction efficiency widens against the field rather than holding, the readable signature of Andromeda compounding rather than catching up, Aguete’s frame becomes the consensus read for the upfront 2027 cycle. If the IAB and PwC mid-year update, due later in the year, ticks the social share above its 40 percent 2025 baseline, the cohort of 2026 “independent” stack-builds confirms as managed retreat.
