Alphabet, Meta & Amazon Control 57.6% of Ad Revenue Outside China
If your growth plan is buying more ads, you do not own a brand. You rent customers from three companies that already control 57.6% of ad revenue outside China.
Most businesses do not have a marketing strategy. They have a credit card plugged into Meta.
That was manageable when digital ads were cheap, attention was plentiful and a decent product could hide behind aggressive retargeting. It is a much uglier proposition now. Alphabet, Meta and Amazon command 57.6% of advertising revenue outside China, up from 43.8% five years earlier. The top 25 sellers take roughly 75% of the industry’s revenue. If that does not make you nervous as a founder, investor or operator, you are not paying attention. ([axios.com](https://www.axios.com/2026/06/16/ad-market-forecast))
The real story is not AI ads — it is who owns the customer
The fashionable version of this conversation is that AI will make advertising more efficient. Of course it will. Every platform is racing to sell that dream: better targeting, automatic creative, cleaner measurement, more conversions with fewer people involved.
But the important shift is more brutal than that. Advertising is moving from selling broad attention to monetising consumer intent — what someone is likely to want next. Axios reported that the latest product wave around Advertising Week showed just how quickly budgets are shifting toward the platforms best positioned to capture that intent. ([axios.com](https://www.axios.com/media-trends-membership/2026/10/06/ai-advertising-intent-attention))
That changes the game.
Attention was imperfect, but at least it left room for a clever operator. You could buy a cheap audience, make a sharp ad, build a landing page that did not stink, and turn the difference into margin.
Intent is a different beast. If the platform knows what the buyer wants before they reach your site, it gets to sit between you and the customer at the most valuable moment. It can charge for the introduction, measure the outcome, improve its model and charge you more next quarter. Lovely business if you own the platform. Less lovely if you are one of the businesses feeding it.
The lazy response will be to hand even more of the marketing department to AI and call it innovation. That is not innovation. That is outsourcing your judgement to the landlord.
L’Oréal understands that marketing is becoming infrastructure
L’Oréal is a useful case study because it is not pretending that AI is merely a content toy. Fortune reported that beneath its ads, TikTok activity and virtual try-ons sits a far more serious layer of data, technology and AI — what chief digital and marketing officer Asmita Dubey calls the “pipes and plumbing.” ([fortune.com](https://fortune.com/2026/10/06/as-ai-reshapes-the-consumer-journey-loreal-is-rethinking-its-marketing-engine/?trk=article-ssr-frontend-pulse_little-text-block))
That is the bit most founders miss.
They see an AI tool that can produce 50 Instagram captions before lunch and think they have solved marketing. They have solved the cheapest, least defensible part of marketing: producing more words and pictures. Congratulations. The internet was not short of those.
The valuable work is knowing which customer you are serving, what job they are trying to get done, what they hesitate over, where trust breaks, how your product is discovered, and whether the promise made in the ad survives contact with the product.
That requires customer data you can actually use, sensible measurement, fast feedback loops and a brand position your team can repeat without a 48-page deck. It is boring compared with an AI-generated campaign film. It is also where the money is.
L’Oréal has the scale to build serious infrastructure. Most businesses do not. That does not let smaller companies off the hook. It means they need to be more selective. You do not need a global data lake to learn why your last 50 customers bought. You need to bloody ask them, record the answers and change your message accordingly.
Why this gets expensive fast
The concentration of ad money matters because it creates a structural tax on every business that depends on paid acquisition.
Alphabet, Meta and Amazon are not competing to make your company famous out of generosity. They are competing to own the demand signal, the shopping journey and the transaction. If your only differentiator is your bidding strategy, you are playing poker against the casino while proudly explaining your spreadsheet.
This is particularly dangerous for venture-backed businesses that grew up on cheap digital distribution. Their models quietly assumed that targeting would remain affordable and that acquisition could be scaled by adding budget. That assumption has been cracking for years. AI may improve campaign execution, but it will not magically make the platforms less powerful or auction inventory less competitive.
Adobe’s own recent earnings call tells you where the market is heading. The company highlighted tools for self-serve, on-brand creative inside commerce media networks and brand-visibility tools aimed at understanding how companies appear across AI-driven discovery services. ([fortune.com](https://fortune.com/company/adobe-systems/earnings/q3-2026/?utm_source=openai))
Read that carefully. Creative is becoming faster and cheaper to make. Distribution and credible discoverability are becoming more valuable.
That is why a brand is not your logo, colour palette or some agency’s nonsense about “owning a moment.” A brand is the memory that stops a buyer from treating you as interchangeable when the algorithm presents five options.
The overlooked angle: better targeting can make you more forgettable
Here is the contrarian view: hyper-personalised advertising can weaken a brand if you let it.
When every ad is dynamically adjusted to suit an individual, companies are tempted to say whatever works in the moment. Different promise, different price cue, different emotional angle, different offer. Conversion might lift. But over time, the market may have no coherent idea what you stand for.
You end up with a performance machine that wins clicks but cannot earn preference.
That is dangerous because platforms change. Costs rise. Attribution breaks. A competitor copies your offer. The moment any of that happens, a business with no distinct mental real estate is exposed.
The answer is not to reject AI or performance marketing and retreat to vague “brand building.” That is just another form of corporate cosplay. Use the machines for the work machines are good at: variant testing, production speed, segmentation, workflow and pattern detection.
Keep humans accountable for the few decisions that compound: who you serve, what you promise, what you refuse to be, and what one idea customers should remember after they close the app.
The AI era will reward companies that can operate at speed without becoming generic sludge. That is a higher bar, not a lower one.
What this means for you
If you run a business, do these five things this week.
1. Calculate your platform dependency. Work out what percentage of new customers comes from Meta, Google, Amazon or any single marketplace. Include the customers you think arrived “organically” after seeing paid activity. If one platform can wreck your quarter with a policy or algorithm change, call that what it is: risk.
2. Interview 10 recent customers yourself. Not a survey. A real conversation. Ask what they were trying to solve, what alternatives they considered, what nearly stopped them buying and the exact words they would use to describe the result. Your next campaign should sound more like them and less like your marketing team.
3. Write a one-sentence promise. If your people cannot state, in plain English, why a customer should choose you over the obvious alternative, no amount of AI creative will save you. Make it specific enough to exclude somebody.
4. Build an owned audience every day. Email list. SMS list. Customer community. Repeat-purchase program. Referral engine. Useful content people deliberately return to. Paid platforms are fantastic accelerants; they are terrible foundations when they are your only route to market.
5. Measure profit, not applause. Track contribution margin after acquisition, repeat behaviour, refund rates and retention by channel. A cheap lead that never becomes a proper customer is not a win. It is a well-designed way to lie to yourself.
The platforms will keep getting smarter. Fine. Let them.
Your job is to build a business that is also getting harder to replace — because customers remember it, trust it and choose it without needing to be chased around the internet like a lost sheep.