Temu’s $962M Meta Ads and Fake Creators
Temu may have spent up to $962 million on Meta ads while 73 of its 100 most-used creators were likely fake accounts. That is a brand tax, not marketing.
Temu may have spent as much as $962 million on Meta partnership ads while 73 of its 100 most-used creators were likely fake accounts.
That is not creator marketing. It is a brand tax disguised as performance marketing.
The $962 million warning hiding in Temu’s Meta ads
A recent investigation by Fortune, drawing on research from Czech non-profit Online Risk Labs (ORL), found that Temu may have spent as much as $962 million on Meta partnership ads in the United Kingdom and 27 European Union countries from January 2025 to April 2026.
Partnership ads are meant to be the more believable flavour of advertising. A creator posts content, gives a brand permission to amplify it, and the brand turns that creator’s post into paid distribution on Facebook or Instagram. Done properly, it is a useful system. The brand gets a human voice; the creator gets paid; the customer gets something less sterile than another polished corporate commercial.
But ORL’s analysis found that, among the 100 creators most used in Temu’s partnership ads during the first four months of 2026, 73 were likely fake accounts. Those 100 creators reportedly accounted for nearly three-quarters of Temu’s partnership-ad activity across the market studied. One account, according to the reporting, was used in 109,541 Temu ad campaigns over 16 months and was viewed more than 1 billion times across Facebook and Instagram. ([fortune.com](https://www.fortune.com/2026/08/31/temu-962-million-meta-partnership-ads-fake-creators/))
Let that sink in. This is not a rogue intern posting dodgy memes at 2 a.m. This is industrial-scale performance marketing running through people who may not be people.
There is an important caveat, because facts matter: there is no evidence in the reporting that Temu itself created the accounts or knowingly used fraudulent ones. Temu did not respond to Fortune’s requests for comment, and Meta declined to comment. Two advertising experts who reviewed ORL’s analysis told the publication that its data and estimates were plausible. ([fortune.com](https://www.fortune.com/2026/08/31/temu-962-million-meta-partnership-ads-fake-creators/))
Still, whether Temu was complicit or merely careless is not the comforting distinction its executives might think it is. If you spend hundreds of millions through a system and cannot tell whether your supposed advocates exist, you do not have a clever marketing program. You have a governance problem wearing a growth-hacker hat.
Meta has built a $10 billion machine for borrowed trust
This story matters because Temu is not some weird edge case on the internet’s rubbish tip. It sits inside one of Meta’s fastest-growing ad products.
Meta said in its first-quarter 2026 prepared remarks that partnership ads had passed a $10 billion annual revenue run rate, more than doubling year on year. Meta reported total first-quarter revenue of $56.3 billion. That tells you exactly why every brand, agency and creator manager is being pushed toward this format: it works well enough at scale to become a very serious revenue line. ([s21.q4cdn.com](https://s21.q4cdn.com/399680738/files/doc_financials/2026/q1/META-Q1-2026-Prepared-Remarks.pdf?utm_source=openai))
The trouble is the product has a built-in temptation. It lets brands buy the thing traditional advertising struggles to earn: the appearance of a recommendation.
Customers know an ad is an ad. They may roll their eyes, but at least the deal is honest. Creator marketing muddies the water. The content borrows the body language of a mate giving you a tip. It looks like discovery, not distribution. And when that “mate” is a synthetic persona with a nonsense biography, recycled footage and a shopping link, the entire proposition becomes a bit grim.
The numbers found by ORL are staggering: 31 official Temu pages reportedly ran more than 9 million ads across the EU and UK between January 2025 and April 2026, generating a cumulative reach of 134 billion. Partnership ads represented 65.8% of those ads and 49.6% of their reach. ([fortune.com](https://www.fortune.com/2026/08/31/temu-962-million-meta-partnership-ads-fake-creators/))
That is not creator marketing as most operators understand it. It is a content factory feeding an auction system.
And once you see it that way, the real issue becomes obvious: the people in the middle are expendable. The creator can be fake. The product video can be AI-generated slop. The brand voice can be a dozen disconnected accounts. The platform still collects the money if the ads keep clearing the auction.
The background every operator needs: performance has eaten the brand department
I am not anti-performance marketing. I like measurable results. I have built businesses; I know that “brand awareness” can become a polite excuse for a marketing team to spend money without owning an outcome.
But the pendulum has swung so hard that many businesses now treat brand as a conversion-rate optimisation problem. Find cheap impressions. Run 500 variants. Amplify the winner. Repeat until the dashboard turns green.
That works—right until it does not.
A brand is not your logo, your colour palette or the nonsense on your values page. It is the shortcut a customer uses to decide whether you are safe to buy from. It is a compounding asset built through thousands of tiny moments: the product arriving when promised, the refund being painless, the message matching reality, and the feeling that you are not trying to sneak one past them.
Fake-creators-at-scale attack every part of that. They might produce clicks in the short term. But they train the market to associate your offer with dodgy claims, faceless content and disposable accounts. The cheap click becomes expensive the moment you need customers to trust you with a higher-value purchase, a repeat order or a recommendation to their friends.
The irony is that Temu’s alleged problem may not be that the system is too sophisticated. It may be that it is too lazy. One expert quoted by Fortune suggested Temu could be pursuing cheaper or more efficient ads, while another called the approach potentially “lazy, not sketchy.” That distinction is not much of a defence. Lazy marketing at massive scale is still a decision. ([fortune.com](https://www.fortune.com/2026/08/31/temu-962-million-meta-partnership-ads-fake-creators/))
The overlooked angle: this is not mainly a Temu problem
Here is the uncomfortable bit for founders and marketers congratulating themselves on being more premium than Temu: plenty of respectable brands are one agency spreadsheet away from the same stupidity.
Most businesses have weak creator due diligence. They check follower count, engagement rate, audience location and maybe whether the last few posts look normal. Tick, tick, tick. Then they send product, sign a usage-rights clause and turn on paid amplification.
That is not verification. That is admin.
The warning signs in this case were not subtle. ORL found frequent account-name changes, thin identities and geographic oddities among the leading accounts. Only eight of the top 100 creators were verified with a real identity, according to the data reviewed by Fortune. ([fortune.com](https://www.fortune.com/2026/08/31/temu-962-million-meta-partnership-ads-fake-creators/))
A serious brand should be able to identify those risks before spending real money. If it cannot, then it is outsourcing its reputation to platforms and intermediaries whose incentives are not aligned with its own.
The second overlooked angle is regulatory. The EU’s Digital Services Act gives researchers and regulators a degree of visibility into advertising on very large platforms that America simply does not provide in the same way. The public ad data helped make this analysis possible. The European Commission describes the DSA as imposing special obligations on platforms and search engines with more than 45 million monthly EU users. ([digital-strategy.ec.europa.eu](https://digital-strategy.ec.europa.eu/en/policies/dsa-vlops?utm_source=openai))
So do not make the rookie mistake of assuming this is a European anomaly because it was discovered in Europe. It may be a European visibility advantage. If you cannot see a problem in another market, that does not mean it is not there.
The contrarian view: AI content is not the villain
Everyone will want to turn this into another sermon about AI ruining advertising. That is too easy, and it misses the point.
AI can make a small business faster, sharper and more useful. It can help a founder test product explanations, localise messaging, create first drafts and serve customers without hiring a bloated team. I am all for that.
The problem begins when AI is used to simulate trust rather than improve service.
Use AI to write better product education? Fine. Use it to create a dozen versions of a real customer story, with permission? Fine. Use it to fabricate a fictional human being who appears to independently recommend your product? That is not clever marketing. It is reputational leverage taken out against future you.
The businesses that win the next decade will not be the ones producing the most content. They will be the ones with the clearest proof behind it: real people, real products, real delivery, real accountability.
What this means for you
If you run marketing tomorrow, do these five things.
First, audit every creator you pay to amplify. Require identity verification, a consistent history of original content, sensible audience geography and clear disclosure of commercial relationships. Do not delegate the final call entirely to an agency.
Second, separate content volume from trust. A thousand cheap videos are not an asset if they make your business look like a dropshipping scam. Set a quality floor before you set a volume target.
Third, track brand damage alongside conversion. Watch refund reasons, customer-support complaints, repeat purchase, branded search, review language and the comments underneath the ads. If sales rise while trust indicators rot, you are borrowing revenue from the future.
Fourth, own the proof. Build a library of genuine demonstrations, real staff, real customers and product evidence. You can use AI to make that material more efficient. You cannot use it to replace the underlying truth.
Finally, ask one brutally simple question before approving any ad: would I be happy for my best customer to discover exactly how this was made?
If the answer is no, kill it.
Marketing is not the art of getting away with something. It is the commercial discipline of making a clear promise and keeping it often enough that people choose you again. Temu’s $962 million warning is that a platform can sell you reach by the billion. It cannot sell you a reputation. That part is still your job.