TL;DR
- Meta's H1 2026 EU Digital Services Act filing reports 297 million monthly active Instagram users in the European Union against 264 million for Facebook.
- Threads reported 41 million EU users out of a global base of more than 500 million monthly actives.
- Meta cut EU content moderation staff from about 7,700 in its March report to about 5,900 in this one, consistent with previously outlined plans to move to AI powered systems.
- Instagram reports of terror related content rose almost 38% and child sexual exploitation reports about 22% against H2 2025; Facebook reports fell (hate speech 19%, spam 27%, bullying 30%, child exploitation 17%).
- Social Media Today notes the Facebook declines could partly reflect the platform's usage drop in the region rather than improved conditions. No Thai figures were reported.
Meta told European regulators that Instagram had 297 million monthly active users in the European Union during the first half of 2026, against 264 million for Facebook. The figures come from Meta's latest reporting under the EU Digital Services Act, which requires very large online platforms to publish user counts and content moderation data twice a year.
Social Media Today, which covered the filing on 30 August 2026, notes that Facebook's usage in the region has been relatively flat for some time and has declined in some previous reporting periods, while Instagram has risen steadily and is on track to pass 300 million users in the region sometime next year. Threads, the newest of the three surfaces, reported 41 million EU users out of a global base of more than 500 million monthly actives.
A regulator made Meta publish these numbers, and that is the point
Almost every platform user figure a marketer ever sees arrives through a company blog post, an earnings slide or a media kit. Those are selective by construction. The company chooses the metric, the period and the framing, and it publishes when the number flatters it. The Digital Services Act works on a different basis. It designates a class of very large online platforms and obliges each of them to publish average monthly active recipients of the service in the EU on a fixed six month cadence, together with data on the content they acted on and why. Meta does not choose which of its apps appear in that filing, and it does not choose the reporting window.
That is what makes the 297 million and 264 million worth reading twice. They are not audited in the accounting sense, and no regulator has signed off on them line by line. But they are filed under a legal obligation, on a defined methodology, for a defined period, with a known penalty regime sitting behind them. For anyone allocating a paid social budget, that is a firmer footing than a press release, and it is one of the very few platform numbers marketers get that the platform did not volunteer.
Instagram 297 million, Facebook 264 million, Threads 41 million
These are the three headline user counts Meta reported for the European Union in the first half of 2026, with the direction of travel Social Media Today attached to each.
| Meta surface | EU monthly active users, H1 2026 | Reported direction |
|---|---|---|
| 297 million | Rising steadily, on track to pass 300 million in the region next year | |
| 264 million | Relatively flat for some time, down in some previous reporting periods | |
| Threads | 41 million | A small share of more than 500 million monthly actives globally |
The gap between the two older apps is 33 million people, and it is the direction rather than the size of the gap that carries the information. Instagram is now Meta's largest European surface by users, and the filing shows it getting there by growing while Facebook holds roughly where it was. A single reporting period would prove nothing. A run of periods in which one line rises and the other does not is a trend, and Social Media Today describes exactly that run.
The Threads number deserves a second look for a different reason. At 41 million EU users against more than 500 million monthly actives worldwide, Europe accounts for a modest slice of the app. That ratio is a warning against reading global Threads growth as evidence of Threads relevance in any particular market. It is also the only one of the three figures where the filing gives a clean global comparison to sit beside the regional one.
Meta cut EU moderation headcount from about 7,700 to about 5,900
The same filing disclosed a reduction in Meta's EU content moderation staff, from about 7,700 people in the March report to about 5,900 in this one. Meta has previously outlined plans to replace human moderation staff with AI powered systems, and the reduction is consistent with those stated plans. That is roughly 1,800 fewer moderators covering the European Union across a six month gap between filings.
For a brand, this is a planning input rather than a policy argument. Paid social inventory does not exist in isolation from the organic content around it. Feed placements, Reels placements and in stream formats put an advertiser's creative next to whatever the ranking system served that user a moment earlier, and the volume of human review sitting behind that content is one of the variables that determines what gets through. The filing does not say whether the automated systems are catching more or less than the staff they replaced. It reports the headcount and nothing about the outcome, and it would be dishonest to fill that gap by guessing.
Reports rose on Instagram and fell on Facebook
Meta also published the volume of user reports it received, broken out by category. The two apps moved in opposite directions. On Instagram, reports of terror related content rose by almost 38% against the second half of 2025, and reports of child sexual exploitation content rose by about 22% over the same period. On Facebook, every category Social Media Today highlighted fell by double digits across the same two periods.
| Facebook report category | Change, H2 2025 to H1 2026 |
|---|---|
| Hate speech | Down about 19% |
| Spam | Down about 27% |
| Bullying and harassment | Down about 30% |
| Child exploitation material | Down about 17% |
Read quickly, that table looks like a platform cleaning itself up. Read carefully, it is ambiguous, and the ambiguity is the most useful thing in the whole filing.
Why the Facebook declines are hard to interpret
Social Media Today makes the point that matters most: those declines could partly reflect Facebook's overall usage drop in the region rather than improved conditions on the platform. The mechanism is simple. A report is generated when a person sees something and chooses to flag it. Fewer people on a surface means fewer sessions, fewer impressions of borderline content and fewer hands reaching for the report button. A 30% fall in bullying reports is compatible with a platform that got safer, and it is equally compatible with a platform that got quieter.
There is a second measurement problem underneath the first. Reports measure user behaviour, not prevalence. They tell you how often people chose to flag something, which depends on how visible the report control is, how much confidence users have that flagging achieves anything, and how much of the harmful material was removed by automated systems before a human ever saw it. Content taken down proactively never becomes a report. A platform that improves its automated detection can show falling report volumes while the underlying volume of violating content is unchanged, and a platform where users have given up reporting shows the same pattern for the opposite reason.
Set that against Instagram, where reports of terror related content rose almost 38% and reports of child sexual exploitation content rose about 22%. The same logic runs in reverse. A growing user base produces more sessions and more reports without any change in the rate of violating content. Meta did not publish an explanation for either movement in what was reported, and neither number should be converted into a claim about which app is safer.
What the filing did not say
Being precise about the holes is part of using the data honestly.
- No Thai user figures were reported. The Digital Services Act covers the European Union, and Meta files EU numbers because EU law requires them. Nothing in this filing is a measurement of Thailand.
- No country level breakdown within the EU appeared in the reporting, so the 297 million and 264 million are regional aggregates and nothing finer.
- No explanation was given for why Instagram reports of terror related content and child sexual exploitation content rose over the period.
- No statement was made about whether the AI powered moderation systems are performing better or worse than the staff whose numbers fell from about 7,700 to about 5,900.
- No prevalence figures accompanied the report counts, so there is no way to convert a change in reports into a change in how much violating content actually existed.
What this means for Thai marketers
Thailand has no Digital Services Act equivalent and therefore no legally mandated, periodically published user count for any platform operating here. That absence is the reason these European figures are worth a look at all. They are an EU proxy, explicitly labelled as such, and they are not a Thai measurement. Anyone who quotes 297 million in a Bangkok pitch deck without that qualifier is misrepresenting the source.
Used as a proxy, two things follow. The first is about planning. A good number of Thai media plans still treat Facebook as the default Meta buy and Instagram as the secondary placement it inherits budget from. The European filing shows the two surfaces crossing over on users, and while the Thai mix may well be different, the burden of proof has shifted. If your Facebook advertising allocation has not been tested against an Instagram first split in the last year, the European direction of travel is a reasonable prompt to run that test rather than assume the answer. The same logic applies across the wider paid social mix, where placement level splits are usually inherited from an old plan rather than chosen fresh.
The second is about measurement. A European regulatory filing cannot tell a Thai brand where its own audience is. Platform reported placement performance and your own site side data can, and the two often disagree. Before moving budget on the strength of a headline from Brussels, pull the placement breakdown in Ads Manager for the last twelve months and set it against what your analytics shows for sessions and conversions from each surface. If those two views tell the same story, act on it. If they do not, fix the measurement before you move the money.
The moderation headcount reduction adds a third consideration for regulated categories. Brands in health, finance and anything with a compliance officer attached tend to run brand safety controls at the placement and topic exclusion level. The filing does not say those controls are now less effective. It says the human review layer behind EU content is smaller than it was in March. That is a reason to review exclusion settings on a schedule rather than to leave them where they were set two years ago.
Frequently asked questions
Does this mean Facebook is losing users in Thailand?
The source did not say anything about Thailand. Meta's filing reports European Union figures because the Digital Services Act requires EU numbers, and the 264 million Facebook count applies to the EU only. Facebook's Thai user base could be growing, flat or falling, and this filing contains no evidence either way.
Are these user numbers audited?
They are legally mandated rather than audited in the accounting sense. The Digital Services Act obliges very large online platforms to publish average monthly active recipients in the EU twice a year on a defined basis, which removes the platform's freedom to choose the metric and the timing. No regulator has certified the figures line by line, so treat them as a disclosed number under legal obligation rather than a verified one.
Does the drop in moderation staff change where my ads can appear?
The filing reports headcount, not placement policy, so nothing in it changes the placement controls available in Ads Manager. Meta disclosed a reduction from about 7,700 EU moderation staff in March to about 5,900 in this report, consistent with previously outlined plans to move to AI powered systems. Whether that changes the content mix next to which ads appear is not something the filing answers.
Do the falling Facebook report numbers mean the platform got safer?
Not necessarily, and Social Media Today flags this directly. Hate speech reports fell about 19%, spam about 27%, bullying and harassment about 30% and child exploitation material about 17% between the second half of 2025 and the first half of 2026, but those declines could partly reflect Facebook's overall usage drop in the region. Reports also measure how often users flagged content, not how much violating content existed.
Do I have to do anything right now?
No. This is a transparency filing about the European Union, not a product change, a policy update or an account setting. The reasonable response is to check whether your own placement split still matches where your audience actually is, using your own data rather than a European aggregate.
If you want a second read on how your Meta budget is split across Facebook and Instagram placements, and whether your own numbers support that split, our team is happy to look at the account with you.
