Bluesky has around 46 million registered accounts, 10.4 million monthly actives, and roughly 3 million people who open it on a given day. Nearly every guide you'll read quotes the first number.
That gap isn't a rounding issue. It's the whole story — and it points in the opposite direction to how this platform is usually described. Bluesky's app has been shrinking through 2026, not growing, while the article you read last month probably called it the fastest-growing social network.
Here's where the three text platforms actually stand, using the most recent figures available, and what that implies for a channel decision you might be about to make.
The three-way, with current numbers
Similarweb data reported in August 2026 gives the clearest recent picture.
| Platform | Daily active (mobile) | Direction |
|---|---|---|
| Threads | 147 million | Up 21.3% year on year |
| X | 123.7 million | Down 7% in July |
| Bluesky | ~3 million | Down 25.6% year on year |
Two things in that table are worth sitting with.
The first is that Threads now exceeds X on daily mobile users, and the two are moving in opposite directions. X is not collapsing — it still has around 302 million mobile monthly actives as of June 2026, and its web visits actually rose 5.3% year on year in July to 4.7 billion. But on daily app engagement, it has been passed.
The second is the scale difference between the top two and Bluesky. This is not three comparable options. It's two large platforms and one small one, and any strategy that treats them as peers is mispricing the smallest by a factor of forty.
Bluesky's app is shrinking while its sign-ups grow
The detail that most coverage hasn't caught up with.
Bluesky's mobile monthly actives were 10.4 million in June 2026, down 27.2% year on year. Against a quarterly average of roughly 22.1 million in Q4 2024, that's a decline of about 52% — the app has lost more than half its monthly users from its post-election peak. Daily actives fell 25.6% year on year to around 3 million.
Registered accounts, meanwhile, keep climbing towards 46 million. Which is exactly how you end up with a marketing ecosystem confidently citing a 40-million-user platform that around 3 million people actually open.
The number to distrust Registered accounts measure historical curiosity. Daily actives measure current habit. When those two diverge for eighteen months, the larger number has stopped describing anything you can market to.
The interpretation is fairly clear: a lot of people who joined during the post-election migration from X never formed a daily habit, and have drifted. Platforms that grow on political migration often follow this pattern once the original grievance cools.
The signal that matters more than the numbers
Usage figures fluctuate. Company strategy is a harder signal, and Bluesky's has visibly changed.
Toni Schneider — founding chief executive of Automattic, the company behind WordPress — replaced Jay Graber as CEO in March 2026, and had "interim" removed from his title in July. Under him, the stated priority has moved away from the consumer app and towards the AT Protocol, the open standard intended to power an ecosystem of independent apps and services. Projects like BlackSky, Eurosky and Skylight are the visible output of that shift.
That's a reasonable bet for a company. It's a different thing for a marketer to read.
If you're weighing whether to invest a year of consistent posting into a channel, the fact that the channel's owner has redirected its attention elsewhere is material. Not disqualifying — but it belongs in the decision, and it appears in none of the platform guides currently recommending Bluesky as a growth opportunity.
Worth adding: Bluesky has no advertising and no clearly stated revenue model. That isn't automatically bad — it's part of what the audience likes — but it does mean the platform's long-run shape is unsettled in a way Threads' isn't.
Wati
WhatsApp Business API platform for team inbox, broadcasts, chatbots and automated customer conversations.
Best for: SMBs automating sales & support on WhatsApp
The honest counter-argument
Everything above is the bear case. The bull case is real and deserves stating properly rather than dismissing.
Stickiness holds up. Bluesky's ratio of daily to monthly actives sat around 29% in June — roughly level with Threads. The people who stayed are engaged. A shrinking user base that retains its stickiness is losing tourists, not residents.
The audience is unusually concentrated. Technology, journalism, academia, developer and policy communities are heavily over-represented. For a B2B software brand or a publisher, 3 million people weighted that way may be worth more than 30 million weighted at random.
The live-event angle is genuinely differentiated. Bluesky's 2026 roadmap leans hard into real-time conversation — a Live Event Feed, custom feeds treated as first-class, and cross-app integrations like the "Live Now" badge that links a profile to an active Twitch or Streamplace stream. The company points to more than 600,000 baseball posts during the World Series and a 30% traffic bump during Game Seven as evidence the behaviour is already there.
Custom feeds are a real structural difference. Discovery isn't controlled by one ranking algorithm, so a niche community feed can put you in front of exactly the right small audience, and the posts themselves stay discoverable in search rather than vanishing into a ranked timeline — closer to how building an engaged community works than to broadcast social.
What Threads passing X actually means
Less than the headline suggests, and more than nothing.
Threads' advantage is structural rather than cultural: Meta can distribute your post to people who don't follow you, using the same recommendation infrastructure that powers Instagram. That makes growth from zero materially easier than on platforms where reach depends on an existing following. Its web visits rose 112% year on year to 471.6 million, which suggests it's becoming a destination rather than only an app.
The trade is the one you'd expect from algorithmic distribution: less control over who sees you, and a feed environment that rewards the same broad, agreeable content that performs on Instagram. If your material is specialist, the reach advantage may not convert.
X's position is more nuanced than "declining." Mobile engagement is softening, but total monthly reach remains the largest of the three and web traffic is growing. For categories where X still carries the professional conversation — and it does in several — leaving on the strength of daily-active headlines would be premature. The picture there is closer to the one described in the state of LinkedIn for B2B: audience composition matters more than aggregate size.
The production cost nobody prices
A practical point that decides more of these debates than strategy does.
X allows 280 characters on the free tier. Bluesky allows 300. Threads allows 500. Those aren't close enough to write once and post everywhere — write to 280 and you waste the room Threads gives you; write to 500 and you're truncated on two platforms.
So each text platform genuinely needs its own version, which means the cost of adding one isn't "another account." It's another draft, another review, another thing to be inconsistent about. That's the arithmetic behind why brands are cutting their channel lists rather than extending them, and it's why the craft of writing to each limit is a resourcing question before it's a creative one.
Three text platforms done properly beats five done automatically. If you can't staff three distinct versions, run two.
What the protocol question means, and when
The open social web — AT Protocol, ActivityPub, the general idea that your identity and audience shouldn't be locked to one company — is a real technical shift and a plausible long-term future.
It is not a 2027 marketing decision.
What's worth doing now is small: hold your handle on the platforms you might later want, keep your owned audience — email list, site — as the thing you actually control, and treat any platform audience as rented. That's not new advice, but the protocol conversation is a good reminder of why it's right.
What's not worth doing now is building a strategy around infrastructure that doesn't yet have an audience attached. Budget against where people are; watch where the plumbing is going.
Where the attention actually went
One thing the three-way framing obscures: the people who left Bluesky mostly didn't return to X, and the growth in Threads isn't obviously made of them either.
X's own monthly actives were down only around 3% year on year in June, which is not the pattern you'd see if a few million Bluesky users had come back. And Threads' growth is large enough that it plausibly reflects people who never used a Twitter-like service at all, arriving through Instagram rather than migrating from anywhere.
The likely answer is less satisfying than a migration story: text social simply isn't where a lot of that attention went. It went to video, to private groups, and to the discovery surfaces that behave more like search — which is why the visual platforms covered in Pinterest and visual search and the shifts in brand-creator deals matter more to most channel decisions than which text app is ahead this quarter.
It also argues for treating text platforms as listening posts as much as broadcast channels. Three million concentrated daily users is a small audience to publish to and a genuinely useful one to listen to, and the second job is far cheaper to staff than the first. If publishing capacity is the real constraint, that is a content production question rather than a platform question.
How to decide
- Start with the overlap test, not the growth chart. Audience composition beats audience size at every scale below "everyone."
- Count the production cost honestly before adding a channel, using the character limits above as the unit.
- If you add Bluesky, add it for a specific community, engage in custom feeds, and treat live moments as the entry point rather than scheduled broadcasting.
- If you add Threads, expect the Instagram dynamic — distribution to strangers, rewarding broad content — and judge it on that basis.
- Don't leave X on a headline. Check your own referral and mention data first, using the process in running a proper social audit.
- Re-run this decision every six months. Every number in this article moved by more than 20% in a year, and the guides quoting 2025 figures are already wrong.
If you're building the channel mix from scratch rather than adjusting one, building a social media strategy from scratch is the right starting point, and a realistic content calendar is what stops the answer being "all of them." The same discipline applies to the visual platforms — the shifts in TikTok's organic reach and in video engagement data deserve the same overlap test rather than a default yes.
The short version
Bluesky has around 46 million registered accounts and roughly 3 million daily users, and its app has lost more than half its monthly actives since late 2024 — so most articles describing it as fast-growing are quoting sign-ups. Its stickiness is nonetheless comparable to Threads, and its audience is concentrated enough that it can be worth it for the right brand. Its company has redirected attention from the app to the AT Protocol, which is a signal worth weighing. Threads has passed X on daily mobile users at 147 million against 123.7 million, though X retains a large lead on total reach and its web traffic is growing. And the character limits differ enough that every platform needs its own version, which is the real cost of adding one.
The question was never which platform is winning. It's which three million people you need, and whether they're in the room.
Running four text platforms and unsure which two are earning their keep?
We check where your audience actually is, then cut the channels that only cost you drafts.
Review my channel mix →Frequently asked questions
How many people actually use Bluesky in 2026?
Far fewer than the headline figure suggests. Registered accounts stand near 46 million, but Similarweb data reported in August 2026 put mobile monthly active users at 10.4 million in June, down 27.2% year on year, with daily active users around 3 million in July, down 25.6%. The app has lost more than half its monthly actives since its late-2024 peak. Almost every article quoting 40 million-plus is quoting sign-ups rather than people who show up.
Is Threads bigger than X now?
On daily mobile usage, yes. Threads reported 147 million daily active users in July 2026, up 21.3% year on year, while X's mobile daily active users fell 7% in July to 123.7 million. X remains larger on other measures — around 302 million mobile monthly actives as of June 2026, and web visits up 5.3% year on year to 4.7 billion. The two are moving in opposite directions on daily engagement while X keeps a substantial lead on total reach.
Is Bluesky worth adding to a brand's social channels?
It depends entirely on audience overlap rather than platform size. Around 3 million daily users is small, but they are concentrated in technology, journalism, academia and developer communities, and stickiness sits near 29% — comparable to Threads. If your buyers sit inside those communities, a small engaged audience can outperform a large indifferent one. If they do not, the channel cost is unlikely to be recovered no matter how well you post.
What is the AT Protocol and does it matter for marketers yet?
It is the open standard underneath Bluesky, intended to let independent apps share identity and content rather than being locked to one platform. Under CEO Toni Schneider, who replaced Jay Graber in March 2026, the company has visibly shifted its emphasis from the consumer app toward growing that ecosystem. For marketers it is a three-year consideration rather than a current channel decision — worth understanding, not worth budgeting against in this planning cycle.
Can you post the same text to Bluesky, Threads and X?
Not without compromise, because the limits differ. X allows 280 characters on the free tier, Bluesky 300 and Threads 500. Writing once to the tightest limit means wasting the space the others give you; writing once to the loosest means truncation. In practice each platform needs its own version, which is a real production cost that should be counted when deciding how many text channels to run.
