Bluesky has roughly 45 million registered accounts and roughly 3 million daily active users. Both numbers are accurate. Which one you cite determines whether you describe a major social network or a mid-sized one.
This post explains where the gap comes from, why it is still widening, and which number to use for what.
The Three Numbers
| Metric | Definition | Mid-2026 | Published by |
|---|---|---|---|
| Registered accounts | Cumulative accounts ever created | ~45 million | Bluesky |
| Monthly active users | Accounts active in 30 days | 12–27 million (disputed) | Third parties only |
| Daily active users | Accounts active in 24 hours | ~1.5–3 million | Third parties only |
Note the asymmetry: the number Bluesky publishes is the one that only ever goes up. Registered accounts is a cumulative counter – it cannot decline, because accounts are not removed from the total when people stop using them.
Where the Gap Came From
1. Migration waves recruit tourists
Bluesky's growth was not gradual. It came in enormous single-event spikes: the 2024 US election aftermath, the Brazil X ban, various X policy changes. Each drove millions of signups in days.
People who join during a news-driven exodus are joining against something rather than for something. A large fraction post an introduction, follow a few accounts, find their friends have not moved, and stop opening the app. The account stays on the ledger forever.
This is not unique to Bluesky – every platform that grew through migration has the same pattern – but Bluesky's growth was unusually concentrated in these events, so the effect is unusually large.
2. Bots and automation
An open protocol with free registration accumulates automated accounts: feed generators, mirrors, art bots, RSS relays, spam. They inflate the registered count and, depending on methodology, inflate MAU estimates too. Nobody knows the true share.
3. The first-week problem
Bluesky's onboarding weakness is well documented and well acknowledged. A chronological Following feed with nobody in it is an empty room, and the Discover feed historically was not good enough to fill the gap. Users who do not get past that first week never become active, and they are permanently counted as registered.
This is precisely why the 2026 roadmap led with the Discover feed and follow recommendations. It is a retention fix aimed at the top of the funnel.
4. Multiple accounts per person
Alts, project accounts, bot accounts run by humans. A meaningful share of registered accounts belong to people who already have another one.
Why the Ratio Keeps Getting Worse
Here is a piece of arithmetic worth internalizing before you read anyone's "Bluesky is collapsing" analysis.
Registrations grow by ~1.5 million a month. Daily actives are roughly flat. The ratio is DAU divided by registrations – so if the numerator holds steady and the denominator grows every month, the ratio declines mechanically, even when nothing about actual engagement has changed.
A headline reading "only 8% of Bluesky users are active, down from 14% last year" is describing successful user acquisition as if it were a failure. Both facts are real; the framing is wrong.
The number that actually tells you whether engagement is deteriorating is absolute DAU, not the ratio. And absolute DAU is below its June 2025 peak of ~4.1 million but broadly stable through 2026 – a genuine problem, but a different and smaller one than the ratio implies.
Is an 8–9% Ratio Bad?
It is low, and honesty requires saying so. Mature social networks run considerably higher.
Two pieces of context:
- Cumulative registration counters always look bad eventually. Any platform reporting lifetime signups against current activity produces a widening gap. Most platforms report MAU precisely to avoid this.
- The active core is unusually engaged. The few million people who stayed post, reply and read at high rates. Bluesky's reply-to-post ratio and the depth of its conversation compare well against much larger platforms. A small engaged network is a real thing, not a failed large one.
Which Number Should You Use?
| You are | Use | Because |
|---|---|---|
| Estimating reach for a campaign | DAU | It is the ceiling on who can see a post today |
| Comparing platforms | DAU or MAU, consistently | Never compare Bluesky registrations to another platform's MAU |
| Sizing infrastructure | Event volume, not accounts | Firehose throughput tracks activity – see our streaming guide |
| Citing an official figure | Registered accounts | The only number Bluesky publishes – label it clearly |
| Assessing platform health | Absolute DAU over time | Ratios mislead when the denominator is cumulative |
The Comparison That Actually Matters
The most common analytical error is comparing Bluesky's registered accounts to a competitor's monthly actives. That comparison is meaningless and it is everywhere, in both directions – boosters use it to make Bluesky look large, critics use the inverse to make it look empty.
Compare like to like. On DAU, Bluesky is a few million. That is a real, valuable, hard-to-reach audience with unusually low competition for attention. It is not a mass-market platform, and pretending otherwise sets everyone up for disappointment.
Frequently Asked Questions
How many real people use Bluesky?
Roughly a few million daily and somewhere in the low-to-mid tens of millions monthly, depending on methodology. Not 45 million.
Does Bluesky delete inactive accounts?
The registered count is cumulative and does not decrease as accounts go dormant.
Why doesn't Bluesky publish MAU?
It has never committed to doing so. Publishing MAU would invite direct comparison to much larger platforms on their strongest metric.
Is the gap unique to Bluesky?
No. Any platform reporting cumulative signups shows this. Bluesky's is wider than most because its growth was so concentrated in migration events.
Measure What's Actually Happening
- Trending Hashtags – live posting activity, not account counts
- Keyword Alerts – track real conversation volume
- Skyscraper for iOS – free Bluesky client