Tony Wang9 min readThe Most Prolific Poster on X Is a Machine — and the Runners-Up Sell Convenience-Store Snacks
We pulled 325,553 public X profiles and ranked them by lifetime posts. The top account has 131.6M posts. Ten of the next twelve are Japanese brand bots.
We pulled 325,553 public X (Twitter) profiles into a search index — bio, join date, follower and following counts, and lifetime post count for each — and did the most obvious thing you can do with a table like that: sorted it by who has posted the most.
The answer is not a person. It is barely even a rounding error away from being entirely non-human.
The top of the leaderboard is automated
Here are the twelve accounts with the highest lifetime post counts in the corpus, as of the July 2026 crawl:
| Rank | Account | Lifetime posts | Posts/day since joining | Followers |
|---|---|---|---|---|
| 1 | @grok | 131,619,111 | 133,515 | 8,860,012 |
| 2 | @akiko_lawson | 50,464,988 | 8,423 | 9,876,090 |
| 3 | @suntory | 37,461,057 | 6,311 | 2,632,033 |
| 4 | @subwayjp | 21,530,624 | 3,523 | 1,111,825 |
| 5 | @asahibeer_jp | 19,982,456 | 3,845 | 1,959,040 |
| 6 | @housefoods_now | 13,949,915 | 2,353 | 817,715 |
| 7 | @AmazonHelp | 13,787,483 | 2,258 | 558,317 |
| 8 | @sanyobussan | 11,836,119 | 2,140 | 282,466 |
| 9 | @gusto_official | 11,653,544 | 2,164 | 1,099,264 |
| 10 | @koikeya_cp | 9,430,158 | 2,087 | 1,056,122 |
| 11 | @mos_burger | 8,948,684 | 1,451 | 2,147,547 |
| 12 | @donki_donki | 8,819,627 | 1,487 | 1,334,910 |
@grok, xAI's chatbot, sits at the top with 131,619,111 posts — because it auto-replies to anyone who tags it, all day, forever. Divide by the days since it joined in November 2023 and you get 133,515 posts per day, or 1.55 every second, held up without a break for 986 days. No human maintains that. That one account is responsible for 2.8% of the 4.69 billion posts in the entire corpus.
Then it gets strange. Positions two through twelve are almost entirely Japanese consumer-brand accounts running reply automation — Lawson and FamilyMart (convenience stores), Suntory and Asahi (drinks), Subway and Mos Burger (fast food), Gusto (family restaurants), Don Quijote (discount retail), Koikeya (snacks). These are campaign accounts: follow us, quote the tweet, a bot replies instantly with your prize result. Japanese brands adopted this "instant-win" mechanic years before it was common elsewhere, and the post counters show it — each of these brands has out-posted every journalist, politician and celebrity in our sample by a factor of a thousand or more.
We flag it plainly because it's a real property of the data, not a scandal: automated accounts are legitimate, and they dominate the volume leaderboard so completely that a human doesn't appear until you are well down the list. "Who posts the most" is a question about software.
Volume and reach are barely related
The intuitive model — post more, get more followers — does not survive contact with the numbers. The cleanest way to see it is followers earned per post:
@grok converts effort to audience at 0.067 followers per post. Taylor Swift converts at 92,401 — she has posted 887 times in her life and is followed by 82 million people. That is a 1.37-million-fold difference in efficiency between the busiest account in the sample and one of the most-followed. They are not doing the same thing with different intensity; they are doing different things.
One honest caveat rides on this chart: the post count is what currently survives, not everything ever written. X's counter goes down when you delete, and several of these low-post celebrities have wiped their history at some point — Taylor Swift famously cleared her account in 2017. So the "few posts, vast following" end is partly a story about deletion, not just restraint. It cuts the same direction either way: for the highest-reach accounts, the running post total is close to irrelevant to how many people follow them.
Almost nobody posts much
Step back from the extremes and the median X account in our data is quiet. Lifetime posts divided by account age gives a posting rate, and the distribution is steep:
Half of all active accounts post fewer than 0.61 times a day averaged over their lifetime. Ninety percent post fewer than six. The firehose is produced by a tiny minority: the top 1% of accounts generated 35.3% of all posts, and the top 100 accounts alone — three hundredths of one percent of the sample — produced 12.3%. Posting on X is more concentrated than following: a small automated core writes most of it while the long tail lurks.
The blue check follows the volume
Here is the finding that lands somewhere between funny and pointed. Sort accounts by how much they post and look at how many carry X's blue check:
| Posting rate (lifetime) | Accounts | Blue-check share | Median followers |
|---|---|---|---|
| Under 1 / day | 199,271 | 11.4% | 796 |
| 1–5 / day | 90,379 | 27.2% | 9,778 |
| 5–20 / day | 28,973 | 42.1% | 28,356 |
| 20–100 / day | 6,374 | 50.3% | 71,762 |
| 100+ / day | 557 | 68.4% | 379,654 |
The check rises monotonically with posting volume. Among accounts posting 100 or more times a day for their entire existence — the machine-scale tier — 68.4% are verified, against a 19.4% baseline across the whole corpus. The badge that verification was originally supposed to provide (this is a real, notable human, not an automated account) now correlates most strongly with the accounts that post like automated accounts, because a paid subscription is exactly what a high-throughput brand or bot account buys. We are not claiming the badge is meaningless — only pointing out where, empirically, it clusters. (For the flip side of that story — how the check tracks willingness-to-pay rather than notability — that's a separate cut of the same data we may write up later.)
What this shows, and what it doesn't
What we think holds up: the volume leaderboard on X is dominated by automation, and not narrowly — a human doesn't crack the top of our list at all. Posting rate and follower count are close to independent, diverging by six orders of magnitude at the extremes. Posting is heavily concentrated in a small automated core. And the blue check rises cleanly with posting volume.
What limits it:
- This is a seeded corpus, not a random sample of X. The 325,553 accounts were discovered from independent sources — people notable enough for a Wikidata entry, public GitHub developers, startup founders, journalists — and then their live X profiles were pulled. It over-represents established, Western, developer-adjacent, and notable accounts and under-represents brand-new and purely-personal ones. So treat every absolute percentage as descriptive of this population, not of "X users" in general. The shape — automation dominating volume, volume decoupled from reach — is what travels.
- Post counts are net of deletions. X's counter decrements when posts are removed, bulk-delete tools are common, and the effect is largest for exactly the high-volume accounts. Read
postsas "currently surviving posts," never "everything ever written." - Follower and post counts are a single-snapshot read. X exposes exact integers here (we checked — these are true counts, not the rounded "8.9M" the web profile shows), but they drift over time and X has documented that its counters can momentarily report zero under load.
- "Machine" is an inference from behavior, not a label X provides. We call an account automated when it posts thousands of times a day, every day, for years — @grok, @AmazonHelp and the brand-campaign accounts fit that unambiguously, but we did not audit each account's tooling.
None of these change the picture at the top. The busiest account on X is a language model answering mentions, and the next tier down is convenience-store marketing software. The people are further down, posting a couple of times a day, if at all.
Sources
Methodology
We built a search index of 325,553 unique public X profiles, seeded from independent public sources and then refreshed against each account's live profile through Crawlora's X profile endpoint. Each record holds username, display name, bio, join date, follower count, following count, lifetime post count and verification flag. We pulled the full corpus by paginating the X users dataset across account-creation date windows, and reconciled the census against the dataset's facet counts — the pulled total matched the independent aggregate to within 0.01%, which is our check that the sample isn't silently truncated.
Posting rate is lifetime posts divided by days since the account's join date. Followers-per-post is followers divided by lifetime posts. "Machine-scale" is a behavioral threshold (100+ posts/day sustained over the account's life), not a platform label. Every figure is exact-integer data from the profile JSON, not the rounded values X's web profile displays; we verified this by confirming that fewer than 0.1% of high-follower accounts sit on round numbers, which rules out display-rounding artifacts.
Update, August 2026: a new discovery channel (Common Crawl) has since added 324,523 more profiles to this index with a very different composition.
We publish aggregates only — no individual profile records. Want to run your own cut? The how to scrape Twitter/X guide walks through the exact calls, the X users dataset is queryable directly, and pricing covers the free tier. For a related study on X's written reviews, see what app reviews say about X before and after the Musk acquisition.
Frequently asked questions
Which X account has posted the most times?
In our corpus of 325,553 public X profiles, the most prolific account is @grok, xAI's chatbot, with 131,619,111 lifetime posts as of the July 2026 crawl — about 1.55 posts per second sustained since it launched in November 2023, because it auto-replies to everyone who tags it. It alone accounts for 2.8% of every post in the dataset. The next-highest accounts are almost all automated Japanese consumer-brand campaign accounts: Lawson (50.5M posts), Suntory (37.5M), Subway Japan (21.5M) and Asahi (20.0M).
Does posting more on X get you more followers?
Barely. Posting volume and follower count are close to independent in the data. @grok earns 0.067 followers per post; Taylor Swift, who has posted 887 times, earns 92,401 per post and is followed by 82 million people — a 1.37-million-fold gap in efficiency. The highest-volume accounts are automation, and the highest-follower accounts post rarely, so the running post total tells you very little about reach.
How often does the average X account post?
Rarely. Across the 318,263 accounts in our sample with at least one post, the median posts fewer than 0.61 times per day averaged over the account's life, and 90% post fewer than six times a day. Posting is highly concentrated: the top 1% of accounts produced 35.3% of all posts, and the top 100 accounts alone — 0.03% of the sample — produced 12.3%.
Are verified (blue-check) X accounts more likely to be automated?
Verification rises steeply with posting volume. Among accounts posting 100 or more times a day for their entire life — the machine-scale tier — 68.4% carry X's blue check, versus a 19.4% baseline across the corpus. A paid subscription is exactly what a high-throughput brand or bot account buys, so the badge now clusters in the accounts that post like automation rather than distinguishing humans from it.
Is this a representative sample of all X users?
No. The 325,553 accounts were seeded from independent public sources — people notable enough for a Wikidata entity, public GitHub developers, startup founders and journalists — then refreshed against their live X profiles. It over-represents established, Western and developer-adjacent accounts, so the absolute percentages describe this population, not X overall. The shape of the finding — automation dominating post volume, and volume decoupled from reach — is what generalizes. Post counts are also net of deletions, so they reflect currently surviving posts, not everything ever written.