X Ended Creator Revenue Sharing. The Word It Deleted Is the Whole Story

A creator's phone showing the X app open to a post analytics screen, with the X logo visible on a laptop in the background

X is replacing Creator Revenue Sharing with the Original Content Rewards Program on September 8. The new program pays on impressions from Premium subscribers, some of whom pay specifically not to see ads. X stopped funding creator payouts out of advertising revenue in November 2024. What it just removed was the last stated percentage of anything.

Every creator payout program eventually reaches the same fork. You can keep the formula and police the content, or you can keep the content policy and quietly stop publishing the formula. X just picked the second one, and it announced the choice as a quality upgrade.

On August 7, X told creators it was winding down Creator Revenue Sharing and launching the Original Content Rewards Program in its place. The coverage ran the obvious frame: platform cracks down on slop, rewards real work. That frame is not wrong. It is just the least interesting true thing about the announcement.

Here is the more useful reading. Since 2023, X has moved creator payouts three times, and each move put more distance between what creators get paid and any number a creator can independently check. The 2023 version paid a share of advertising revenue from ads served in your replies. The 2024 version paid out of Premium subscription payments instead. The 2026 version pays “qualified impressions” at a rate X does not disclose, judged by a model X says it will keep tightening.

The word that disappeared from the program name is the same word that disappeared from the payout math.

What Happened

X stopped accepting new entrants into Creator Revenue Sharing immediately on August 7. Existing participants keep earning through September 7 and receive three final payouts, on August 14, August 28 and September 11. From September 8 they can apply to the Original Content Rewards Program, with the first payment under the new system landing around September 25. Payouts stay on a two-week cycle.

The eligibility rules:

RequirementDetail
Age and location18 or older, account in good standing, in an eligible country
SubscriptionActive X Premium Basic, Premium, Premium+ or Premium Business
FollowersAt least 500 verified followers
ReachAt least 500,000 qualified Home Timeline impressions in 90 days
ContentMust clear X’s originality standard

A “qualified impression” is doing heavy lifting in that table. X defines it as a unique impression, from a Premium subscriber, on the Home Timeline, where at least half the post was visible. Replies are excluded. Duplicate, paid, promoted and artificially generated impressions do not count.

On the content side, X will pay for original reporting and analysis, photos and video you shot, graphics and memes you designed, and commentary that adds something. It will not pay for posts copied from another account, media downloaded and reuploaded, aggregated feeds, automated posting, or reposts without what X calls “meaningful original value.” Accounts whose output is mostly monetization coaching are also out.

X’s Allegra Jacchia said the old program’s incentives had become misaligned, and that adding more rules and exceptions was the worse option compared with starting over.

The Backstory

The program launched in July 2023 as a genuine advertising revenue share. Ads ran in the replies to your posts, X counted the ones served to Premium subscribers, and you got a cut. Creators could at least see the inventory their money came from.

That lasted about fifteen months. On October 9, 2024, X announced its largest change to date: payouts would no longer be tied to ads in replies at all, and would instead be calculated from engagement with your content by Premium users. The stated pool became a slice of Premium subscription payments, reported at the time as up to 25%, with engagement from higher tiers worth more. The change took effect November 8, 2024.

The reason was structural. X had made Premium+ fully ad-free in August 2024. The more successfully X sold subscriptions, the more of its most engaged audience saw no ads, and the thinner the reply-ad pool got for exactly the creators driving the most engagement. Rebasing payouts onto subscription revenue solved that. It also handed creators a new instruction: generate replies from paying subscribers, by any means available. TechCrunch flagged the rage-bait risk the day it was announced.

Then came 2026 and three attempts to patch the thing from inside:

  • March 3. X said accounts posting undisclosed AI-generated content about armed conflict would lose payout access for 90 days.
  • March 24. Head of Product Nikita Bier announced that impressions from a creator’s home region would count for more, to stop creators worldwide from farming US and Japanese attention. Creators outside those markets revolted within hours. Musk paused it the same night.
  • April 12. X said it was cutting payments to aggregator and clickbait accounts. Popular accounts profiting from the existing rules complained again.

Three patches in five months, one reversed before it shipped. That is the context for the decision to scrap the program rather than keep amending it, and it is a fair defense of the move. What it does not explain is why the replacement stopped publishing a rate.

The Plan

Read the new rules as a set of instructions to creators rather than as a content policy and the design gets clearer.

Only impressions from paying Premium subscribers count. Engagement from free users, however large, is worth nothing. You must be a paying subscriber yourself to participate. You need 500 verified followers, which on X means 500 followers who are themselves paying subscribers. Under the previous system, higher subscription tiers were explicitly worth more per engagement.

Stack those and the job description is not “make good content.” It is “acquire and retain X’s paying customers, and get them to look at your posts on the main feed.” X is paying a commission on subscription reach and calling it a content program. The originality gate is quality control on the sales force.

There is a second thing originality buys, and this one is a reading rather than a stated fact. X now sits inside SpaceX, which trains Grok on the platform’s corpus. A rule set that refuses to pay for copies, reuploads, aggregation, bot output and untransformed reposts is, functionally, a specification for deduplicated, human-authored, provenance-clean text and media. That is what a data buyer would write. X has not framed it that way and may never. But the incentives point the same direction whether or not anyone says so out loud.

The Business Model Angle

Two numbers explain why the formula went away.

The first is the pool. X’s advertising revenue was $367 million in the second quarter of 2026, disclosed for the first time inside SpaceX’s AI segment. That is down from $426 million a year earlier and roughly a third of the $1.08 billion Twitter reported for the same quarter in 2022, the last one before Musk’s takeover.

Bar chart of X quarterly advertising revenue showing $1,080M in Q2 2022, $426M in Q2 2025, $343M in Q1 2026 and $367M in Q2 2026, a 66% decline from the last quarter Twitter reported as a public company

The second is the audience. SpaceX’s filing put X at 4.4 million paid subscribers as of March 31, 2026, against a platform X describes in the hundreds of millions of monthly users. Premium is under one percent of the base. And the top tier of that one percent pays to see no advertising at all.

Put those together and you get the structural problem X has been managing since 2024. The users whose attention determines creator payouts are disproportionately the users who generate the least advertising revenue. Under a true revenue share that is arithmetically impossible: no revenue on an impression means no share of it. Under a rewards program it is fine, because the rate is whatever the platform decides it is. That is not a side effect of the rename. It is what the rename makes possible.

Now the part nobody in the coverage has run. Compare the two eligibility bars honestly.

The old threshold was 5,000,000 organic impressions in three months. The new one is 500,000 qualified impressions in 90 days. On the surface, a 90% cut. But the denominators are not the same animal. To bank one qualified impression you need a unique view, from a paying subscriber, on the Home Timeline, with replies excluded and at least half the post on screen.

Work the ratio. If Premium subscribers are under one percent of users but consume, say, ten times as much timeline as an average user, they account for roughly six or seven percent of Home Timeline impressions. Strip out repeat views from the same account, strip out partial views, strip out the entire reply surface that the original program was built on, and 500,000 qualified impressions plausibly requires somewhere between 7 million and 15 million raw ones.

The old bar was 5 million raw. The headline number fell by 90% and the actual bar probably went up.

That estimate rests on an assumption X could settle in one sentence and has not: what share of Home Timeline impressions is delivered to Premium accounts. If the real figure is closer to 20%, the new bar is roughly equivalent to the old one rather than tougher. If it is closer to 5%, the new bar is meaningfully harder. Either way, the interesting thing is that the number cannot be checked from outside, which is the pattern this whole piece is about.

TikTok ran this sequence first. The Creator Fund launched as a fixed pool, diluted as more creators qualified, drew years of public complaints from creators watching per-view rates fall, and was replaced in 2023 by a program that paid more per view but narrowed what counted as a qualifying view and scored originality directly. Platforms rarely announce a rate cut. They redefine the unit.

The Risk

The strongest case against this reading is that X’s stated reason is simply true.

The October 2024 change did produce exactly the engagement-farming problem observers predicted on day one, because it told creators that replies from subscribers were the unit of account. The March and April 2026 patches were real attempts to fix it inside the existing formula, and one of them was politically dead within hours. At some point “start over” is the correct engineering call, not a cover story. An originality gate is a defensible answer to a real problem, and refusing to pay for reuploaded video is straightforwardly good.

Discretion is not automatically worse for creators, either. TikTok’s move away from a fixed pool to a performance-scored rate raised reported per-view earnings by an order of magnitude. A rate the platform sets can go up.

Here is the falsifier, and it is a clean one. If X publishes a rate card, a stated pool percentage, or a named revenue line that Original Content Rewards draws from before applications open on September 8, the argument in this piece weakens considerably. Publishing a number is cheap if you have one. Twenty-two months of calling something “revenue sharing” while sharing no advertising revenue suggests the number is the part X would rather not print.

Quick Questions

Is the new program still funded by advertising? X has not said. The last published basis, from October 2024, was a share of Premium subscription payments rather than ad revenue. The new announcement names no revenue source and no percentage.

Does the lower impression threshold make it easier to qualify? Probably not. The threshold dropped from 5 million impressions to 500,000, but the definition narrowed to unique views from paying subscribers on the Home Timeline with replies excluded. On plausible assumptions the effective bar is flat to higher.

Do free users’ views count for anything? No. Only impressions from X Premium, Premium+ and Premium Business subscribers qualify, and only on the Home Timeline.

Does commentary on someone else’s news qualify as original? Yes, if you add something. X explicitly allows analysis and reaction. What it refuses to pay for is reposting or lightly editing material you did not make.

What happens to current participants? They earn through September 7 and receive final payouts on August 14, August 28 and September 11. Applications for the new program open to them September 8, with first payment around September 25.

Does passing X’s originality check mean I own the rights? No. Those are separate tests. Clearing the program’s originality standard says nothing about copyright in footage, images or text created by someone else.

The Business Model Analyst Take

Creator payouts are the clearest window into what a platform actually thinks it is buying.

YouTube has run the same 55% advertising split for close to twenty years, and that stability is not generosity. It is a contract creators can price against, which is why the supply side kept building businesses on top of it. A stated share transfers risk to the platform: when ad revenue falls, YouTube’s payout falls automatically and nobody has to announce anything. A rewards pool transfers the risk the other way. The rate only changes when the platform changes it, which means every future cut arrives as news, as a policy, and as a fight.

X has now made that trade three times in three years, each time trading formula transparency for operating freedom. In 2023 it bought a creator supply base with a real revenue share. In 2024 it repointed the payout at subscriptions because its ad-free tier had broken the original math. In 2026 it stopped naming a basis at all, right as the disclosed advertising line landed at $367 million and shrinking.

That does not make Original Content Rewards a bad program. It makes it an unpriceable one. If you build a business on X, you are now a supplier to a customer who sets the price after delivery, publishes no rate, and has said in writing that it intends to raise the bar over time. Price that risk accordingly, and treat anything you earn there as a bonus on top of an audience you can reach some other way.

The tell will come in October, when the first full cycle of payouts lands. Watch whether creators publish their effective rate per thousand qualified impressions, and whether X lets them.

UNLOCK THIS FREE DOWNLOAD

DOWNLOAD NOW

Fill Your E-mail to Receive this Download Directly in Your Inbox.

RECEIVE OUR UPDATES

The Biz Model Club

Get daily, no-fluff insights on the latest business models, startup strategies, and trends delivered straight to your inbox.