The clone economy is already dying

The clone economy is already dying

As of 16 September 2026. We said most of these domains would be parked by winter. The dying started earlier than that — the week the copies outnumbered the spectators.

We named the clone economy when Outbid.lol printed six figures and hundreds of tables arrived the same week. We asked how long it would last and split the clocks: the swarm, the original, the factory.

Four weeks after launch, the swarm is already showing the tell. Not a crash. A quiet empty. Boards still load. Stripe still works. Nobody is bidding.

That is what death looks like in the clone economy. The site does not go down. The screenshot stops happening.

The numbers were a eulogy

A live census put the copywave at 448 verified pay-to-rank boards. Revenue could be read on 310 clones. The median of those copies had collected $17.50. More than four out of five stayed under $100. Twenty-nine had taken nothing.

The average looked less embarrassing — about $206 — because a handful of early mutants pulled it upward. The richest measurable clone was roughly 39% of all clone revenue. The top ten took about two-thirds. The top fifty took more than 90%.

That is not a category. That is a lottery with 400 losing tickets and a table that still says #1.

Timing made it worse. Among clones with usable registration data, the earliest quarter of the wave collected thousands. The last quarter collected a rounding error. The median fell from about $25 for the people who shipped in the first hours to about $6 for the people who arrived when the kits were already on sale.

The whole wave happened in roughly two days. Being several hours late was already enough to miss the only thing that mattered: the crowd.

A smaller sample of boards that publish the actual price of first place was bleaker. Nine had never taken a dollar. The median top spot was $3.50. First place in front of nobody is worth nobody, whatever the rank card says.

The tells are already on the page

You do not need a scraper to see it. Open a late clone.

• The #1 row has not moved since launch week. The race ended when one wallet sat above the rest — or when nobody showed up.
• The live counter is a number from August. It is dressed up as a pulse.
• The "recent bids" column is empty, or it is the operator bidding against themselves so the page does not look abandoned.
• The domain still resolves. The X account posted twice, then a "we shipped" thread, then silence.
• Recap accounts stopped launching clones and started dunking on them. That is the feed telling you the format is content now, not a market.

Stripe is still real. If you pay, your name probably appears. The shape is the same cheap-scam energy we wrote about in week one, except the FOMO ran out. Late money is not exit liquidity anymore. Late money is a donation to a parked idea.

Outbid itself cooled the way a hit cools. The first 24 hours were about $21,500. The second day exploded. By the end of the first week the live counter was past $200,000, and then the pace dropped. That is a viral attention event, not a $100,000-a-day business you can annualize. The original at least had the audience. The copies had the prompt.

Death is not the original going offline

People will point at Outbid still adding listings and say the clone economy is fine. That misses the clocks.

The original can keep a long tail. Daily boards, categories, a cheaper fight than the all-time #1 at five figures — that is a format trying to become inventory. Some advertisers already posted trials, demos, and click spikes that look like ads that worked. First-mover plus a million visitors is allowed to survive.

The economy we named was never the original. It was the industrial process: see a counter, prompt the table, buy a .lol, post your own counter, wait for a fool with a credit card. That process is already unprofitable for almost everyone who ran it.

When 80% of the copies make less than a hundred dollars, the factory is still running and the product is already dead. Makers will keep prompting boards the way people kept minting JPEGs after the rooms emptied. Throughput is not demand. AI is excellent at both. It does not know the difference.

Winter will make it obvious. Domains expire. Kits stop selling. The meta-directories will list ghosts. We do not have to wait for winter to call it. The dying started when clone #200 shipped into a feed that had already seen the joke.

What dies, what repeats

Most of these sites were never supposed to meet a user on a Tuesday. They were supposed to catch a screenshot. The screenshot moved. The tax is the same one we wrote down in September: wasted maker hours, wasted token hours, a cheaper "I shipped," a rarer "this is useful."

A few mutants will remain — luck boards, maps, caps, fights, anything that changed a constraint enough to teach you something by using it. Those were never the swarm. The swarm was cashiers.

The factory will outlive this format. The next one-sentence product that prints will get the same night of copies. The half-life will be just as short, maybe shorter, because the feed has now watched a clone economy die in public. People will still ship. They will just ship into a more tired joke.

I would not build on a row that exists because a chatbot could scaffold Stripe. And I would not read "the original is still up" as "the wave is still a business." We already lived through a cycle where minting was mistaken for making. This time the jpeg was a leaderboard. The floor went to zero on schedule.

Both a three-day cashier and a Tuesday product can make money. Only one is still software after the feed leaves.

That is also why we build ThinkReview the way we do: a copilot on the pull request — GitHub, GitLab, Azure DevOps, Bitbucket — meant to still be useful on a Tuesday, not a format that expires when the quote-tweets do. AI should help you judge what got shipped, not only help you ship the next clone. Pick a model in Model selection. Run it on a real diff.

As of 16 September 2026. Revenue and clone counts cited here are public snapshots from operator counters and third-party scrapers in the weeks after August 19, 2026. They move constantly and are not audited. This is part three of our clone-economy series — after the swarm and how long it lasts. Ready to review the next PR with a tool that is not a three-day cashier? Install ThinkReview.