TL;DR / Key Takeaways
- A side-by-side comparison of the AI 'wait-state' ad platforms that appeared in June 2026 — Kickbacks.ai, IdleAds, Idlen, Sponsoric, and picoads.
- We rank them on revenue share, estimated CPM, integration model, and security risk, run the honest earnings math, and give a clear recommendation by user type.
- The takeaway: the rev-share percentage matters less than whether you own the surface the ad appears on.
Every AI coding agent shows a status line while it thinks — a spinner, a verb, a shimmer. "Wait-state" advertising is the practice of selling that moment to advertisers and sharing the revenue with whoever displays it. Kickbacks.ai made it famous in the second week of June 2026; by the end of that week there were at least five ways to do it, ranging from "patch your editor" to "embed an SDK in your own app" to "settle clicks in stablecoin." They look interchangeable. They are not.
The scorecard
| Platform | Dev share | Est. CPM | How it integrates | Risk |
|---|---|---|---|---|
| Kickbacks.ai | ~50% | $8–15 | Patches the Claude Code / Codex editor bundle; edits CLI settings | Medium–High |
| IdleAds.dev | 70% (goal 90%) | $8–14 | Claims server-side verification, no editor patching | Medium |
| Idlen.io | 70% | $20–42 | Widest surface: VS Code, Cursor, Windsurf + browser ext. (ChatGPT/Claude/Perplexity/Gemini) | Medium |
| Sponsoric | 70% | $15–35 | Publisher-native API/SDK — you embed it in your own app | Low |
| picoads | CPC, paid in USDC | — | Aimed at MCP / autonomous agents, not human-in-the-IDE | Low |
The platforms, one by one
Kickbacks.ai — the one that started it
The first mover and the loudest. Kickbacks replaces the Claude Code / Codex spinner with an auctioned ad and pays ~50%. Its terminal path is clean (it edits the official `spinnerVerbs` and `statusLine` settings), but its VS Code path patches Anthropic's extension bundle, persistently weakens that bundle's Content-Security-Policy even after you disable it, and auto-updates every 90 seconds with no signature verification. Best launch story, lowest rev share, highest platform risk. Our full profile is on the Kickbacks.ai page, and the strategic backstory is in Get paid for waiting: is the spinner gold rush real?.
IdleAds.dev — the cleaner clone
Launched within 24 hours of Kickbacks, IdleAds pitches a 70% share (with 90% as a stated goal) and a technically cleaner approach: server-side impression verification and, it claims, no patching of the editor. Worth one caveat about trust: a comment in the Kickbacks Hacker News thread plugging IdleAds came from the same username as its founder, with no disclosure — first-degree astroturfing in a market 24 hours old. The product may be cleaner; the launch wasn't.
Idlen.io — the widest net (and the biggest data ask)
Idlen plays a different game: the broadest surface area of anyone, with extensions for VS Code, Cursor, and Windsurf plus browser extensions targeting ChatGPT, Claude, Perplexity, and Gemini. 70% share, and the highest estimated CPMs in the category. The trade-off: the team is fully anonymous, and to target ads, Idlen reads your projects' dependency files. When an anonymous product wants that much visibility into your workspace, the trust question changes shape — weigh the higher CPM against the wider access.
Sponsoric — the grown-up option
Sponsoric is the one that escapes the category's original sin. Instead of an extension that patches someone else's tool, it's a publisher-native API/SDK you embed in your own app. That single change fixes the two deepest problems: the audience is no longer the payee (you show ads to your users, not to yourself), and there's no landlord who can evict you, because you own the surface. 70% share, lower risk, but it's only relevant if you're building an AI app — not if you just want your personal Claude Code to pay rent.
picoads — the agent-native, crypto-rail outlier
picoads targets a different buyer entirely: MCP servers and autonomous agents rather than humans watching an IDE. It prices on clicks (CPC) and settles in USDC. If your "developer" is actually a fleet of agents doing work unattended, this is the only entrant designed for that world. If you're a person in a code editor, it's not for you.
The earnings math, honestly
Strip out the hype and the numbers are modest. The most-cited real data point is a Hacker News tester who logged 407 impressions in about three hours on Kickbacks for $4.43 — roughly an $11 CPM. Extrapolated to typical usage (about four hours of agent-heavy coding a day), independent modeling lands here:
- 1Kickbacks (50%): ~$21–36/month
- 2IdleAds (70%): ~$30–50/month
- 3Idlen (70%, higher CPM): ~$50–100+/month, especially if you run its widest surface
- 4Sponsoric (70%): depends entirely on your app's traffic, not your personal coding hours
Enjoying this? Get one like it in your inbox each morning.
one email a day · unsubscribe in two clicks · no third-party tracking
The risk the rev-share number hides
Two risks matter more than any percentage on the scorecard:
- 1Security. Anything that patches another vendor's product (Kickbacks' VS Code path) or auto-updates without signing introduces attack surface you don't control. The SDK/publisher-native options (Sponsoric) and the settings-only paths avoid this.
- 2The landlord. Tools built on Claude Code's surface live on Anthropic's land with no lease — and Anthropic ran a Super Bowl ad saying "Ads are coming to AI. But not to Claude," and has enforced its terms against third-party tools before. The publisher-native model (your own app) is the only one immune to this, because you own the surface.
So which should you use?
There's no universal winner — the right pick depends on who you are.
| If you are… | Best fit | Why |
|---|---|---|
| A solo dev on Claude Code daily, risk-tolerant | IdleAds (or Kickbacks) | Higher share than Kickbacks with a cleaner integration story; Kickbacks if you only want the reversible CLI path |
| Building your own AI app | Sponsoric (or Idlen SDK) | Publisher-native: you own the surface, the viewer isn't the payee, no landlord risk |
| Corporate / risk-averse | None yet — wait | Patching third-party tools and unsigned auto-updates won't pass a security review; wait for publisher-native, vendor-blessed options |
| Running MCP servers or autonomous agents | picoads | The only entrant designed for agent traffic, with CPC settled in USDC |
Frequently asked questions
What is a 'wait-state' ad platform?
A wait-state ad platform sells the status line an AI coding agent shows while it's working — the spinner, the 'thinking' verb — to advertisers, and shares the revenue with whoever displays it. The category was created in June 2026 by Kickbacks.ai and now includes IdleAds, Idlen, Sponsoric, and picoads.
Which AI wait-state ad platform pays the most?
On headline revenue share, IdleAds, Idlen, and Sponsoric all advertise ~70% versus Kickbacks' ~50%, and IdleAds names 90% as a goal. But effective earnings depend more on CPM and integration model than on the split: Idlen claims the highest CPMs ($20–42), and Sponsoric's publisher-native model can out-earn all of them — but only if you're embedding it in an app with real traffic, not running it on your personal editor.
Is it safe to install a spinner ad extension?
The risk varies by integration model. Extensions that patch another vendor's product (such as Kickbacks' VS Code path) or auto-update without signature verification add attack surface you don't control. Settings-only CLI paths and publisher-native SDKs (Sponsoric) are lower risk. Always check your employer's extension policy and confirm payouts are actually live — several launched with earnings accruing to a counter that couldn't yet be withdrawn.
Will these platforms last?
The editor-patching ones are fragile: they sit on surfaces owned by Anthropic and OpenAI, both of which can close or replicate the channel — Anthropic has publicly committed to keeping Claude ad-free and has enforced its terms against third-party tools before. The npm ecosystem ran this exact experiment in 2019 (the `funding` package) and the platform banned terminal ads within weeks, then shipped its own alternative. The publisher-native model (Sponsoric) is the most likely to survive, because it doesn't depend on a landlord's tolerance.