Three collective infrastructure projects past $1k — and why the payer is never who you expect

This week's qualifying cases come from a corner of Open Collective that the previous two issues barely touched: projects that aren't maintained by a named solo developer but by a loose committee, a specification body, or a two-person shop running something the whole web depends on. JSON Schema, Mocha, and jsDelivr all crossed $1k/month. None of them got there by convincing end users to donate.

A fourth candidate — Mage-OS, a Magento community fork with €31,790/year in estimated annual budget — doesn't qualify on the core criterion. Its Open Collective page shows €126,353 sitting in reserves with only €1,000 ever disbursed. That's an association treasury, not maintainer income. It's included as a diagnostic at the end.

JSON Schema is the specification that defines how JSON data structures are validated — the schema language behind OpenAPI, JSONForms, VS Code's IntelliSense, and dozens of other tools. It isn't a library or a framework; it's a standard. The spec is developed by a volunteer committee of four: Ben Hutton, Austin Wright, Benjamin Granados, and Amaechi Onyedikachi, operating through the json-schema-org GitHub organization. 1

~$2,134/month, based on an estimated annual budget of $25,603 on Open Collective as of June 11, 2026. The project runs under Open Source Collective as fiscal host. 1

The top all-time contributors are AsyncAPI Initiative ($20,250), Google Summer of Code ($7,000), and Airbnb ($5,800). One paid community manager position runs at $12,000/year. The tier structure runs from Individual Backer at $5/month up to Golden Sponsor at $1,000/month, with Bronze/Silver in between. 1

Notice who's on that list. AsyncAPI is a messaging specification project — it directly depends on JSON Schema to define its own schema syntax. Google Summer of Code routes funding to open infrastructure projects as part of its mentorship program. Airbnb uses JSON Schema internally and in open-source tooling it maintains. These aren't users who fell in love with the project. They're participants in a shared technical stack who calculated that an unmaintained or stagnant JSON Schema spec would cost them maintenance work on their own downstream tools.

The committee has also held multiple editions of JSON Schema Conference, which functions partly as a community development event and partly as a fundraising mechanism — enterprises attending or sponsoring the conference become aware of the project's funding needs in a direct way that a GitHub badge doesn't replicate.

Pure tiered Open Collective sponsorship, no commercial layer. The unlock here was structuring the project as a specification organization rather than a software project — holding a conference, publishing a roadmap, hiring a part-time community manager — which gives enterprise adopters a recognizable "this is a real institution" signal that justifies writing a check through a finance department.

If your project is a specification, protocol, or standard — something other projects depend on at the schema or contract layer — this is the most replicable template here. Corporate funders aren't paying for features. They're paying to keep the spec active because their own tooling would need rework if it stagnated.

If your project is a library rather than a spec, the same logic applies at smaller scale: do your users face institutional risk if you stop? A five-person startup won't cut a $1,000/month sponsorship check. A company with 500 engineers who've built tooling on top of yours might.

Mocha is a JavaScript test framework that's been around since 2012 — older than Jest, older than Vitest, older than most of the frameworks it's tested. It has 22,900+ GitHub stars, ~12 million weekly NPM downloads (per the project's own page), and 4,046 commits. The admin team currently includes Mark Wiemer and Josh Goldberg (who also administers TypeScript ESLint — covered in a prior issue at ~$600/month there). The project explicitly states it is "not endorsed or administered by the OpenJS Foundation," despite Mocha being an older tool in the JavaScript ecosystem. 23

~$2,037/month, based on an estimated annual budget of $24,447.74 on Open Collective as of June 11, 2026. 2

993 individual Backers at $5/month, and 9 organizations in the Sponsor tier at $500/month. The top all-time contributors are Localize ($38,000), clay ($21,500), Salesforce ($10,185), and Icons8 ($9,546). 2

The pattern is corporate test infrastructure: these are companies running large JavaScript CI pipelines where Mocha is the test runner. Localize is a localization SaaS company with engineering teams that presumably have thousands of Mocha tests. Salesforce has entire product lines with JavaScript test suites. These companies aren't sponsoring Mocha out of goodwill — they're paying to keep a tool their CI infrastructure depends on from going unmaintained.

The project has also adopted Collective Funds Guidelines 0.1, a published standard for how Open Collective funds are distributed — which, similar to JSON Schema's conference structure, signals to potential corporate funders that there is a governance framework around the money. 2

Tiered Open Collective sponsorship — 993 individual Backers plus a concentrated corporate Sponsor tier. No dual license, no hosted version, no pro features. The $2k/month figure comes primarily from 9 organizational sponsors, not from the nearly 1,000 individual backers.

Do the math: 993 × $5 = ~$4,965/year from individuals. 9 sponsors at $500/month = ~$54,000/year from organizations. The individual tier is roughly 9% of revenue.

Mocha is useful precisely because it's old. No viral growth, no recent launch. It's $2k/month because it's embedded in enterprise CI pipelines that won't swap it out without a painful migration.

If you maintain a mature tool enterprises run in production — something that's been in build scripts for years, not a shiny new library — corporate sponsorship can work without growth. The question is: does replacing this project cost a company more than $6,000/year in engineering time? If yes, you have a pitch.

One data transparency note: Josh Goldberg is an Admin on both Mocha and TypeScript ESLint (~$600/month, prior issue). Mocha's collective has no public breakdown of how funds are distributed among admins, so the $2,037/month figure refers to the project's collective, not any individual's take-home.

jsDelivr is a free public CDN for npm packages and GitHub repositories — the infrastructure layer that serves JavaScript libraries directly to browsers without requiring end-user bundling. It's been running since 2012. Dmitriy Akulov and Martin Kolarik run it as a two-person operation. As of June 2026, it serves over 100 billion requests per month and 4,500 TB of data. 4 Tracxn (March 7, 2026) confirms it as an unfunded company with no VC capital. 5

~$1,800/month, based on an estimated annual budget of $21,600 on Open Collective as of June 11, 2026. Current balance: $20,067.85. Total disbursed since inception: $2,422.27 — suggesting the two maintainers draw very little from the fund. 4

Google contributed $18,000 (starting January 2026). BairesDev contributed $5,850. The next-largest contributors are far smaller. 4

That $18,000 from Google is 83% of the estimated annual budget. This is the same concentration-risk structure that appeared with Actual Budget (74% Peakford Ltd) in the prior issue — one entity accounting for the majority of revenue. If Google stops contributing, jsDelivr drops from $1,800/month to roughly $300/month.

Why does Google fund jsDelivr? The clearest explanation: Google's own developer ecosystem (Angular, Polymer, and hundreds of community projects) benefits from a fast free CDN for npm packages. When developers prototype with a Google-adjacent library, pointing to cdn.jsdelivr.net is friction-free. A degraded jsDelivr is a degraded experience for the JavaScript ecosystem Google has significant stakes in. This is the same corporate-risk logic as core-js and Mocha, just concentrated into a single funder.

Open Collective sponsorship with a very flat tier structure (Backer at $10/month, Sponsor at $150/month). The $1,800/month figure is almost entirely a single large organizational contribution. The low disbursement ($2,422 total) suggests the maintainers are either running minimal expenses or haven't drawn personal income from it — the fund may serve primarily to keep the project's operational costs covered rather than fund maintainer time.

jsDelivr is instructive as a warning more than a template. Serving 100 billion requests/month on $1,800/month with 83% from one source is precarious. The $20,067.85 balance is roughly 11 months of runway if Google stops.

The question for any project considering this path: is there one company so large and so dependent that it will write a five-figure annual check? If yes, this can work. If the answer is "many companies benefit but none see us as a critical dependency" — you'll get 17 Backers at $10/month.

Mage-OS is a community fork of Magento Open Source, maintained by the Mage-OS Association, an 8-person nonprofit board led by Vinai Kopp. Its Open Collective page shows an estimated annual budget of €31,790 (~$2,850/month) — which would make it the highest-earning project in this issue. 6

It doesn't qualify. Total disbursed: €1,000. Total in reserves: €126,353.71. 6

That's an association reserve fund. The €31,790 annual budget figure reflects incoming contributions from enterprise partners (MDOQ at €20,000, three Silver Partners at €12,000 each), but the money sits. Nothing has been paid to maintainers in any material amount. Whether this is an intentional endowment strategy or a governance bottleneck, it's structurally different from what this series tracks — actual income reaching the people who build and maintain the project.

The Mage-OS case is useful precisely because it shows what "big Open Collective number" can look like when it doesn't correspond to maintainer pay. The disbursement rate is the check: if total disbursed over the project's lifetime is a small fraction of what's in the account, the money hasn't been used.

All three qualifying projects this week fit a single archetype — collective infrastructure — where the payer is a company that depends on the project as background infrastructure rather than a featured product. That's a different dynamic than the values-community funding (Molly IM) or dependency-risk corporate funding (core-js) from prior issues.

The pattern that threads through JSON Schema, Mocha, and jsDelivr: the funder isn't paying for access to a feature they want. They're paying to avoid a cost they don't want to absorb. A stagnant JSON Schema spec means rework downstream. A Mocha that stops getting security patches means a CI pipeline that can't upgrade. A slower jsDelivr means a worse developer experience for a platform that benefits from ecosystem health.

If your project can be framed as "you are currently relying on this without paying for it, and here is what it costs you if it breaks" — that's the pitch. Not all projects can make that argument honestly. But the ones that can are consistently the ones crossing $1k/month through corporate sponsorship.