Our One — The Math of the Commons
For the skeptics. For the practical minds. For everyone who needs more than a manifesto.
Before you read this
We know what you are thinking.
You have heard utopian arguments about technology before. You watched the early open-source advocates promise that freely shared code would liberate humanity. You watched the Web 2.0 evangelists promise that user-generated content would democratize media. You watched the blockchain believers promise that decentralized ownership would end financial inequality. You watched all of them get absorbed into the existing power structure or collapse under their own contradictions.
You are right to be skeptical.
So we are not going to ask you to believe in anything.
We are going to show you the arithmetic.
What it actually costs to run a social network
Let us start with the question no incumbent wants you to ask: what does it actually cost, technically, to connect a human being to a global social network?
Not what Meta spends. What it costs.
At the scale of 100 million users, a text-and-image social network running on modern cloud infrastructure — compute, storage, database, CDN, bandwidth — costs somewhere between $0.50 and $2.00 per user per year if you build it honestly. A well-architected platform with a small, skilled engineering team, using open-source infrastructure and modern tooling, sits at the lower end of that range.
Even at the generous end — $2.00 per user per year — that is 17 cents a month.
For a billion users, economies of scale push the cost lower still. The incremental cost of adding the billionth user to a well-built platform is close to zero.
This is not speculation. You can triangulate it from public data:
Wikipedia serves roughly 20 billion page views a month — more traffic than most social networks — with an annual hosting budget of around $3 million. That is a fraction of a cent per user.
Mastodon and Bluesky server operators — people running social networks for thousands of users on their own infrastructure — consistently report costs of between $0.10 and $0.30 per user per month. That is $1.20 to $3.60 per year, and that includes a generous hardware margin for a small operator without the scale advantages of a large platform.
Discord serves 200 million monthly active users on a relatively lean infrastructure team. Their engineering blog has described architectural choices specifically designed to keep per-user costs minimal.
The conclusion is not controversial among engineers. Infrastructure at scale is cheap. The hard part of running a network is not the servers. It is the people, the moderation, the governance, and the trust.
What Meta actually charges you
Meta does not charge you a subscription fee. You use Facebook, Instagram, and WhatsApp at no direct financial cost.
But you pay.
In 2024, Meta generated $164.5 billion in revenue. Virtually all of it — 97% — came from advertising. There are approximately 3.3 billion daily active users across their platforms.
That works out to roughly $50 per user per year globally.
In the United States and Canada, the number is dramatically higher. American users generate approximately $68 in revenue per quarter, which is over $270 per American per year.
These figures come directly from Meta's quarterly earnings reports. They are not estimates.
To put them in context: the infrastructure to serve those users costs roughly $1 per user per year. Meta collects $50.
The extraction premium
The difference between those two numbers — $1 in infrastructure cost versus $50 in revenue collected — is what we call the extraction premium.
The $49 difference is not pure profit. Meta employs roughly 70,000 people, many of them extremely well paid. It spends tens of billions on AI research, metaverse development, data centers, legal teams, lobbying, marketing, and executive compensation. In 2024, Meta's total expenses were approximately $95 billion.
But here is what matters: when you strip away the advertising infrastructure (the machinery required to track, profile, and auction your attention), the lobbying and regulatory teams (required to prevent accountability for the tracking and profiling), the metaverse division (a separate bet that has lost over $50 billion and generated almost nothing), the executive compensation, and the layers of corporate overhead that exist primarily to maintain and grow the extraction operation itself — what remains is a relatively small, efficient operation providing a communication service to billions of people.
That efficient operation, without the extraction machinery, costs a small fraction of what Meta spends.
This is not conjecture. It is what the numbers look like when you separate infrastructure from extraction.
What this means for Our One
Our One products do not run advertising. They do not track users across the web. They do not maintain behavioral profiles for auction. They do not employ the machinery of surveillance.
This means the cost structure is radically different.
Consider a professional network at the scale of LinkedIn — roughly 1 billion users. LinkedIn is owned by Microsoft and generates several billion dollars per year in revenue, primarily from premium subscriptions and hiring tools.
An Our One equivalent, built on honest infrastructure and maintained by a skilled steward team of perhaps 200 people (not LinkedIn's 20,000+), would have costs that look roughly like this:
| Cost category | Annual estimate (1B users) |
|---|---|
| Cloud infrastructure | $600M — $1B |
| Steward team (200 people, competitive salaries) | $40M — $60M |
| Security, legal, governance | $20M — $30M |
| Reserve and improvement fund | $40M — $60M |
| Total | ~$700M — $1.15B |
Per user, that is $0.70 to $1.15 per year. Less than $0.10 per month.
Even with a generous buffer for unexpected costs, a community-owned professional network at LinkedIn scale could sustain itself on $2 to $3 per user per year — a small transparent subscription that most working adults would not notice — and remain healthier, better-funded, and more resilient than the advertising-dependent model it replaces.
At 100 million users, the numbers are proportionally similar. The costs are lower. The required subscription is comparable.
The comparison that matters
Let us make this concrete.
You currently pay LinkedIn approximately $40 per month for a premium subscription if you use it for job searching. That is $480 per year.
Or you use the free version and pay with your data, your attention, and the behavioral information that LinkedIn (and its parent Microsoft) uses to train AI systems and sell advertising.
An Our One professional network would cost you between $2 and $3 per year to sustain at full scale — and you would own it. Not metaphorically. You would have constitutional protections that prevent it from being sold, from being converted to advertising, from changing the rules that protect you.
The difference is not between free and paid. The difference is between being extracted from and being an owner.
The stewards get paid
We want to be direct about this because the cooperative model often fails here.
Our One stewards are not volunteers. They are not underpaid idealists held together by mission alone. The economic model explicitly includes competitive compensation for skilled people.
A steward team of 200 professionals — engineers, designers, product leads, governance specialists, security experts — at competitive market salaries might cost $50-60 million per year. That is a real number, and a sustainable one, funded by transparent user contributions rather than hidden extraction.
The comparison to existing companies is instructive. LinkedIn has 20,000 employees. Meta has 70,000. Most of that headcount does not make the product better for users. It makes the advertising machine more efficient, the lobbying operation more effective, and the extraction operation more profitable.
A constitutionally bound product does not need that machinery. It can spend its resources on the actual product, on the actual users, and on the people who actually maintain it.
What users get that they cannot get elsewhere
An Our One product funded by transparent community contributions, not advertising, offers things that advertising-funded products structurally cannot:
No algorithmic manipulation. The algorithm's only job is to show you what is relevant. Not to maximize time-on-platform. Not to provoke engagement through outrage. Not to identify your psychological vulnerabilities and exploit them for advertiser benefit. The incentive to manipulate is gone because there are no advertisers to please.
Real privacy. No behavioral tracking. No cross-site profiles. No sale of data to third parties. Not because the founders are virtuous, but because the constitution prohibits it and there is no financial incentive to do it anyway.
Genuine ownership. If the product is ever sold, you would need to consent. The constitution protects against quiet conversion into private property. Your investment — of time, relationships, content, and trust — cannot be taken from you by a board vote you were not part of.
An aligned team. Stewards succeed when users are genuinely well-served. Not when engagement metrics are optimized. Not when ad revenue per user is maximized. When the product is actually useful, trustworthy, and worth paying for. The incentives point the same direction as the users' interests.
The honest limits
We will not pretend this is simple.
Network effects are real. Building a new network from zero when everyone is already on the incumbent is hard. The first years of any Our One product will be the hardest, not because the economics are wrong, but because adoption curves are slow and network value is low until enough people arrive.
Community governance is genuinely complex. Not as complex as the Web3 vision of voting on every decision, which is why we have stewards for daily operations. But constitutional governance — setting limits, ratifying major changes, holding stewards accountable — requires active community participation that some people will find more demanding than passive consumption.
And some products — particularly those with heavy video infrastructure and AI features — have higher per-user costs than the simple estimates above. Those products require transparent community conversations about what is actually worth paying for and how much.
None of these are insurmountable. All of them require honesty, which is exactly what the existing model does not provide.
What we do not promise
We do not promise every user a fantasy income for participating. We do not promise that every product can be free forever. We do not promise that all communities monetize the same way. We do not promise that governance removes all conflict.
We promise something stronger: the economics will be honest. The rights will be clear. The surplus will not be captured by default. The users will not be converted into a resource mine for the convenience of distant owners.
The question worth asking
When you look at these numbers — $1 in infrastructure versus $50 in extraction, $3 per year to own versus $270 per year being taken — the question is no longer whether user-owned, constitutional products are economically viable.
The question is why anyone would choose the alternative, if given a real choice.
The answer, for twenty years, has been: because there was no real choice. The network was already there. The product was already built. The switching cost was too high. The alternative did not exist.
We are building the alternative.
The math has always worked. What was missing was the will to do it, and the tools to do it cheaply enough.
The will exists. The tools exist.
What remains is the building.