Can You Open Source the Dollar?
OpenUSD aims to open-source stablecoin issuance. Every layer that opens strengthens the one layer that can't: the dollar.
Venture capitalists like to analyze technology in terms of waves. For them, each new wave is an opportunity to reposition, find interesting companies before anyone else starts paying attention, and generate precious alpha as a result.
The current wave, for instance, is the rise of agents, enabled by everything AI now contributes to software development. Before that came others: cloud computing, mobile internet, social media, B2B SaaS, the metaverse, and many more. Some produced great companies and outsized returns, while others proved disappointing.
In the early 2000s, one massive wave helped pave the way for a Silicon Valley renaissance: open source software. The dotcom crash of 2000 had been interpreted by many as a clear signal that there was no “there” there, that the idea you could build profitable businesses on the internet was merely an illusion. But then a small vanguard of software engineers started to share software libraries and lines of code with their peers for free, with license to reuse and modify. Over time, that movement contributed to both a massive collapse in the cost of software development and a dramatic improvement in the quality of code. It eventually fueled the rise of a new generation of startups and led to a decade during which, to borrow Marc Andreessen’s words, software would eat the world. A massive success!
One reason open source became a movement is that, in 2004, Tim O’Reilly, a respected publisher and conference host, brought it into perspective by explaining why sharing code for free made economic sense. His landmark article Open Source Paradigm Shift, based on a keynote he first delivered at the annual technology conference of private equity firm Warburg Pincus, drew on historical parallels and business strategy concepts.
Above all, it explained how you could still make money from something shared for free: open sourcing a layer in any value chain commoditizes that link and displaces competition toward another layer where one or more players hold a sustainable advantage. If you have something that others don’t, your best strategic move is to open source an adjacent layer and force everyone to compete with you within the layer where you hold the upper hand.
O’Reilly was writing about software, but the logic he described is indifferent to its raw material. One could say it even applies to money in general, and stablecoins in particular. And this is exactly the reason why we have been revisiting O’Reilly’s landmark article recently. On June 30, 2026, a consortium of more than 140 companies, including Visa, Mastercard, Stripe, Coinbase, BlackRock, and BNY, announced OpenUSD: a stablecoin nobody owns, with no fees to mint or redeem, no volume caps, and the reserve yield handed back to partners rather than kept by a single issuer. In other words, an open source stablecoin.
The announcement immediately drew two reactions. First, the trade press read it as a shot at Tether and Circle, and the market agreed with unusual speed, sending Circle’s stock down more than 17% within hours. Then came the backlash: several named “partners” turned out to have signed nothing; the governance is a curated board of incumbents, not a commons anyone can join by showing up; the word ‘open’, critics argued, is doing marketing work rather than describing a structure.
Both reactions are fair, but they are also too small. Whether OpenUSD ships, stalls, or collapses under its own consortium politics is a question about one product only. What we’re interested in here is the broader perspective. Stablecoins were already open-source money of sorts: they run on permissionless, open-source blockchains — Ethereum, Solana, and their peers — that nobody owns and anybody can build on. By contrast, what had not been open until now was the layer immediately above the rails: issuance, that is, where Tether and Circle sit.
Whatever OpenUSD’s fate, it is the first serious attempt to push the openness one layer up the stack, and to make issuance open source. The question worth asking is not whether this particular consortium succeeds. It is: how far up does the openness travel — and what happens to the money underneath it, and to every currency that refuses to play the same game?
The commoditization play, twenty years on schedule
When O’Reilly wrote his essay back in 2004, the goal was merely to explain why Linux was eating the software industry from below. His core claim was that software turns into a commodity the moment it has to speak a shared protocol, whether HTTP, SMTP, or TCP/IP.
Two decades on, the article reads as a template for building a competitive advantage on top of a commoditized layer in any value chain. A commodity, by definition, is something interchangeable, sourced from more than one producer, that a customer can swap for a rival without noticing the difference. Once a layer becomes a commodity in this sense, the profits don’t disappear into thin air; instead, they typically relocate to an adjacent layer.
O’Reilly borrowed the mechanism of that relocation from the late Clayton Christensen, then a professor at Harvard University and the famed author of the book The Innovator’s Dilemma. One of Christensen’s well-known theoretical concepts, which he labeled the “law of conservation of attractive profits”, holds that when a product becomes modular and commoditized at one stage of a value chain, attractive profits reappear at an adjacent stage.
This happened several times in the history of computing: IBM commoditized the personal computer in 1981 by opening its architecture to clone manufacturers, then lost control of the industry to Microsoft, which owned the layer just above the hardware. Dell later commoditized the hardware itself and won the market by embracing thin margins rather than fighting them. And of course, open source software later did the same to proprietary software. By commoditizing the act of writing code, it shifted competition to the adjacent, user-facing layer, where firms compete on user experience and network effects.
Stablecoin issuance today sits where proprietary software sat in 2003: it is closed, and it is concentrated, with Tether and Circle holding roughly 83% of the market between them. Their business model is disarmingly simple: take deposits, buy (mostly) short-dated Treasuries, keep the yield — whether to fund other ventures as Tether does, or to redistribute it to partners as a marketing fee as Circle does. Operating the issuance layer is lucrative precisely because it has not been commoditized: USDT and USDC are not perfectly interchangeable, and switching costs in integrations, liquidity, and compliance let both issuers hold on to reserve income that a true protocol-level market would compete away.
That reserve margin is exactly what OpenUSD is aiming at. The consortium doesn’t even need to be truly open — a genuine commons where anyone can join and shape the rules — to do the damage. It only has to open the standard wide enough to break the current Tether–Circle duopoly. A fixed set of members who then compete against each other inside walls they built together would be enough to turn issuance into a commodity.
Of course, OpenUSD is, for now, only an announcement. Still, Circle’s stock fell the instant the market grasped which way issuance was heading — toward the losing side of Christensen’s law, where a service becomes a commodity and its margin collapses. Nothing about how issuance works has changed, and nothing has yet shipped; the market is simply pricing in what issuance might become. Watching investors reprice Circle in a single afternoon, on the strength of a press release, was watching the shift O’Reilly described twenty years ago arrive exactly on cue.
How far up does the openness go?
Again, the stablecoin rails were open from birth. Permissionless blockchains are open-source software in the strictest sense: public codebases, forkable, maintained by distributed communities, owned by no one. The monetary instruments that run on them, however, are the opposite of open source. Tether and Circle built proprietary stablecoin businesses on a commons — the oldest move in the open-source playbook, and a profitable one for as long as nobody opens the layer you occupy.
OpenUSD’s value proposition is to open that layer too. Even if it fails, the idea will not die with it; the economics that produced it — 83% concentration, reserve yield held by two firms, switching costs as the only moat — will produce another attempt with a different logo. Then there will be attempts to commoditize even beyond issuance.
The agent frameworks now being built to let software transact autonomously are, for the most part, already open source. Settlement and compliance infrastructure could follow the path of web servers. Wallets and distribution, in turn, could be commoditized in much the same way that web browsers were. In the end, if these software-dominated layers are all commoditized through open source, Christensen’s law suggests that profits will migrate to whichever adjacent layer still depends on scarce resources. That would confine scarcity to the physical world: robotics, logistics, energy, and the machines that carry out what agents decide and stablecoins settle. As stablecoins become open infrastructure, layer by layer, they will become embedded in those physical systems, serving as the default settlement rails for real world commerce while competition and value creation shift to the assets and operations that cannot be commoditized.
This leads to the final question. If the rails, issuance, wallets, settlement, compliance, and agent frameworks all succumb to commoditization through open source, what becomes of the currency itself? Is the dollar simply the next layer in the stack, or is it something fundamentally different?
