There’s a quote that circulates constantly in XRP discussion: that Ripple co-founder Arthur Britto once said XRP was “designed to reach $10,000.” It resurfaces every time Britto — famously private, virtually silent in public for over a decade — so much as posts an emoji. There is no verified primary source for the statement. No paper, no interview, no on-record moment. It traces back to a secondhand 2019 video referencing an alleged 2017 remark, and it has never been confirmed by Britto, Ripple, or anyone who worked alongside him.
We’re not going to build on that quote. It isn’t sourced, and a price target was never the interesting part of the story regardless. The interesting part is what Britto and his collaborators actually built, why they built it that way, and what their design didn’t — and couldn’t — anticipate.
What the architects actually built
XRP doesn’t have a whitepaper in the Bitcoin or Ethereum sense — a document explaining the asset. What it has instead is The Ripple Protocol Consensus Algorithm, published in 2014 by David Schwartz, Noah Youngs, and Arthur Britto. It’s a paper about a settlement mechanism, not a monetary thesis. That framing matters, because it tells you what problem these three engineers thought they were solving.
The problem was structural, not speculative: cross-border payments were trapped by Nostro/Vostro accounts — pre-funded balances banks maintain in foreign currencies, sitting idle across a global correspondent banking network, just so money can move when needed. It’s capital-inefficient and slow. The XRP Ledger’s federated consensus model — validators reaching agreement every three to five seconds, without mining or staking — was built to let value move between any two currencies without that trapped capital. XRP was cast as the neutral bridge asset in the middle of that exchange.
Britto was reportedly direct about what this meant for the asset itself. In a 2013 internal note, he’s on record suggesting that XRP’s price was beside the point — what mattered was the spread, and that most institutions using the ledger would barely think about XRP as an asset at all. Whether or not he ever forecast $10,000, that earlier and better-documented statement is the actual thesis: XRP was infrastructure for moving value, not a target for holding it.
Designed for one class of counterparty
Here’s the part worth sitting with. The RPCA paper was written in 2014. On-Demand Liquidity, the product built on top of it, was designed around one kind of counterparty: financial institutions — banks, remittance firms, payment processors — moving money on behalf of people. The entire architecture assumes a human-initiated, human-approved transaction somewhere upstream. That wasn’t a limitation anyone would have flagged at the time, because it wasn’t a limitation. It was the only kind of economic actor that existed.
That’s not a flaw in the 2014 design. It’s just the edge of what was conceivable in 2014.
The counterparty nobody designed for
Twelve years later, a second class of counterparty has shown up: the autonomous agent. Software that pays for its own compute, settles its own invoices, and completes transactions without a human approving each one. Ripple’s own language on this, from its June 2026 launch of the XRPL AI Starter Kit, states plainly that most existing payment rails were built for people to initiate, approve, and reconcile — and that autonomous systems need something else entirely.
The starter kit is Ripple’s answer: an XRPL documentation server built on the Model Context Protocol, wallet and payment tooling for Claude and comparable agent frameworks, and support for the x402 machine-payment standard so an agent can pay for an API call or a model inference the same way it would settle any other transaction — in XRP or in RLUSD, XRPL’s dollar-denominated stablecoin. The same day, Mastercard named Ripple a launch partner in its own Agent Pay for Machines network, placing XRPL at the settlement layer of a card-rail giant’s agentic commerce stack.
Why the architecture holds up anyway
This is the part that actually vindicates the 2014 design rather than exposing its age. An autonomous agent needs almost exactly what a bank’s correspondent desk needed: settlement that finishes in seconds, not days; fees that are fixed and knowable in advance, not auctioned; and a way to move between assets without routing through an external bridge. XRPL’s federated consensus, its fixed transaction costs, and its native on-ledger DEX weren’t built with AI agents in mind — they were built to solve a liquidity problem for institutions. It turns out those same properties are precisely what a machine-native counterparty requires.
The original design wasn’t wrong. It was scoped to the transacting parties its architects could see. The omission wasn’t a design flaw — it was a horizon problem, and the horizon has since moved.
Where this leaves the treasury view
For a principal-only digital asset treasury, this distinction is the whole point. We’re not holding XRP because of a number attributed to a co-founder who never confirmed it. We’re holding it because the settlement properties that made it useful to correspondent banks in 2014 are the same properties now extending its utility to a second class of counterparty that didn’t exist when it was built. The thesis hasn’t changed. The set of participants it applies to just got larger.
This piece reflects the internal research and market view of Stellae Liquiditas LLC, a principal-only corporate digital asset treasury. Nothing here constitutes investment advice or a forecast of asset prices.