The $1B Settlement Graveyard

xRapid asked banks to hold a volatile token, XRP, as a liquidity bridge. Libra/Diem proposed a basket-backed currency that triggered monetary-sovereignty alarm in every G7 capital. Bank consortia such as IBM World Wire, Fnality, and Contour ran out of runway before network effects kicked in. CBDC cross-border pilots addressed domestic modernization, not the actual cross-border problem: correspondent retreat, FX scarcity, and compliance fragmentation. The projects looked different on the surface but shared a common architectural failure: misalignment between what the project asked of counterparties and what counterparties needed.

xRapid's value proposition required banks to hold XRP as a liquidity bridge. This required banks to accept price volatility between payment initiation and settlement, take on XRP custody and operational risk, navigate unclear regulatory treatment of XRP in each jurisdiction, and replace a proven operational model with a novel one. The SEC's 2020 enforcement action alleging XRP was an unregistered security, partially rejected in 2023, added multi-year regulatory friction. Even without that action, the model asked banks to take risk they had other ways to avoid.

Most banks that engaged with Ripple used RippleNet, the messaging layer with no XRP required, rather than xRapid, the liquidity layer with XRP required. The messaging product found some adoption. The liquidity product, which was the economic core of the thesis, did not scale.

Libra's original design proposed a basket-backed currency that would function as a new unit of account. That is precisely what central banks cannot permit: an instrument that could challenge domestic currency as a unit of account directly attacks monetary sovereignty.

The eventual pivot to single-currency Diem coins addressed this but came too late. The regulatory damage was done, and the political coalition that had formed to block Libra did not disband. Payment infrastructure that positions itself as complementary to sovereign currencies gets regulatory tolerance. Infrastructure that positions itself as replacing sovereign currencies gets regulatory opposition.

Bank consortium projects face a specific cold-start problem. Every participating bank must integrate to the shared network, often as a multi-quarter project. Participants must agree on governance, compliance standards, and fee structures. The network must reach critical mass before it delivers meaningful value to any individual bank. During the build phase, each bank continues to operate its existing correspondent network, which competes with the new network for volume.

The result: many projects launch with three to five participating banks and limited corridors. They deliver some value at that scale but not enough to justify the investment. Additional banks join slowly because the value is limited. The network does not reach the inflection point where growth becomes self-sustaining.

CBDC projects were designed to modernize domestic payment systems. The cross-border benefit was theoretical: if multiple countries had CBDCs, inter-CBDC bridges could enable fast cross-border settlement.

But the core cross-border problem is not domestic modernization. It is correspondent banking retreat, FX scarcity, and compliance fragmentation. CBDCs address none of these directly. They provide a domestic digital cash instrument, which is useful domestically but does not solve the international settlement problem.

The eNaira is the clearest case: launched partly to reduce foreign currency dependence, it did not reduce USDT demand because USDT solved the cross-border problem while eNaira solved a different, domestic problem.

Every successful cross-border payment innovation has this property. Wise did not ask banks to hold a new asset; it built on top of local banking. SWIFT gpi did not replace correspondent banking; it added speed to existing relationships. Stablecoins did not create new currencies; they tokenized existing ones.

The graveyard projects all tried to replace something. The survivors all tried to upgrade something. This is the pattern future cross-border infrastructure designers should internalize.

The simplification case: USDC and USDT effectively replaced the function of money market funds and bank deposits for payment liquidity, and they scaled. Fnality and Partior tried to upgrade existing bank-to-bank settlement and have struggled to reach scale. So replace-versus-upgrade is not a clean predictor.

The counter-counter: USDC and USDT replaced an instrument category but kept the underlying claim structure intact. They are USD-denominated claims on issuers, structurally similar to money fund units in user experience. They did not ask users to accept a new monetary unit, which was Libra's mistake. Fnality's challenge is closer to bank consortium governance drag than to architectural mismatch. The dichotomy is a useful heuristic, not a deterministic rule.