What Silvergate and Signature Taught Us About Relationship Risk

On 8 March 2023, Silvergate Bank announced it was winding down operations. It had been the primary banking partner for a significant portion of the US digital asset industry. Exchanges, payment businesses, and cross-border operators had come to depend on its infrastructure for daily settlement.

Four days later, Signature Bank was seized by regulators.

Silvergate and Signature had concentrated exposure to digital asset clients. For Silvergate, the FTX collapse in late 2022 was fatal. Signature followed weeks later, seized by regulators as panic spread through the banking sector in the wake of Silicon Valley Bank.

Since then, the picture has sharpened: the SEC settled charges against Silvergate's former executives in 2024 for misleading investors about the bank's compliance programme, and the US regulatory posture toward digital assets has shifted considerably — the GENIUS Act became law in July 2025. What looked like a contested regulatory story in 2023 has since resolved into something more straightforward: a concentration and governance failure, in a sector that was changing faster than the institutions serving it could manage.

Two major banking partners exited within four days, and the businesses that had alternatives fared better than those that didn't.

What happened in the aftermath was instructive. For years, these businesses had treated their banking relationships as infrastructure. They believed them to be stable, background, and permanent.

It’s a very human thing to do. We discount unlikely events because managing for them is costly, uncomfortable, and rarely rewarded until the moment it becomes essential. The businesses that had diversified their banking relationships before March 2023 looked cautious for years. Then they looked prescient for about a week. Then nobody talked about it anymore, because they were fine and we’d moved on to the next news cycle.

Correspondent banking has always worked this way. The relationships feel bilateral, but the risk is not. A bank extends access based on its strategy, regulatory posture, and read of the political environment. That frame is always shifting. Sometimes it shifts slowly, through a tightening of compliance requirements, longer onboarding timelines, more questions about transactions. Sometimes, it shifts in a week.

The businesses most exposed to this are the ones running high-volume, cross-border flows through a single banking relationship. Building that relationship took years, and once it was working, there was every reason to leave it alone. The relationship became the business. And when the relationship changed, so did the business.

Building redundancy early isn’t being pessimistic, it’s recognising that each counterparty has its own interests, constraints, and relationship with regulators.

What March 2023 did was compress the timeline. Events that might have played out over months or years, such as limiting access or a pull back of services, happened in days. The businesses that were already working on alternative rails weren't spared the disruption, but they had somewhere to go. The ones that weren't had to find somewhere to go under pressure, which is a different and worse problem.

Building redundancy isn't really about predicting the specific failure. It's about acknowledging that concentration itself is a form of risk.

The right moment to diversify correspondent relationships is before you need to, because the option to move has a cost, and that cost is lowest when there's no urgency.

---

If you'd like to explore alternatives before you need them, FiveWest provides licensed cross-border settlement infrastructure across African and Asian corridors.

Send us a message to get started.