Three forces are simultaneously reshaping banking: AI as infrastructure risk, tokenization of the deposit base, and t...

Three forces are simultaneously reshaping banking: AI as infrastructure risk, tokenization of the deposit base, and the opacity building in private credit. Follow them to their logical conclusion and they start to describe the bank of 2030.

"It ain't what you don't know that gets you into trouble. It's what you know for sure that just ain't so."

The mistake is to treat them as separate agendas: tokenization as a liquidity solution, private credit as a contained exposure, and AI as a productivity tool. Each of those certainties deserves scrutiny.

Recent regulatory concern around AI models and cyber vulnerabilities shows why AI is moving from productivity topic to infrastructure-resilience issue: a model that can accelerate threat discovery can also accelerate systemic exposure.

Tokenization of the deposit base has moved past experimentation. Ervinas Janavicius, writing for Capco, draws on work with global banks to show the GENIUS Act gave executives "the signal many were waiting for to move from exploration to implementation."

Tokenization is the proposed solution to fragmented liquidity but without the underlying operating model, adoption adds a new layer of fragmentation: multiple chains, misaligned regulatory regimes, hybrid legacy and on-chain systems in parallel.

Private credit has grown rapidly into a multi-trillion-dollar market, but its exposures remain harder to see, monitor and model across the banking system. New supervisory checks on banks’ exposures, suggest regulators still lack sufficient visibility into layered leverage structures.

Bank of England Deputy Governor Sarah Breeden named the convergence: "The thing that really keeps me awake at night is the likelihood of a number of risks crystallising at the same time — a major macroeconomic shock, confidence in private credit goes, AI and other risky valuations readjust."

Most institutions are still managing each force separately. That gap is Macro-Blindness: the failure to see that risks managed separately can crystallise together.

Follow those forces forward, and a plausible 2030 bank starts to emerge.

The deposit base is programmable. Banks manage tokenized flows across settlement rails, with AI optimizing liquidity. Credit intermediation is increasingly shared with private markets, but the bank holds the position private credit cannot: regulated counterparty certainty in a fragmented system.

As Ervinas identifies, "early movers who operationalize multi-jurisdictional compliance orchestration will become the natural bridges." That position, and the fee income from settlement orchestration and trust licensing that comes with it, goes to institutions that treated 2026 as a design brief.

The bank of 2030 will be programmable, AI-operated, connected to private markets, and valued for its ability to provide trust across fragmented infrastructure. The institutions that capture that value will be the ones that treated these forces not as separate risks, but as a blueprint.