When banks build custody infrastructure on Bitcoin, they replicate the fractional-reserve logic on the only asset with a mathematically fixed supply in history.
According to the Bitwise Crypto Market Review Q3 2026, individuals hold 66.1% of Bitcoin’s maximum supply – roughly 13.9 million BTC out of 21 million. Meanwhile, the Bitcoin Banking Adoption Index compiled by Strategy assigns 25 large financial institutions a composite score of 32% across five dimensions: custody, trading, investment products, lending, and management support. The standard reading of these figures is reassuring: banks are adapting, adoption is advancing, the system is learning. The structural reading is less comfortable.
The traditional banking model is built on fractional reserve: the bank collects deposits, holds a fraction in reserve, and lends or invests the rest. The depositor believes they own their money; in reality they own a claim against the bank. As long as everyone does not demand their funds back at the same time, the system works. When they do, it is called a bank run. This architecture rests on a silent premise: the underlying asset is expandable. If there is a liquidity shortage, the central bank can create new base money. The relief valve is always available.
Bitcoin removes that valve. 21 million units, decreasing issuance, no authority capable of altering the emission schedule. It is precisely this rigidity that makes Bitcoin interesting as a store of value – and precisely this rigidity that makes it structurally incompatible with the logic of fractional reserve. If a bank collects BTC in custody on behalf of third parties, it issues IOUs on an asset it cannot create. When redemptions exceed actual reserves, the valve does not exist. Only insolvency remains.
The standard response at this point is that banks operate under “segregated custody” and that clients retain “beneficial ownership”. Technically accurate, at least in the early phase. But the history of finance shows a recurring trajectory: it starts with pure custody, then securities lending on custodied assets is added, then derivatives, then leverage. Banks can monetise custody, lending, and trading relationships even when the client formally retains beneficial ownership of the BTC involved. The economic incentive systematically pushes toward greater use of the custodied asset, regardless of the good faith of the parties involved.
There is also the dimension of individual sovereignty, which the debate on institutional adoption tends to ignore almost out of reticence. Bitcoin was designed as a peer-to-peer transfer system that requires no trust in an intermediary. Entrusting one’s BTC to a bank means giving up its most relevant property. Price exposure is gained; sovereignty over the asset is lost.
It is also worth considering the asymmetry of scale. When a growing share of the 13.9 million BTC held by individuals migrates onto banking rails, counterparty risk concentrates on an asset whose total supply is fixed by definition. Every BTC that enters a bank custody system carries with it a potential uncovered claim – the incentive to lend against the custodied asset exists, is documented, and is precisely the reason banks are building that infrastructure. The 32% average score on Strategy’s index measures how much banks have already built. It also measures how much they still have to build.
Institutional adoption of Bitcoin, in the dominant narrative, is framed as “maturation” of the asset. It is a reading that mistakes the direction of the flow. Institutions are building the infrastructure to absorb Bitcoin into the system that Bitcoin exists to replace. The end result – if the process completes its logical arc – is a market in which most exposure to Bitcoin is mediated by intermediaries, property rights are claims against custodians, and the mathematical scarcity of the underlying asset coexists with a theoretically unlimited supply of claims on that asset.
