From Buzzword to Banking Stack

If you spend enough time around blockchain conversations in financial services, you start to notice a pattern: people talk about “tokenization” like it is one thing. It is not. It is three distinct categories of work, each with different risk profiles, different operational demands, and different paths to production. Tokenization will matter most when it stops acting like an experiment.

From where I sit as Head of Institutional Finance at Ava Labs, the team building the Avalanche blockchain, the clearest way to deconstruct what is actually happening inside banks and financial institutions is to separate the use cases into distinct categories: crypto custody, stablecoin payments, and traditional assets coming on chain.

Custody is where activity shows up first because it starts with client demand and maps cleanly to a service banks already understand. Early momentum came through private banking, and the pattern is consistent: banks go where the demand is. That is why custody has moved from niche to mainstream, with major custodians launching digital custody platforms and expanding them. It is a contained problem with familiar controls, even if the assets are new.

Stablecoins are different. They touch the core of what banks do: move money. Cross border payments, settlement speed, and competition from non-banks are pushing this from curiosity to strategic urgency. Banks are responding by exploring tokenized deposits as an alternative, partly to protect their deposit base and partly to stay relevant in the payment stack. I do not view this as a winner take all debate. There is likely a role for both stablecoins and tokenized deposits. What matters is that banks are becoming the bridge between crypto and fiat, and that bridge is where a lot of real financial services value will be layered.

The third category is the one most people in lending care about: mortgages and other Real World Assets. And here, the truth is simple. The most exciting use cases exist, but we are still early. Yes, you have examples like on-chain HELOC origination, and you have emerging platforms bringing commercial real estate and other bank assets into a more portable, machine readable format. But broad adoption across primary mortgages is not imminent because the hard part is not tokenization. The hard part is changing how assets are issued, serviced, governed, and moved through their full lifecycle.

That distinction matters more than most conversations admit. Tokenization is often used as a blanket term, but there is a meaningful difference between digitizing the wrapper and digitizing the asset. If you tokenize the ownership record, the cap table equivalent, you get a “casino chip” that moves around and represents interest in something. You can get to market quickly and still unlock value, especially around collateral mobility, but the deeper efficiencies come when the asset itself is digitally native. This is when lifecycle events such as repayments and amortization are executed on-chain, and where the chain becomes a true system of record rather than a mirrored representation of an off-chain process.

The challenge is that lending is not going to wake up tomorrow with a single digitally native loan origination system. We will be issuing loans in traditional systems for the foreseeable future. There are too many issuers, too many workflows, and too much embedded infrastructure. That means the near-term value will come from reliable onboarding of assets originated in a Web2 environment, validating the data, maintaining that data over time through integration with servicing systems, and making those assets more portable and composable for downstream capital markets use.

This is where the “democratization” promise of blockchain becomes practical. Regional banks fund a meaningful share of the economy, yet many of their loans live in formats that are not portable and cannot be recycled efficiently. They do not have capital markets takeout desks. Their capital gets trapped. If you can bring those assets into a validated, standardized digital form, you can route them into participation vehicles, warehouse structures, or securitizations, with automation around concentration limits and other functions that have historically required heavy operational overhead. Done correctly, that does not just benefit the largest institutions with the biggest innovation budgets. It can help smaller and regional institutions compete, but only if we are intentional about design and access.

Of course, none of this matters if it cannot live inside a bank’s existing risk and control framework. That is where I think some of the fear is misplaced. The principles banks rely on are not new. Defense in depth, role-based access controls, separation of duties, governance and permissioning, and strong encryption are all familiar concepts. They simply need to be applied to new control surfaces.

In digital assets, one example is that the most dangerous key is not necessarily the one that grants access to an individual holding. It is the key that controls the smart contract itself, because that impacts every holder of the asset. That means institutions need clear separation across wallets that control mint and burn, whitelisting, pause and freeze functions, and especially contract upgrades. The control environment has to be designed with the same rigor banks already understand, just mapped onto a different technical substrate.

This is also why the public versus private chain debate has never been academic. Financial institutions have two jobs. The first is “don’t go to jail.” They have fiduciary obligations to their clients, and legal obligations to their regulators. The second is “don’t miss out.” They want to deliver best execution to their clients. Privacy, compliance, and safety push institutions toward permissioned environments. Liquidity, innovation, and network effects pull them toward public permissionless environments. The first wave of private chains solved for control but failed on connection, largely because they became silos. The next wave is racing toward connection, but it cannot ignore why privacy and governance mattered in the first place.

What I believe Avalanche gets right is that you do not have to choose between control and connection. Complex financial transactions contain multiple trust boundaries and multiple data sensitivity layers. Some elements belong in a permissioned context. Others benefit from public settlement and liquidity. An architecture like Avalanche’s that supports both, without relying on third party bridges, opens a much broader design space for regulated finance: control without isolation, and connection without compromise.

Regulation is often framed as the gating factor, and it does matter. My view, though, is that we have centuries of regulatory principles we can reuse. We will need tweaks and specific implementation patterns, but we do not need to reinvent the concept of custody, governance, and control from scratch. The more interesting comparison right now is between Europe’s clarity and America’s experimentation. Europe has harmonized regimes and earlier supervisory structure, but that has also created constraints and slower product velocity. The US allowed experimentation to run ahead of formal clarity, and that helped markets and products develop. It is now codifying the learnings into formal regulations like GENIUS and CLARITY. The key is finding the balance where innovation can move without leaving institutions paralyzed by uncertainty.

If I had to make one five-year prediction, it is that some of the most controversial debates today will look overly precious in hindsight. We are spending enormous effort splitting hairs over control in principle versus control in practice, especially in collateral mobility and related structures. Those questions matter, and they will be resolved. But I suspect that in five years, the industry will view some of that debate as a level of legal and regulatory anxiety that extended further than it needed to, because the operational reality of control and liquidation will prove more stable than the fear suggests.

The real work is not the terminology. It is turning tokenization from a wrapper into a workflow, and from a pilot into a banking stack.