a16z Crypto: Traditional finance wants blockchain, not DeFi

null

Author: a16z Crypto

Translation: Jiahuan, ChainCatcher

In the crypto industry, there’s an imagination about the future that has almost become the standard answer: DeFi and TradFi move toward convergence, unpermissioned liquidity meets institutions’ distribution capabilities, and ultimately a graceful hybrid emerges that captures the best of both—replacing the old system.

That story sounds reassuring, but it’s basically wrong.

The more honest version is: as long as blockchains can make existing businesses perform better, traditional finance will use them. Not because they’ve embraced decentralization, but because the cost calculus works out. This technology compresses costs, improves settlement, expands distribution—and it lets institutions keep tighter control of customer relationships.

This means institutions are not “merging” with DeFi. They’re simply picking out the parts of DeFi that fit their operating constraints, discarding what doesn’t, and then reassembling everything according to institutional requirements. The end product won’t be like traditional finance, nor like today’s DeFi. We’re witnessing a new category emerging: programmable financial infrastructure that runs on the blockchain rails but is optimized for institutional constraints.

As regulatory frameworks mature, this landscape may change. Legislation like the CLARITY Act may make it easier for institutions to directly access permissionless systems in the future. But no matter how open the legal environment becomes, traditional finance’s risk appetite won’t be reset overnight. Institutions evaluate technology with a constant focus on cost, risk, control, and operational fit. That’s exactly why there are two opportunities in front of the industry—not one.

The first opportunity is to help institutions adopt the infrastructure they are ready to accept today. Each time an institution adopts a component—atomic settlement, programmable money, or tokenized collateral—it’s effectively validating the technology, refining the shared rails, and bringing real transaction volume and capital on-chain.

The second opportunity is to keep building the open, crypto-native financial system that institutions aren’t prepared to use yet.

These two paths are not mutually exclusive. They can coexist and, if done well, even reinforce each other. Open networks will continue producing new components, markets, and innovations, which institutions will ultimately put to use. If both succeed, “convergence” will happen naturally: not one side absorbing the other, but both sides becoming increasingly dependent on the same underlying infrastructure.

What traditional finance is actually doing

When traditional finance adopts a component, it must satisfy two conditions at the same time: first, it must improve costs, risk, or distribution; second, it must not break control and accountability mechanisms. The components that institutions discard—like permissionless access, anonymity, and trustless execution—can pass the first test, but not the second.

So institutional adoption is predictable, not random—and founders can treat it as a design test. In other words, if the value of a feature can only be achieved by taking away an institution’s control, then no matter how elegantly it’s designed, it’s almost destined to be modified or rejected.

Let’s run a few components through this test. Atomic settlement eliminates the time gap between trade execution and final settlement, reducing counterparty risk, and it also frees up the collateral institutions lock up for unsettled transactions. A shared ledger turns the biggest hidden cost in the back office—reconciliation—into something trivial.

Programmable money allows interest payments, margin top-ups, and corporate actions to be executed automatically as code, instead of relying on a chain of manual instructions. Once the permissionless “shell” around AMM curve math is removed, it transforms into an on-chain pricing engine for FX and for net asset value of tokenized money market funds.

Each of these components can improve some number on an institution’s P&L, or eliminate an operational risk and its associated costs—but none of them requires institutions to “believe in decentralization.”

So we need to be clear: JPMorgan’s permissioned chain for institutional deposits, and BlackRock and Franklin Templeton’s tokenized money market funds—these projects are not enterprises “testing DeFi.” They’re using blockchain to do things they already do, such as bank-to-bank payment settlement, managing fund subscriptions, and distributing yield-bearing instruments—just with a better pipeline.

These deployments use blockchain’s technical attributes: programmability, transparency, and atomic settlement. At the same time, they intentionally discard the attributes that allow native DeFi to function: permissionless access, anonymity, and trustless execution.

This is neither failure nor compromise. It’s a deliberate architectural choice—and it clearly tells us which direction things are going.

Buyers differ, rules differ

If you think institutional adoption is simply opening up a bigger distribution channel for existing DeFi infrastructure, you’re wrong. The way institutions evaluate protocols is completely different from how crypto-native users do.

From an institution’s perspective, this is selecting software vendors and infrastructure partners. They must consider operational risk, compliance controls, and the long-term ownership of key systems—everything follows their own standard workflows. The result is that success in DeFi cannot be automatically converted into success in the institutional market.

Companies rarely buy the “best” technology. They buy technology that best fits their real-world conditions—existing workflows, risk models, procurement processes, and so on.

Any technology entering a heavily regulated, heavy risk-control, extremely responsibility-averse institutional environment will be reshaped by that environment. The internet went through it (enterprise firewalls, internal networks). Cloud computing went through it (private clouds, VPC, FedRAMP certifications). AI is going through it now (on-prem deployments, data residency requirements, model governance). Blockchain won’t be an exception.

This reshaping unfolds along two axes:

First is compliance. KYC, AML, sanctions screening, investor accreditation, and regulatory reporting are non-negotiable for most institutions. Permissionless systems by design can’t meet these requirements. Institutions need the ability to freeze assets, revoke transactions, and identify counterparties.

DeFi wasn’t designed with these in mind from the beginning. Meeting them often means making major architectural changes. This may loosen in the future—for example, the CLARITY Act could allow institutions to access permissionless systems while satisfying regulatory requirements. But today, when most institutions evaluate blockchain infrastructure, they still look primarily at control, accountability, and operational risk.

Second is enterprise value delivery. This axis is often underestimated. Institutions adopt blockchain not because they believe in permissionless principles, but because it compresses costs, reduces reconciliation friction, opens new distribution channels, or lets them embed even more deeply in customer relationships. The value proposition must be expressed in these terms—otherwise it won’t even pass the procurement stage.

Stablecoins may be the clearest example. Banks, payments companies, and fintech firms increasingly treat them as useful settlement infrastructure because they can move dollars faster across networks and regions. But only a few are truly embracing the ideology of permissionless finance. They adopt programmable dollars because it’s useful—not because they want to rebuild the financial system according to DeFi principles.

Circle’s evolution is a telling case. Its Arc network reflects how blockchain infrastructure is being packaged and sold to institutional buyers: emphasizing compliance, operational control, credible counterparties, and integration with existing workflows—rather than permissionless access and composability.

It isn’t selling “permissionless” itself. It’s selling faster settlement, global reach, and higher capital efficiency—delivered in forms that institutions can actually use.

Even organizations like SWIFT increasingly view blockchain from this angle. Its attempts at tokenized asset interoperability aren’t meant to replace existing financial institutions; they aim to help those institutions collaborate better using the SWIFT network. The same pattern keeps showing up: blockchain adoption reinforces existing financial networks rather than replacing them.

This is how strong technology meets a mature market—it always evolves in this way.

Two opportunities in front of founders

At the industry level, it’s wrong for everyone to give up one opportunity to squeeze into the other. At the company level, it’s also wrong to try to grab both.

Institutional adoption and open network ecosystems can mutually reinforce each other, but for most teams, they are two fundamentally different businesses. Doing institutional business requires understanding procurement, compliance, internal controls, channel partners, and long sales cycles. Doing open network business requires optimizing around developers, liquidity, composability, and network effects.

Who the customers are, how distribution works, what the product must deliver, and how success is measured—these often differ completely between the two.

This isn’t to say one opportunity is better. It only requires founders to think clearly about which market they’re serving, while remembering that what connects the two is the common underlying rail: public chains as a neutral settlement layer.

Working with institutions and building a parallel financial system are not mutually exclusive. If done well, they can amplify each other’s value. The permissioned layer brings transaction volume, legitimacy, and capital, while the open layer continuously produces the components that the permissioned layer will adopt next. If “convergence” happens, it will occur at the rail layer—not by one side surrendering to the other.

The role of public chains as settlement rails may become increasingly important, even if the applications running on top become more and more permissioned.

Built for programmable financial infrastructure

To build this new programmable financial infrastructure, there are two options: build from scratch, or adapt existing products.

First, look at a network like Canton. It hasn’t tried to retrofit existing DeFi infrastructure; instead, from the beginning it was designed around institutional requirements for privacy, compliance, and controlled interoperability. Its goal isn’t to bring banks into DeFi. It’s to use blockchain-based collaboration mechanisms while preserving the governance, confidentiality, and operational control institutions require.

But successful institutional strategies don’t necessarily all require a ground-up rebuild. Morpho takes the opposite approach. It didn’t discard its DeFi components; it focused on making those components easier for institutions and asset issuers to use.

For example, Apollo’s ACRED fund incorporates Morpho into its on-chain lending strategy—pairing a DeFi-native lending component with institutional-grade distribution, compliance, and fund structure.

In the end, the result is neither pure DeFi nor a fully isolated institutional technology stack, but a pattern where institutions selectively adopt existing crypto infrastructure and then repackage it to fit their own requirements for control, compliance, and distribution.

This new category is built specifically for institutional constraints. It draws from DeFi, but operates in a more permissioned and more compliant way—so it must be different from anything that exists today.

Teams like Morpho that successfully adapt crypto-native infrastructure for institutional use cases do exist. But founders shouldn’t treat that as the default playbook. Institutions are an independent customer group with unique needs. In many cases, designing from the start around those needs will be more effective than adapting products originally built for open networks.

Opportunities to keep building in DeFi

None of the innovations institutions are adopting today were born inside banks, asset management firms, or existing financial infrastructure. They all come from open networks—the places where entrepreneurs can freely experiment with new market structures, new collaboration mechanisms, and new financial components.

This distinction matters. Institutions are not the primary source of innovation in this industry; the permissioned layer is usually downstream of the open layer.

This leads to a more critical strategic judgment: if the whole industry is focused on selling to banks and asset managers, we might mistakenly treat a big customer segment as the whole opportunity. TradFi is an important customer, but it isn’t the only customer.

Designing for institutional needs is a legitimate and valuable path, but it’s only one lane, not the entire highway. Companies that last are the ones that always know exactly who they are building for. Institutional adoption may be a huge opportunity, but it’s not a simple extension of DeFi. Success in one market doesn’t guarantee success in another.

If you’re building for institutions, commit fully. Don’t assume that success in the crypto-native market automatically translates into adoption by enterprise customers. Learn your customers, understand procurement processes, and intentionally design around institutional needs.

If you’re building for open networks, keep going. Don’t abandon your vision just because institutions are the loudest buyers in the current market.

Remember: these two paths are complementary, not competing. One is responsible for taking verified innovations and adapting, commercializing, and scaling them. The other is responsible for discovering these innovations.

Some version of this technology will almost certainly become part of the financial plumbing in the existing TradFi system—but that isn’t the only future being built. Open networks are still the industry’s most important testbed and innovation source, and many of the components that tomorrow’s institutional infrastructure depends on will most likely be created there first.

TradFi isn’t adopting DeFi. It’s selectively adopting the parts that fit its own patterns.

The opportunity for founders isn’t to chase every market at once, but to think clearly about which market they’re building for—and then execute accordingly. The future may indeed run on top of institutional infrastructure, but the most important innovations will still keep coming, continuously, from open networks.