How Decentralized is DeFi?

Decentralized finance (DeFi) once sold the promise of an alternative financial system built on crypto with no middlemen, no gatekeepers, no regulators, and, most of all, no entrenched corporations or governments that'd be able to erect barriers to entry or toll booths. 

That story was always quite emotionally satisfying to crypto natives, and it was sometimes even technically accurate. But, it quietly smuggles in an assumption that investors cannot afford: That DeFi is actually sufficiently decentralized on average to confer the positive qualities that the evangelists claim it has.

As-of notes: BTC supply distribution uses River’s ownership research (Aug. 2025) as reported in CoinDesk and River’s own research page. ETH, XRP, and SOL non-retail buckets mix (a) public-company and government treasuries from CoinGecko, (b) ETP/managed-product AUM by asset from CoinShares (as a proxy for “funds/ETFs”), and (c) foundation/issuer disclosures where available. “Individual/Retail control” for non-BTC columns is a residual estimate from the other rows, so treat it as a best-effort guesstimate, not a directly observed measurement.

Source 1 | Source 2 | Source 3 | Source 4 | Source 5 | Source 6 | Source 7 | Source 8

That assumption is not one to take on faith. So investors are left to either treat decentralization as something measurable, or let the market measure it for you, most likely during the worst possible hour. We propose the former. Here's what you need to know, and what you need to do. 

What Is Decentralization Anyway?

Before we go layer by layer about how to probe DeFi decentralization or lack thereof, it behooves investors to notice a recurring pattern. 

1. DeFi tends to decentralize the parts of a financial process you can point to on-chain

2. It then recentralizes the parts that users might not always think about explicitly, like upgrade pathways, data oracles, and transaction plumbing.

From there, we can posit that a useful definition of this issue treats decentralization as a set of properties or a spectrum rather than a binary label. A cryptocurrency could easily utilize highly decentralized data oracles or transactions, and still rely on fully centralized development of its platform technology, and that would imply a different set of risks than a different asset which is collaborative in its development process but which uses fully centralized transaction processing.

Therefore, a simple map here splits DeFi into three planes. 

• Settlement: Where transactions finalize.

• Control: Who can change or pause the system. 

• Access: Who can reach it in practice. 

Use the table below as a forcing function, as it turns the word "decentralized" into questions with observable answers.

Next we can look at the first place decentralization often thins out: liquidity.

So How Decentralized is DeFi, Actually? 

The reality is it depends on how you look at it. Let’s use Bitcoin as an example:

Who Owns Bitcoin (Rough Averages)

When looking at government holdings, the threat of centralization is relatively slim; worldwide, government bodies hold less than 3% of Bitcoin in supply.  By contrast, “wealthy individuals” (individuals with more than 100 bitcoin in their wallet) seem to hold most of the power here. When adjusting that parameter, however, data from BitInfoCharts shows a more stark delineation in the top holders:

Bitcoin Wealth Distribution

The top 2% of Bitcoin addresses control 95% of the entire Bitcoin supply, and Bitcoin's Gini coefficient (a measure of inequality where 1.0 represents perfect inequality) sits at 0.842, making it one of the most unequal asset distributions ever measured.

Even more complicated, however, is when we look at the difference between control and influence. Consider these examples from 2025 data:

Individual Control Analysis

These individuals and their close networks demonstrate how crypto's "decentralization" claim obscures highly concentrated control. Michael Saylor through MicroStrategy alone owns 3.197% of all Bitcoin ever to exist, which is more than most nation-states. Meanwhile, Trump's ability to influence U.S. regulatory policy affecting crypto gives him effective control over far more Bitcoin than he personally owns.

None of this is to say that DeFi as a whole is in danger; as a reminder, the analysis here is restricted to Bitcoin due to the transparency of available information). It is to say, however, that investors need to be aware of the gap that exists between risk management and wishful thinking as they move forward.

Liquidity Concentrates By Default

Crypto markets are permissionless and still end up fairly concentrated, because there are strong financial incentives which have the effect of concentrating them over time. That brings us to the most honest parts of DeFi's decentralization story; capital goes where it can earn the highest yield for the same unit of risk, and capital also goes to where execution is best. 

DeFi is permissionless in the sense that anyone can deploy code and anyone can trade. Markets, however, almost never reward players equally. Three forces drive this organic consolidation:

• Incumbency Effects: Early movers capture liquidity and keep it

• Capital Efficiency: Deeper pools of capital offer tighter spreads on trades and more opportunities to generate yield, thereby attracting even more capital

• Order Flow Capture: Winners attract more volume, compounding their advantage

The nature of the game is that the most efficient (and highly centralized) places to process transactions will ultimately hoard liquidity, as those with the most capital will benefit the most by basing their holdings there. In practice, it's very hard to go against this dynamic and still make a meaningful amount of money when working with anything more than five or six figures worth of starting capital. 

Still don't believe it's like this? Consider that 20% of the pools can account for more than 90% of a DEX's trading volume. Liquidity is always a choke point, and choke points turn local failures into systemic ones. Knowing this, many DeFi projects offering very different sets of financial services leave the door open to reducing their decentralization even further in a crisis by borrowing from (or lending to) other operators. 

Consider that if these arrangements are factored into a project's planning process (like the idea that an emergency credit line might be available) the effect is a reduction in effective financial decentralization across the sector. 

You can see the same dynamic at the protocol layer. A snapshot of top protocols by TVL repeatedly shows a small set of names absorbing a big share of user capital across cycles. Protocol concentration can thus hand you the opposite of diversification, with correlated exposure to a handful of governance systems, codebases, pools of capital, and operational teams. 

Control And Governance Are Where Decentralization Falters

Protocols don't run themselves, even if they're technically operated by a DAO. They get upgraded, patched, reparameterized, and sometimes rescued, all by living humans, many of whom have other things to be doing as their main vocation in life. 

Put differently, the mechanisms that make a system maintainable can also create human control points, and control points are a form of counterparty risk that's especially pronounced in crypto, where the human capital can be just as volatile as the coins.

The table below lays out common centralization vectors and what an investor can check quickly.

Upgradeability is probably the easiest example because it's explicit. If a protocol uses a proxy that can be upgraded by an admin, your investment thesis should include the behavior of whoever controls that admin key, plus the operational security of the signer set. Most investors can't even name the chief technical people at the chains whose native tokens they hold, but they'd gain some value if they did. 

Governance adds another twist, namely that even when important decisions are decided on-chain, voting power can and usually does compress into a minority of holders who are both invested and vocal. When turnout is low, as with many altcoins during bear periods, the effective control group gets even smaller. 

Access is Where Decentralization Becomes a User Experience Problem

Most users experience DeFi through interfaces, wallets, hosted RPC endpoints, and transaction routing infrastructure. Even if the smart contracts are permissionless, the journey to those smart contracts often is not, which creates another point where decentralization can become a slogan while practical usage of it is optional or irrelevant.

Interface screening makes the point crisply. If a major front end blocks wallets flagged as illicit, users may still be able to interact with contracts directly, but the default path closes.

MEV infrastructure is another part of the access story. When transaction routing concentrates, inclusion and ordering become mediated by a smaller set of actors than the settlement layer suggests. Unlike other decentralization issues, MEV farmers and bots are already notorious for extracting vast amounts of value from hapless investors who usually don't even understand that it's happening. Paradoxically, this same problem happens in the highly-centralized stock market, but often with less-detrimental results for investors.  

That raises another issue, namely a version of the white-hat versus black-hat split, but in miniature. 

• White hats want many independent routes from DeFi apps back to the chain because it improves resilience and supports the credible neutrality of the plumbing. 

• Black hats, on the other hand, want chokepoints, because chokepoints are targetable, and, despite what the name implies, systems that implement relatively wide chokepoints with good throughput are actually far more efficient than decentralized ones, giving them a large advantage in attracting capital -- and making them a more juicy target for attackers. 

So if you're investing, you want to know which side your protocol's architecture accidentally helps, because that's going to be who uses it the most.

The last piece is a bit uncomfortable because it is social. Narratives about decentralization have incentives behind them, and those incentives shape what gets repeated. Generally, projects and their aligned promoters benefit when users believe decentralization is already solved, because it lowers perceived counterparty risk and raises their willingness to deploy capital. 

So, In Closing, How Decentralized Is Defi Really? 

It's typically only as decentralized as what there's capital to support, and capital tends to prefer at least some amount of centralization in some relevant areas on average. Armed with that knowledge, as an investor your job is to find the weakest link when it comes to decentralization, because that's where stress shows up first, and it's also how protocol-breaking problems are likely to cause financial damage to you.

Sources

• DeFi risks and the decentralisation illusion

• Concentration of DeFi's liquidity

• Top Protocols

• Contracts API: Proxy

• Concentration of voting rights and low voting activity for decentralized finance protocols

• Chainlink: A Full-Stack Institutional Platform

• USDC Risk Factors

• Address Screening Guide

• MEV-Boost Introduction

• Relay Payloads Percent

• OFAC Compliant Blocks

• U.S. Treasury Sanctions Notorious Virtual Currency Mixer Tornado Cash