Crypto removed the middleman. Then it rebuilt him.

The Bitcoin white paper is nine pages long and its argument fits in a sentence: you can move value between two people without a trusted third party in the middle. Not a better middleman, not a regulated one — none. That was the claim, and it was genuinely new.

Seventeen years on, look at how people actually use this technology. They buy through an exchange that holds their coins. They hold dollars as a token issued by a company. They stake through a pool. Increasingly they get exposure through a fund that trades on a stock market and never touch a wallet at all. At almost every layer where an institution was removed, one grew back.

This is the strongest criticism the field faces, and it is usually made badly — as a gotcha, as if the whole thing were a con. It is more interesting than that, because the middlemen came back for reasons that are entirely rational, and what remains after they did is still not nothing.

Where they came back

Custody was first. Self-custody means holding a key nobody can recover for you, and the consequence of losing it is total and permanent. Most people, correctly assessing their own filing habits, decline. So exchanges hold coins for hundreds of millions of users, which makes a balance there a claim on a company rather than property you control — the thing “not your keys, not your coins” is warning about. The warning is correct and mostly unheeded, because the alternative asks people to become their own bank at a moment when they wanted to buy fifty dollars of something.

Consensus followed. Bitcoin mining was meant to be broadly distributed; it is now dominated by large operations and pools, because the economics of specialised hardware and cheap electricity reward scale the way they reward scale everywhere else. Proof of stake did not escape this either. Running an Ethereum validator takes 32 ETH, so most stakers delegate to a pool or an exchange, and influence pools accordingly — one of the criticisms our comparison of the two mechanisms takes seriously.

Then money itself. The most-used dollar in crypto is not a decentralised construct — it is a token issued by a company holding reserves you cannot inspect, redeemable on terms you probably do not qualify for. Stablecoins are the settlement layer of the entire market, and the dominant designs are corporate liabilities. An industry built on removing the need to trust an issuer now runs on trusting an issuer.

And most recently, the wrapper. Spot ETFs let people own exposure through a brokerage account. A custodian holds the actual coins; the investor holds a share. It is the oldest financial structure in the book applied to the newest asset, and it has been the most successful distribution mechanism the asset has ever had — precisely because it removes every part of crypto that people found difficult.

Why it happened, without the sneering

The lazy reading is that this proves the promise was empty. It does not. It proves something duller and more durable: intermediaries exist because they perform work people do not want to do themselves.

A bank is not only a rent-seeker. It is also a password reset, a fraud department, a phone number, and someone who is liable when things go wrong. Removing it does not delete those functions — it reassigns them to you. Self-custody is not merely holding your own money; it is being your own security desk and your own recovery process, permanently, with no appeal. Most people, offered that trade explicitly, do not want it. That is not ignorance. It is an accurate assessment of how much attention they have.

Scale did the rest. Wherever an activity gets cheaper the bigger you are — buying hardware, sourcing electricity, meeting compliance requirements, running validators reliably — concentration follows. That is not a property of blockchains. It is a property of the world, and no consensus mechanism has repealed it.

What actually survives

Here is where the criticism overreaches, because the interesting thing is not that middlemen returned. It is what happens now when one fails.

When exchanges have collapsed — and several large ones have — the asset did not stop working. The ledger kept producing blocks, self-custodied holdings were untouched, and people withdrew to their own wallets in enormous numbers. Compare that with the failure of a bank, where the money exists only as an entry in the failed institution's database and the resolution is a matter for regulators and a compensation scheme. In crypto the intermediary is a service layer over a ledger that does not depend on it. In traditional finance the intermediary is the ledger.

That is the difference worth defending, and it is narrower than the original pitch: not “there are no middlemen” but “using one is a choice you can reverse, at any time, without anyone's permission”. Exit is available. It costs a transaction fee and some inconvenience rather than a court order. Nothing about the fact that most people don't exercise it makes the option worthless — a right you can use unilaterally constrains the party you might use it against, whether or not you do.

Where it genuinely breaks down

An honest version of this argument has to concede the cases where exit is not actually available.

You cannot exit a stablecoin into anything other than another claim. If the issuer fails, holding your own keys protects your ability to move the token, not its value — the asset itself was the liability. That is a real hole in the story, and it sits under most of the market.

Concentration in consensus is worse still, because it is not something an individual can opt out of. If a small number of pools produce most blocks, that is the network's property, not your account's, and moving your coins to a hardware wallet does nothing about it.

And an ETF holder has no exit at all in the sense that matters here. They own a security, not a key, and the entire proposition is that somebody else handles the part where the technology was novel. That is a legitimate product. It is just not an instance of the thing the white paper described.

The honest summary

Crypto did not remove intermediaries. It made them optional, unbundled them from the ledger they sit on, and made leaving one a decision you can execute yourself. That is a real change and a much smaller one than was advertised.

Whether it is enough depends on what you thought you were buying. If the promise was a financial system with no institutions in it, that has plainly not happened and shows no sign of happening. If the promise was that no institution gets to be the final authority on what the ledger says — that one still holds, and it is the only part that was ever load-bearing.

The useful question is not whether the middlemen came back. They did, everywhere, and pretending otherwise is how people end up surprised when one fails. The question is whether you can still leave. Keep checking that the answer is yes, because that answer is the entire remaining difference — and unlike the price, it is something you can verify for yourself.