Crypto
Most token projects spend their first six months on the part nobody is worried about. The chain. The ticker. The chart. Meanwhile the question every serious buyer asks inside the first two minutes of a call goes unanswered: who is holding the thing, and what happens to me if they stop.
Nobody buys a claim. They buy custody of a claim.
Custody is the unglamorous layer. The vault, the key, the legal wrapper, the named party who is responsible when the wire does not land. There is no narrative in it. It cannot be marketed with a countdown timer. And it is the only part of the stack that decides whether a person with real money participates at all.
The oldest version of this business is close to a thousand years old, and it was not run by a bank.

The Order of the Poor Fellow-Soldiers of Christ and of the Temple of Solomon was founded around 1119 to protect pilgrims on the road to Jerusalem. Within a few generations the protection work had become something else. By the standard account, a pilgrim could deposit coin and valuables at a preceptory in Europe, carry a document instead of the metal, and draw against the deposit at the far end of the journey. The road was no safer. They had simply removed the thing worth stealing from the traveller.
That was the real product. Not the sword. The receipt.
It worked well enough that the French crown kept its treasury at the Paris Temple for most of the thirteenth century. A military order became the most trusted custodian in Europe because it had quietly solved the problem everybody had and nobody wanted to discuss.
A token is a receipt. That is the honest description of it. It says something exists somewhere and that you hold a claim on part of it. Every interesting question lives on the other side of that sentence.

Where is the asset. Who legally holds title. What happens to the claim if the issuer dissolves. Who can be served with a notice. What does redemption look like on a Tuesday afternoon when the person asking is annoyed and not technical.
If you cannot answer the redemption question in one sentence, you are not selling an asset. You are selling a mood.
The failure mode
The Templars also show you how custody dies, and that half of the lesson usually gets skipped.
On 13 October 1307 the arrests began. The Order was suppressed in 1312. The network that had worked for most of two centuries did not fail because the vaults were weak. It failed because one party had come to hold everybody’s trust, and the party holding everybody’s trust is the most attractive target in any system.

That is the honest case for self custody. It is also the reason self custody has not won. Spreading the risk means carrying it yourself, most people will not, and telling them they should is a sermon rather than a strategy.
Centralized custody fails all at once. Self custody fails one person at a time.
Build the boring layer first
If you are putting a real asset on a chain, do the custody work before the brand work. I have spent the last two years on the design and go to market side of tokenized real world assets, and the pattern has never broken: the engineering is tractable, the marketing is tractable, and the thing that stalls a deal is always custody and recourse, because those are the parts where somebody has to accept liability in writing.
Name the entity that holds title. Name the jurisdiction. Write the redemption procedure as a numbered list a tired person can follow. Publish who audits it and how often. Decide in advance what happens when the asset is sold, damaged or reassessed, and put that in the document instead of discovering it live in front of holders. The liquidity problem and the adoption problem both sit downstream of this one.

Then build the button, the brand, the community. Those still matter. They just cannot carry weight that custody has not already taken. The asset comes first, every time.
Solomon put the discipline in one line: Be not thou one of them that strike hands, or of them that are surety for debts. Read forward, that is not an argument against custody. It is an argument against accepting an obligation you have not priced.

The chain proves the record. Somebody still has to hold the thing.
