They All Pay The Same Today. That Is Not What You Are Choosing.

Say your company is holding four million dollars it does not need this quarter. Somebody points out that it is sitting in an account earning nothing, and by the end of the week there are four tabs open and a spreadsheet with a column called "yield."

Right now those tabs say 3.44%, 3.60%, 3.60%, 3.62% and 3.75%.

Best to worst, that whole list is thirty-one hundredths of a percentage point wide. Two of them are the same number to the hundredth. On the four million dollars you are arguing about, the gap between the winner and the loser is twelve thousand four hundred dollars a year, which is roughly what the meetings about it will cost.

So the decision everybody treats as the decision is not really one. Whichever you pick, you get about the same thing today.

Most of that list barely moves. A fund holding treasury bills earns what treasury bills earn, and that drifts slowly and predictably. Coinbase sets a rate, publishes it, and tells you when it changes. Those are honest products and none of this is an argument against them.

One of them is not like that, and the difference never comes up while you are choosing.

When you supply dollars to a lending market, what you get paid is not set by anybody. It is set by how many people want to borrow that morning. More borrowers, higher rate. Fewer borrowers, lower. It updates every few seconds, forever, and nobody writes to tell you.

We read that rate off Aave's own record at one thousand four hundred and sixty four points across the last year, six hours apart, so that nothing here depends on which morning we happened to look.

The 3.62% you were quoted is one frame out of that. The middle of the year is 3.30%. The bottom is 1.56% and the top is 12.60%, and for 83.5% of the year, about 306 days of it, the rate was under 4%.

Six months ago it paid 1.67% for weeks. That is less than half what the products directly above it on your list were paying the same week, to the same dollars, for the same risk you had already decided you were fine with.

The high point is worth a minute, because it explains the shape of everything else.

On the nineteenth of April the rate reached 12.60% and then stopped there. Not near 12.60%. Exactly 12.60%, reading after reading, for hours.

A number that refuses to move has hit a wall rather than found a level. Borrowers had taken out every available dollar, and once a pool is fully lent out the formula that sets the rate simply has nowhere further to go. That 12.60% was a ceiling.

Which means the number that would headline any pitch deck was Aave paying its absolute maximum precisely because it had run out of money to lend. A withdrawal comes out of whatever is sitting idle in the pool, and that afternoon nothing was, so anyone who wanted their dollars back was waiting on a borrower to repay.

Grant the on-chain option its liquidity, because it genuinely earns it. Supplying dollars this way is the most liquid thing on that entire list, and a withdrawal settles in seconds, which no bank sweep on earth can do. That was true on three hundred and sixty five days of the year. It was not true on the one day the screen showed the best number it showed all year.

Of the $2.4 billion sitting in that market today, $307,364,990 has not moved in a year.

No deposit. No withdrawal. Not one transfer in either direction, by anyone, across 366 days. That money was there at 12.60% and it was there at 1.56%, and neither of those did anything to it.

We are careful about what that shows. It shows what happened to the money, which is nothing, for a year. Somebody may well have been watching the whole time and had perfectly good reasons to leave it alone. What we can say is that the position never changed while the thing it was earning changed by a factor of eight.

On four million dollars, picking the best option on that opening list instead of the worst is worth twelve thousand four hundred dollars a year. That is the choice that got the spreadsheet and the meeting.

On the same four million, the difference between the year's middle rate and what that money was actually paid during the spring stretch at 1.67% is sixty five thousand two hundred dollars a year.

Five times the money, sitting in the part that was never a decision at all.

And the reason it was never a decision is not carelessness. There is no moment where anybody is asked. The venue asks you once, on a Tuesday, and then never asks again.

One sentence is worth carrying back to whoever signed off on the original decision, in your own words.

Nobody at your company chose a yield. Somebody chose a venue, once, and in doing so agreed in advance to accept whatever that venue happened to pay for as long as the money stayed there.

What a treasury actually wants to express was never a number anyway. It is a rule. Keep this much liquid. Put the rest wherever it earns the most. Move it when that stops being true.

A rule is a thing you can write down once, on the Tuesday, while you are paying attention and everyone is in the room. Then it is still true at three in the morning in April when the rate falls through the floor and every person who would have noticed is asleep.

That is what Plug is for. You describe the condition and what should happen when it is met, sign it once, and it settles itself afterwards without anybody being at a desk.

We measured 3.59% last night and 3.62% this afternoon while writing this. That is not a correction. That is the entire point.

The rate on your company's cash was never a decision anyone made. It is a decision that keeps being made, every few seconds, by strangers deciding how much they feel like borrowing.