The clearest non-scam answer I’ve found for ‘who the hell is funding these APYs’ is fiat denominated loans against BTC, ETH et al. People with long, appreciated crypto positions want to take profit/add leverage, so they borrow stable coin against crypto at substantial rates, collateralized by the crypto. As long as crypto goes up and to the right, borrower is in the green, can pay interest with some of the borrowed stablecoin (or in the underlying crypto asset) and has some nice stuff. This demand for fiat denominated borrowing can be seen in the differential rates offered to BTC/ETH/… depositors vs stablecoins. An order of magnitude difference in favor of the stablecoin is not unlikely.

The lender is basically long the underlying asset, which they could pay to hedge (likely impairing their 10-20% APY) or just hold, since crypto only goes up. The stablecoin depositor is ‘risk free’ in the sense that loans are secured against other crypto, but exposed to massive exchange rate risk. In a year 2BTC recovered from a default on 100k USD might be pittance or a windfall, but I wouldn’t take a strong position on which, at least not for 1200bps.

In general one wonders how many of these 10-20% APY crypto things are just mis-priced exchange rate risk/exposure. A hypothetical protocol/coin minting 10% if it’s overall volume per year would have no problem paying 10% nominal APY to holders. But, like investing in a bond issued in an foreign, inflating currency, one would expect the exchange rate to decline enough consume most of the nominal yield, leaving some real yield proportional to the other underlying risks. If the mechanism is sufficiently obscured, or lost in the frothy demand , borrowers might actually get outsized returns, for a while. If it’s not, they might be surprised when they try to take profit in fiat.

This is compounded by PoS systems where there are real rewards minted by holders to compensate for the costs of maintaining the network.