@Fippy I might disagree on an average game. Certainly if you build GLH because of it that's going to overcome failgold. But in general I'd say whether failgold or early wonders is more powerful (or important to you winning) depends on what kind of win/loss ratio the person has. Failgold is a get ahead stay ahead mechanic, because if you're too far behind then you can't failgold. Someone with a 30% win rate should find the most value in IND from failgold, because with a good start they can easily stay ahead of the curve enough to access failgold for a bunch of wonders, which keeps them ahead of the curve, and so they win. Someone with a 95% win rate, is going to win with an average or better start regardless. There the value comes from winning those difficult maps, where you're going to be behind and therefore not get a lot of failgold opportunities. A lot of those difficult maps though can be turned into easy wins with the appropriate wonder (coast/islands -> GLH; barb hell -> TGW; iso -> mids).
On the time value of money: yeah things snowball but you still have to do something with it. Hammers tend to snowball quicker than money. Once you've built X you benefit from X (for the most part, military units complicated). Money doesn't give you any snowball effect on its own. It's only when converting to beakers, and even then only when beakers have added up to a full technology. This delay means there's a chance for failgold to come in before you would have finished a tech without it, at which point you've lost absolutely nothing. There's also a chance that failgold takes a while, and the person who built wealth grabs Tech #1 a few turns faster than the failgolder. One way that "snowballs" is by enabling infrastructure or something to improve your economy. But this is going to be a relatively minor effect over such a short timescale compared to a +50% return. The real issue is if Tech #1 is good trade bait, and getting it a few turns later risks the AI acquiring the tech during that time. However, the failgolder will get their failgold eventually. There are more techs to consider. They'll be getting all subsequent techs earlier, meaning it's the wealth builder who is taking the extra risk on missing out on tradebait. And this is assuming they don't get failgold quickly! in which case the wealth builder is slower to even the first tech.
If failgold comes in before the first tech is completed, failgolder is ahead on all fronts. If failgold comes in after mattering for the first tech but before the 2nd tech (most likely), failgolder will be behind on the first tech, but ahead on tech 2, tech 3, tech 4, etc. If failgold is suuuuper late (worst case scenario) failgolder will be behind on both tech 1 and tech 2 before overtaking the wealthbuilder at tech 3+.
Failgolding is better long-term, but may be slightly worse as far as the immediate next tech is concerned. If failgold was a 10%-20% bonus, this might lead to interesting choices. But at +50% the only time I'd considering hedging my bets by building wealth is on teching Liberalism itself. And that would of course only be if you thought Liberalism was at risk. By monitoring the tech chart carefully, much of the time there's no risk in losing the techbait, or the flip side: you've already lost the techbait. Or you've got no one you could trade to! Or no one can trade you anything!
Wealth building is a certain loser over a long enough timespan. It would only make sense over failgold if you thought getting to the next tech first was both at risk, and going to determine all on its own whether you win or lose the game. Maybe you can concoct an example, but all I've got is Liberalism.
This sidesteps the real power of failgolding though, which is we can turn forests and whips into money. You can't do that by building wealth. You can also failgold before Currency, which is important because being able to get to currency before your economy stagnates is important.