Spent some time this month actually digging into why the builder insolvency numbers look the way they do, instead of just noting that they're bad.
The pattern's consistent. Contracts signed in 2021, sometimes early 2022, at fixed prices, before materials and labour costs took off. Then two years of the cost side climbing while the revenue side stayed locked in on paper. Eventually the gap between what a job costs to finish and what the contract pays becomes a loss on every single project, and there's no way to trade out of it.
Newer contracts price in the higher costs, so the risk is mostly concentrated in that 2021 to 2022 cohort working through their backlog now. Doesn't mean it's over. Just means the shape of the problem is understandable, which is at least useful if you're trying to underwrite the risk of a builder failing mid-project.