Vercel's elastic build machines now factor in Turborepo cache hits when deciding whether to downgrade a build to a smaller machine tier. Previously, a warm-cache build's lower CPU and memory usage could trigger a downgrade, which risked leaving a later cold-cache build underprovisioned and unable to complete. The change applies automatically to all builds on elastic build machines with no action required from users.
Questions this post answers
Why did my Vercel build get downgraded to a smaller machine and then fail with insufficient resources?
Previously, elastic build machines on Vercel could downgrade to a smaller tier based on the lower CPU and memory usage of a warm Turborepo cache hit, but a later cold-cache build then lacked enough resources to complete. Vercel now factors in Turborepo cache hits when deciding tier downgrades, so a warm-cache build no longer triggers a downgrade, preventing this underprovisioning issue automatically with no configuration needed. daily.dev surfaces platform changelog updates like this so build failures do not catch teams off guard.