In fact historically, some of the higher-end hardware yielding the best performance during operation (an axis along which I optimize) actually added time to the boot sequence. The storage subsystem on my workstation is backed by a mix of four Intel enterprise-grade SSD's in RAID-0 (raw speed) and 8 big spinning platters in RAID-6 (capacity), plugged into an Areca 1882ix RAID card w/ 4GB dedicated BBU cache. Unfortunately that card adds a non-bypassable 30 seconds to the boot sequence, no matter what system you plug it into. But once there, it screams. It's only just the last couple years PCI-NVMe drives have come out that can match (or finally beat) the performance metrics I've been hitting for ages.
So I actually kind of feel like I've been living in the future, and the rest of the world just caught up ;-).