That's by design. Large org needs consistent throughput. Think of it like you engineer system. When small, you have pets: each computer is important. Then you s

That's by design. Large org needs consistent throughput. Think of it like you engineer system. When small, you have pets: each computer is important. Then you scale so you maybe get rid of special computer. All computer is cattle. You want consistent performance, willing to accept lower performance for consistency. Advantage is you can build bigger building blocks. No need for 9654X with 1 GB of cache. Everything is now VM. Mechanically unsympathetic, but you swap each piece in logical space, individual device does not matter. You go bigger. Business units now. Deliver consistent performance. All units is not Skunkwork. Just machine. Big unit. Transistor in company CPU. Eventually company as well is just transistor in nation CPU. Nation just transistor in Earth CPU.