Why is a largely isolated operation justified for the pharma company's top-secret research area, and why is isolation alone not enough?

Because the research results make the company market leader, the extra effort and lost convenience of isolation are acceptable; but to reduce espionage, theft and loss through ransomware, further security measures are still needed inside the isolated system.

The setup: pharma company P simulates molecules in a high-performance data centre with a research lab. It has strict access control and is completely separated from the network. The only exchange with the rest of the company is a data lock (Datenschleuse), operated by specially trained administrators who check all data technically and by content; they also run the systems in the research area.

Data exchange is reduced to an absolute minimum and strictly controlled, and all operating processes are adapted to that. The price is higher effort and less convenience ("just quickly downloading something" is impossible), which is justified by the value of the research.

Isolation reduces the attack surface but does not remove all threats: data still crosses the lock, and insiders are inside the boundary. So additional measures within the cyber system are advisable against industrial espionage, theft of research results, and loss through ransomware.