Bell Labs (Murray Hill)
The most productive research lab in human history.
Kernel
Bell Labs is the lab that proves Alexandria's model still works in industrial form. From its 1925 founding to its 1984 break-up: the transistor (1947, Shockley/Bardeen/Brattain). Information theory (Shannon, 1948). The first laser (1958). The Unix operating system and C language (Ritchie, Thompson, ~1970). CCDs (Boyle/Smith, 1969). Cosmic microwave background (Penzias/Wilson, 1964 — they discovered the Big Bang's afterglow by trying to clean their antenna). Nine Nobel Prizes. The entire substrate of the information age was, for half a century, an AT&T subsidiary in suburban New Jersey.
Why it worked
A regulated telephone monopoly with guaranteed revenue, a mandate to advance the science of communication, and management discipline (Mervin Kelly, William O. Baker) that protected long-horizon research from short-horizon product pressure. The compact campus, the lunch table, the requirement that everyone publish, and the institutional patience to wait fifteen years for a payoff. Most of these conditions are now structurally impossible to recreate.
Why it declined
The 1984 antitrust break-up of AT&T removed the patient-capital that funded the lab. By the 1990s the research mandate had been narrowed and the talent had migrated to Silicon Valley, Microsoft Research, and university labs. The lab still exists (now Nokia Bell Labs); the institution that produced the transistor does not.