(FORTUNE Magazine) – If Arno Penzias had arrived at AT&T Bell Laboratories in 1991 instead of 1961, he would never have won a Nobel Prize in physics. Today Penzias, 60, is a top manager at Bell Labs, with a substantial research budget. But these days he almost certainly wouldn't fund the kind of heady astrophysics for which he and a colleague got a Nobel in 1978. Their radiotelescope detected echoes of microwaves radiating in deep space that appeared to have been created when the universe was formed -- buttressing the theory that it all began with a Big Bang. The Bell Labs that many Americans cherished as a national treasure is no more. Before the 1984 breakup of AT&T, it was supported by a kind of hidden tax included in America's phone bills from America's monopoly telephone company. Bell Labs' work was meant to benefit the nation. And did it ever: From the labs in the New Jersey countryside emerged the transistor, the laser, the solar cell, the first transistorized digital computer, cellular radio (now used for phones), the Unix operating system for computers, and a galaxy of other seminal technologies that gave rise to whole new industries. Now the labs work only on projects likely to bear fruit for AT&T's stockholders -- and fairly quickly. For $3.5 billion annually, roughly 39% more in real dollars than AT&T spent before the breakup, it expects proprietary product ideas. Intellectual speculation in biochemistry is out. Digital compression algorithms and semiconductor physics are in. Says CEO Bob Allen: ''Anybody who believed that Bell Labs would not change as a result of the breakup of the Bell system must have been smoking something.'' The labs have narrowed their focus, concentrating on areas of research with close connections to AT&T's businesses -- computing, networking (both for computers and telecommunications) , photonics (manipulating light signals, especially important in fiber-optic lines), software engineering, and technologies related to speech and images. A parade of innovative AT&T products suggests that the labs' competence in research, though refocused, remains unimpaired. For example, in 1989 a team led by Jay Wilpon made a crucial inroad in the vexing problem of computer recognition of human speech. By breaking down words into basic sonic ^ components, the team devised a technique known as word spotting. A phone caller might mumble, ''I wanna maga collecall.'' The software can pick out the key word ''collect'' from a mass of unrelated sounds and ask the caller for further instructions, eliminating the need for an operator. Next year word- spotting computers will handle over a billion calls nationwide, for both AT&T and such customers as New Jersey Bell. AT&T has just announced the development of a voice-controlled cellular car phone that uses a related technology. All the good ideas no longer come from Bell Labs, however. When AT&T needs something it didn't invent, it gets it elsewhere. The AT&T videophone, introduced last year, uses a modem and audio compression software invented at Bell Labs. But the central innovation -- software that squeezes video signals onto a regular phone line -- came from Compression Labs, a medium-size company in San Jose, California. Says Penzias: ''Today I no longer feel insulted if we don't do everything.'' It's much more important to be first to market with new stuff, he says. Most of the staff at Bell Labs today doesn't even work for Bell Labs. About 80% of the employees report to AT&T business units, which pay their salaries. Many employees carry two ID cards, one for the labs and one for their business unit. Each product group now has a chief technical officer, who reports both to the group executive and to John Mayo, president of Bell Labs. That way the company grooms more executives who understand both technology and managing. Penzias, Allen, and Mayo all claim there's no reason to worry about America losing a strategic asset. Their argument: The labs work for AT&T, and AT&T can do only as well as its customers -- which still include the vast majority of U.S. telecommunications users, both individuals and businesses. Says Penzias: ''I'm not saying what's good for AT&T is good for the country. But -- without being suicidal about it -- we have to go beyond profit to some extent and make sure society gets the most benefit from the technology.''