(FORTUNE Magazine) – The modern corporation, like modern art, is over. We're seeing more and more mutant forms of company organization: alliances and joint ventures, equity spinoffs, tracking stocks. The big corporation--with public investors, outside directors, and hired-hand managers--made sense as a vehicle for raising funds to acquire and operate expensive assets like steel mills: The directors made sure the managers operated the assets in the best interest of the people who bought them.

The postmodern corporation is different. In knowledge-intensive companies, most of the value is produced not by capital equipment but by talent, which is owned and invested by employees. When the key capital is human, you have to help employees align their best interest with that of the business. That's why traditional talent-based organizations have been partnerships--like law and accounting firms.

Partnership doesn't work for many of today's talent-based companies, such as software producers, Internet portals, etc., which need access to capital markets--so that they can do things like buy other businesses and pay people. Cap Gemini can (and might) buy the consulting business of Ernst & Young, and not vice versa. Hopes of fat stock remuneration explains why senior consulting-firm partners like Andersen Consulting's former chief executive George Shaheen have left ritzy jobs for risky startups.

Partnership has one set of limitations; the public company has another. It can reward the wrong investors, giving too little to its human capitalists. The old structures lead companies into "the innovator's dilemma." That phrase is the title of a fine book by Harvard Business School professor Clayton Christiansen. It argues, essentially, that success hogties innovation. Companies with a leading technology in a market usually do a good job of improving what they have but rarely develop a new technology that displaces the old. Too many vested interests (money, time, talent, assets, mindset) are tied to the existing product line. Moreover, a new technology almost always starts out less capable than a dominant old one. If you own a goose that lays golden eggs, it seems irrational to give its feed to a quail that lays silver ones. Also, big companies have a problem making other-than-big bets. If it's not worth a few hundred million, you can't get the CEO's attention.

Teradyne Corp. has found a way out of the innovator's dilemma. The Boston company is big (1998 sales: $1.49 billion) and dominates its market. It makes automated testing equipment used by electronics and telecom companies to test semiconductor chips, circuitboards, telephone networks, and software. Complex machines can cost $2 million each. But in 1998, Teradyne introduced a tester called Integra, based on a new, low-cost technology; it's not yet good enough for high-end applications (like testing microprocessors) but in a few years will be--at perhaps one-quarter the cost of the old technology. About 250 Integra systems have been sold in its first 18 months on the market, generating sales at an annualized $150 million--the fastest ramp-up in company history--and it is already profitable. Teradyne and Integra are doing what history and professor Christiansen say can't be done.

They did it by tinkering with the governance model. Alex d'Arbeloff, Teradyne's founder (1960), chairman, and till 1997 CEO, has never lost the entrepreneurial itch: Over the years he has served on the boards of a dozen startups. That might be why he conceived of a novel way to start Integra as an ersatz startup, with its head reporting not to a boss but to a board of directors. "We'd never done this before," d'Arbeloff says. "I just thought it was a good idea to set it up as a venture. It was a good argument for recruiting talent, a way to give the team a feeling of independence."

This wasn't wildly radical. The board was internal, its members all Teradyne executives, including d'Arbeloff and George Chamillard, who is now CEO; and Integra wasn't given phantom stock or an out-of-the-ordinary compensation scheme. But it was enough to change the environment. Integra's general manager, Marc Levine, had a business plan and venture capital, not a budget. Says Levine: "The idea was to think of this as a business from the start, not an R&D project. The board setup allows more of a coaching attitude."

Having a board benefited both Levine and Teradyne's leadership. The structure allowed senior management to be highly involved, the way an activist VC board is--without tying up their time or binding Levine's hands. Says Edward Rogas, a Teradyne VP: "Being on the board let us jump out of the daily crap, which is the toughest thing for a division manager to do and easily fills 30 hours in a 24-hour day."

Levine got independence that he used first to recruit both inside the company and out. Says Chamillard: "We could have given the project to a division, but they would never have put their best people on it." Autonomy also allowed him to break with custom--purchasing software outside rather than developing it in-house, as Teradyne usually does, for example. Integra's strategy--to produce a low-cost product for the low end of the market--was fundamentally different from the rest of the company's, and would have conflicted with the strategy of any division in which it was housed. Above all, the board gave Integra enough remove from Teradyne's day-to-day business that Levine could take technological risks without worrying about existing customers. Says Rogas: "A division is always pressed to do the next logical thing--and make it compatible with the existing line. We told Marc: Be aggressive on the technology; do something no one else has done."

This year Integra becomes a "normal" division of Teradyne. Levine will get his budget through the same planning process as everyone else; he'll have customers to serve, improvements to make--the daily crap. A different kind of venture might have been spun off or found some other hybrid form. That's the fundamental point: As business increases in speed and knowledge intensity, structures have to adapt too. Says d'Arbeloff: "In fast-clock-speed industries, you're more likely to see these structures. But everybody's got to have a venture mentality."