But China's obsession with large projects has resulted in massive environmental damage, such as smog in nearly all cities, social ills caused by the migration of as many as 300 million workers, and corruption.

"Large projects put a lot of power in the hands of managers of state-run companies," says Zha. Infrastructure overbuilding already has generated overcapacity. Currently, only about one-sixth of the south-north water-diversion project capacity is used.

The government is curbing some large projects. Public details on the Tsangpo dam program, once envisioned as a 40,000-MW effort to divert water to China's arid regions, now only refer to 2,000 MW of installed capacity."There are reports of subsequent phases that are far bigger, btr no clarity on the actual size of projects," says Joydeep Gupta, South Asia director of an internet publication dedicated to Asia's water crisis.

One suggestion from scientists to use nuclear energy to cut a massive pass in the Himalayas to draw in monsoon winds is under fire from environmentalists and neighboring country officials. "India needs to be very vigilant about Chinese interventions in the Tsangpo," Ramaswamy R. Iyer, former secretary of water resources in the Indian government, told ENR.

While the growing scale of infrastructure development is stimulating better ways to prepare megaprojects, the industry's track record remains poor, says Bent Flyvbjerg, a researcher and consultant at Oxford University and academic director of its business school. Less than 10% of the major international projects he has studied were successful "from beginning to end," he says. The rest fell into the "break-fix" category, in which projects are rescued by injecting good managers.

Dreams Can Become Nightmares

Many megaprojects are undermined by a failure to describe accurately the costs and benefits, a problem that stems from the schemes' proponents needing "to make the project looks good on paper," asserts Flyvbjerg. "They do that by ignoring a lot of the real risks … and they also overestimate the benefits."

Australia's Sydney Opera House ended up 1,400% over budget at its completion 41 years ago, says Flyvbjerg. Over the past half-century, Olympic Games' facility cost overruns averaged 179%. The London Games "were the most expensive Olympics ever when they were built," says Flyvbjerg. But it was a legitimate vanity project, since the U.K. decided, " 'We are going to do it well, and that's going to be expensive,' " he says. "They got very good project managers, and, because they were good, they said, 'We want sufficient funds up front.' "

But with so much management outsourcing, governments now need better project leadership skills.

Oxford's new Major Projects Leadership Academy is providing 12 months of training for all U.K. civil servants charged with managing large investment programs. "Nobody who hasn't gone through the program will be allowed to be responsible for a major project," says Flyvbjerg. "Other governments are very interested in this."

Raymond Levitt, director of Stanford University's Global Projects Center, plays down Flyvbjerg's theory that dream projects are a conspiracy "between politicians who want to cut ribbons and engineering firms [that want to] get more business." He also disputes early reads on Boston's Big Dig project as a failure, noting its impact on urban quality and real estate value. But Levitt emphasizes, "Project governance needs work." In fact, it is a major new area of Stanford's research, particularly related to funding.

"When I look at large programs underperforming, the No. 1 reason is that the owner's organization cannot clearly articulate the objective," says Fluor's Prieto. "Why are we doing this project? Are the owner and the project-delivery team ready?" He advocates a need for better models to measure direct and indirect "cascading" impacts of scope changes.

Wallace sees scale "as a reasonable criteria for a dream project as long as its effects are limited and manageable."

He concedes, "The engineer in me thinks it would be fun to build [a supertall high-rise] just to see if such a building could be built. But just because we want to build it and we can build it doesn't mean we should build it. The litmus test is the effects of the project on society. Because our technology is so potent, the level of scrutiny needs to be very high."

But Prieto says even unbuilt dream projects can produce "a nugget picked up for another project. Dream projects are essential to innovation. They may be impractical today, but they are still relevant."

He adds, "there are people thinking about building shelters for humans on Mars."