When it comes to the number of supercomputing centers it wants to support, the National Science Foundation (NSF) has decided that less is more. Under a new program called Partnerships for Advanced Computational Infrastructure, NSF on Friday announced awards of $170 million each over 5 years to the University of Illinois and the University of California, San Diego, to develop the next generation of high-end computing hardware and applications for the nation's researchers. The decision spells the end of NSF underwriting of supercomputing centers in Pittsburgh and at Cornell University, which have shared their disappointment with federal legislators but pledged to stay open.
NSF began its supercomputing centers program in 1986 to provide access to the megamachines for researchers nationwide. That mission has largely been accomplished, according to a 1995 report from an outside task force, although there remains a need for greater computing power, better tools to manipulate and distribute massive electronic databases, and broader access to such technologies. The Illinois and San Diego proposals "extend our vision beyond what we had even hoped for," says NSF's Paul Young. Toward that end, the two centers have enlisted dozens of regional partners at major research universities, companies, and national laboratories.
The winning entries  had been an open secret for weeks. But last week's decision fleshed out the details: Illinois and San Diego will each receive $29 million in the first year, which starts on 1 October, while Pittsburgh and Cornell will get $11 million over the next 2 years to ease the transition.
On 9 April, NSF will be asked to justify its decision at a hearing before the House Science Committee, two members of which have griped publicly about the planned reduction in centers. And Pennsylvania's senior senator, Arlen Specter (R), has sent NSF a letter containing a veiled threat about possible fiscal consequences to the agency's budget of terminating a program in his state. In the meantime, Richard Zare, president of the National Science Board (NSF's oversight body), strongly defends the board's action, calling it "fair, thorough, and based on peer-based merit review."