Intel leverages universities for research

By Ed Sperling, Editor in Chief -- Electronic News, 7/17/2007

Intel is heavily leveraging relationships at universities to do its core research, dividing each project up into dozens of discrete slices that can be done in pieces.

The approach is radically different from that taken by most of its competitors, which are working in ecosystems with each other on what they consider pre-competitive research—research that is so bleeding edge that turning it into products can take years and often yield different results for each of the contributors.

Paolo Gargini, Intel’s director of technology strategy, said the university infrastructure works “10 times better than Bell Labs.” He said Intel contributes $1 billion annually to the program, but that for Intel to do that research the price tag would be far higher.

A decade ago, research papers from universities were considered substandard compared with the work being done by large companies such as Intel, Texas Instruments, IBM and others. But with rising costs and increased complexity at each new semiconductor process node, most companies are changing direction. Texas Instruments, for one, has offloaded its process development to TSMC, and IBM relies heavily on partnerships in a broad ecosystem it has developed. Gargini said university research is now every bit as good as what comes out of private industry, and possibly better.

At least part of the change has to do with the structure and discipline that Intel has placed on its research. The company works closely with hundreds of universities, some working on programs involving one professor and one student, while others are working broad programs involving a number of universities. Gargini said Intel first began adopting the strategy after the U.S. government shut down the national laboratories in 1996, greatly accelerating its push into the university research world several years ago.

On top of that, Intel Capital invests heavily in startups, many of them spinoffs from universities, as a means of keeping track of new development work.