HILLSBORO, ORE. - Intel Architecture Labs used to develop hot technologies and give them away to other companies. But then the research group smartened up and is now making almost 45% of its costs back from licensing its creations.
"We were naïve on how to write deals to benefit both us and other companies," says Craig Kinnie, director of the labs. Now, he says Intel is aggressive in its licensing deals, either getting money from companies or exchanging rights for access to their technologies.
Intel's $70 million research effort brings in about $30 million from deals with other vendors. Kinnie hopes to increase that figure over the coming years. He adds that the lab funds new projects with the money it brings in.
Ask Intel researchers how they are able to return such a profit and they'll tell you it's simple: They are practical about which projects they choose.
"We don't take on a project we can't see through to the end," Kinnie says. Almost 90% of the projects result in products for Intel or its partners.
Intel has three goals for the labs: make the microprocessor business thrive, help existing business units by developing new features and functions for products, and start new business units.
When the labs started in 1991, researchers focused on LANs. "The 'Net has certainly changed that," Kinnie says. The teams now focus on seven areas: home technologies, broadband, conferencing and telephony, Internet media, information management and the exchange of knowledge, manageability, and scalable platforms.
More than 500 researchers work in small teams as well as with product groups and Intel partners to develop new technologies for each area. There are four steps the researchers take: develop a technology, write the specification, create a prototype and test for interoperability.
It's a process that in some cases takes more than five years, a bit longer than Intel chief Andy Grove had originally anticipated. "The labs were Andy's idea," says Kinnie, a 23-year veteran. "When the PC was becoming obsolete, he said we have to bring leadership to the industry. But he thought we could [finish projects] in one or two years. It takes much longer."
In fact, the labs fund researchers at universities for projects that could be as far as 10 years out. For instance, Intel is funding Lixia Zhang at the University of California in Los Angeles to develop the Resource Reservation Protocol, which lets users schedule a path through the network for packets. It's a bet that Intel is making on the importance of bandwidth management. "The ability to schedule bandwidth has got to come about," Kinnie says.
Kinnie says Grove was surprised by the quick impact the labs were able to make on the industry, helping create technologies such as DVD, v.90 modems, Windows Telephony API, G.Lite digital subscriber line and H.323 conferencing. "And I've been amazed by the amount of influence we've been able to have outside of the PC market," Kinnie says.
While not much phases Kinnie, the possible glut in technical recruits has him concerned. Intel is pouring money into grade school and high school mentoring programs, college scholarships and other opportunities, hoping to create a new crop of talent.
Kinnie is also not shy about plucking Master's students and Ph.D.s straight out of school to give them experience. "We have kids a year out of school working on some of our major initiatives," he says. "Their lack of fear is a benefit."
RELATED LINKS
