Intel, Stanford Engineering share history of technology, talent development

Stanford electrical engineering alumnus Marcian "Ted" Hoff, became Intel employee number 12 back in 1969 and within two years, along with Federico Faggin and Stan Mazor, he had invented Intel's flagship product: the microprocessor. The success was no fluke. For more than four decades, the Stanford-Intel relationship has yielded not only important advances in semiconductors but also hundreds of successful, and sometimes legendary, engineering careers. Today almost 1,000 Stanford alumni have Intel on their resume.

While the company has launched the careers of many Stanford Engineering graduates, there are also esteemed professors back on "The Farm" who recall that Intel supported their work right from the start. Over the years Intel's support has created myriad opportunities for education and PhD-quality research in a majority of the school's departments.

Perhaps the best embodiment of the Intel-Stanford relationship is Craig Barrett, who was a materials science and engineering alumnus and then served as a professor from 1965 until he joined Intel in 1974. This month, the former Intel CEO will retire from his post as chairman of the board. The historic moment provides an opportunity for reflection.

"Across the board from computer architecture, to semiconductor device modeling, to new packaging technologies, to new computer programming algorithms, the work at Stanford has helped move the industry forward," Barrett said in a recent interview.

"We've had great results from the collaboration," added Pat Gelsinger, an Intel senior vice president and the general manager of its Digital Enterprise Group. Gelsinger, a former CTO and head of research at the company, earned a MS degree in electrical engineering at Stanford in 1985. "In almost every area that Intel is doing work we can point to significant collaboration and research projects with Stanford." 

The longtime interaction has not only been beneficial to Stanford, but also to the culture of research and innovation that is crucial to the local economy and the broader technology industry, said Stanford President John Hennessy.

"From the late 1970s when I joined the electrical engineering faculty, to today, Intel has been one of the most reliable and energizing sources of support for Stanford students, faculty and their research," said Hennessy, who is also a professor of computer science and electrical engineering. "This especially close relationship has not only led to new technologies and knowledge, but has also strengthened the model of academic and industrial partnership that sustain the success of Silicon Valley."

Intel is a strong partner for a broad swath of the academic community, added Jim Plummer, dean of the School of Engineering.

"The company's leaders believe in, value, and push for long-range university research," said Plummer, a professor of electrical engineering and a member of the Intel Board of Directors since 2005. "It's had a big impact, not just on Stanford but also on the entire university community that works in the semiconductor space."

Plummer credits not only Intel's direct support of university programs but also its crucial role in the creation of industry research groups including the Semiconductor Research Corporation. Intel has also lobbied the government very effectively in the past to continue funding academic research on technology.

Early, steady support

Long before it became the world's largest semiconductor company (a status it has held since 1992), Intel's founders were willing to make big bets on supporting academic research. Thirty years ago, Stanford launched the Center for Integrated Systems (CIS), an effort to bring together computing hardware and software researchers with industry on an open, pre-competitive basis. Bob Noyce and Gordon Moore personally enrolled Intel as a founding member.

Intel's support for the center in the early days was not just financial, notes electrical engineering Professor Jim Gibbons, who co-founded CIS and was Stanford Engineering's dean between 1984 and 1996. When Gibbons wanted to recruit other companies to join, he could count on Noyce (as well as HP's then-president John Young) to place the persuasive phone calls to other CEOs. Meanwhile, Intel has also donated key pieces of equipment, such as a six-inch wafer furnace back in the late '90s.

Richard Dasher, executive director of CIS for the last 11 years, emphasizes two points about Intel's membership. One is that it has never wavered, and the other is that Intel has understood the long-term nature of university research as complementary to its own.

Albert Yu, who in 2002 retired as Intel's senior vice president and had been general manager of microprocessors in a 30-year career at the company, explains that Intel has always looked to universities to cover the longest-term research. In the late '90s he initiated the Intel Research Laboratories program to complement the company's development efforts, but he still factored the unique role of universities into that strategy.

"We made a conscious decision that we would do something still closer to the development side, and fund universities like Stanford and others to do the more far-out stuff," said Yu, who earned his MS and PhD in electrical engineering at Stanford in the mid 1960s. He is now a consulting professor in the mechanical engineering department.

Paolo Gargini, director of technology strategy in Intel's Technology and Manufacturing Group, explains that CIS is an important program because it brings together researchers from different disciplines to tackle the most forward-looking and fundamental questions.

For the last two years, for example, Intel has supported a new collaboration between Professor Krishna Saraswat, a semiconductor devices expert, and Professor Shan Wang, an expert in magnetism, to explore the futuristic possibility of making chips that compute based on the magnetic orientation of electrons, instead of by pushing electrons through transistors. The magnetic approach might require significantly less energy and work better at nearly atomic scales.

"Now students of the two professors are working together, learning from each other toward creating a new class of components that we hope will become functional in the next 5 to 10 years," said Gargini, an Intel Fellow and former Stanford postdoc, who is on the CIS advisory board and who leads the semiconductor industry's efforts to stay on a technology "roadmap" of future performance benchmarks.

Professors' projects

The magnetic project is hardly the first one Intel has helped start up within the school. Electrical engineering Professor and former CIS director Bob Dutton recalls that Intel gave him some of the earliest support for his work in the early 1980s on creating semiconductor device and process simulation tools. Ultimately the software his group created, programs like SUPREM and PISCES, became standards for the industry as chips became more complex and simulation tools became indispensable.

"Intel provided seed funding for some of this work and they were early adopters of the tools," Dutton said. "In the '80s people were seeing 20-50 percent savings in their silicon experiments by using simulations to cut down on trial and error."

Similarly, when materials science and engineering Professor Reiner Dauskardt first joined Stanford in 1994 it was with Intel support for his research into how thin films of materials adhere to each other. The issue is crucial for the semiconductor industry, which depends on many such layers of metals, silicon and other elements remaining intact and precisely aligned. Dauskardt's method for measuring thin film adhesion remains in wide use.

"In 1995 when Intel introduced the Pentium processor... there was no way to quantitatively measure adhesion," Dauskardt says. "We developed a method to do that. There is not a single microelectronics company in the world that doesn't use this technique."

Over time Intel has backed research in a large number of areas. Through CIS and in other ways, the company has invested to explore new materials and architectures for building transistors. The company has also supported projects in digital imaging and in visual and graphic computing. In 2005, two Intel supported projects at Stanford made headlines. One was the publication in Nature by electrical engineering professors David A.B. Miller and Jim Harris of a key advance in integrating electronic chips with optical signals. The other was the victory of the artificially intelligent Volkswagen SUV "Stanley" in a desert race of self-driving cars. Intel also supported Stanley's successor "Junior," which took second-place in an urban-themed robot race in 2007.

Those are just a few examples among many, but what makes the enterprise particularly productive, Gargini adds, is that Stanford understands better than most universities how to make the most useful results available through technology transfer and mutually beneficial intellectual property licensing.

"Universities are supposed to advance knowledge and industry is supposed to take this knowledge and turn it into practical commercially usable devices," he said. "The professors at Stanford have always maintained that their number one goal was to publish papers ... and as long as there was funding for the programs and they could publish and do something meaningful they were willing to share the benefits with those who were supporting them. Stanford has been outstanding in this respect."

Talent transfer

At least as important as the technology transfer, is transfer of talent. In fact, back in 1984 when Gibbons became dean he set about interviewing Silicon Valley CEOs to find out what they valued most about Stanford as they built their companies. Gibbons recalls Gordon Moore singling out Stanford's graduating classes of hundreds of Masters students.

"They are the new intellectual pool for Silicon Valley," Gibbons said.

Intel's desire to nurture and recruit top engineering students at all degree levels underlies the company's substantial investment in education at Stanford and elsewhere. The company has provided fellowship support to Masters students since 1993. In 2008, nine students were recipients. Intel also furnishes many students with internships in a variety of locations, including overseas.

Many Stanford-educated engineers have gone on to both technical and business leadership positions at the company. Today, for example, many Stanford engineering alumni are Intel Fellows, the designation the company gives to its technical superstars. Meanwhile, Stanford engineering alumni in top executive positions at Intel include not only Barrett, Gelsinger, Gargini, and former Intel senior VPs Albert Yu and Sunlin Chou, but also current VPs Thomas R. Macdonald, Melton Bost, Michael Ricci, and Nasser Grayeli, a materials science and engineering alumnus who is Stanford's chief sponsor at Intel.

Yu says the closeness that Intel has maintained with respect to Stanford Engineering research also helps in recruiting top students.

"Once you know the faculty, they tend to recommend their best students to you," he said. "That's been ongoing for a long time."

Intel also supports a number of teaching efforts in the school, including a new nanotechnology teaching lab in the Department of Materials Science and Engineering (MSE). The company recently donated a very capable but compact, "tabletop" scanning electron microscope that students can use with relative ease to examine materials and structures at the scale of billionths of a meter.

"Intel has provided significant contributions matched by the School of Engineering for the last five years," says MSE chairman Professor Bob Sinclair. "That support has helped us set up a modern teaching lab for nanomaterials and electronics materials processing and testing."

The ability to give students "hands on" experience with materials at the tiniest of scales is a great asset for a teacher, but also an important, if indirect, investment in human capital for a company whose business is to push the frontiers of physics.

As Gargini reflects on the physics of the project that Intel supports for Professors Wang and Saraswat, he notes that electronic and magnetic interactions may produce usefully different effects, but they are fundamentally different manifestations of the same physical force. Electricity and magnetism have a natural duality to them, not unlike research and development, or for that matter, Stanford and Intel.

Timeline

• 1968: Intel founded.
• 1969: Stanford graduate Marcian “Ted” Hoff joins Intel as employee #12. He co-invents the first microprocessor, released two years later.
• 1974: Stanford Professor Craig Barrett joins Intel.
• 1978: Stanford Postdoc Paolo Gargini joins Intel.
• 1979: Center for Integrated Systems is founded at Stanford. Intel is a leading supporter among member companies.
• Early 1980s Intel supports semiconductor simulation research of Professor Bob Dutton.
• 1985 Intel employee Pat Gelsinger earns electrical engineering Masters degree; Dean Jim Gibbons, Robert Noyce and Chinese Premier Li Peng share the stage at a Stanford conference.
• 1993 Intel begins a masters fellowship program at Stanford.
• 1994: Professor Reiner Dauskardt starts influential work on thin film adhesion with Intel support.
• 1997: Intel donates more than 400 computers to Stanford, and helps support the launch of the PBL Lab in which civil and environmental engineers use computers to design buildings.
• 1998: Intel supports the digital camera research of Professor Abbas El Gamal.
• 2000: Materials research building named for Gordon Moore.
• 2005: Intel supported research on optoelectronics published in Nature.
• 2006: Intel helps launch the multiuniversity Western Institute of Nanotechnology.
• 2008: Intel donates hundreds more computers to School of Engineering.