IBM'S SECRET WEAPON

• Article
• Comments ()
• Click-2-Listen
• Videos

By Melita Marie Garza BLOOMBERG NEWS

International Business Machines Corp. credits its No. 2 spot on the Dow Jones Industrial Average this year to more than 3,000 workers like Michael C. Pitman. He's not an Internet wizard or a supercomputer designer. He spends his time studying how the light receptor works in the human eye.

One of the last corporations still emphasizing basic research, IBM attributes part of its stock's gain this year to expanding the development budget when competitors like Bell Labs fell behind.

IBM spent $6.2 billion in 2007, 30 percent more than it did in 2002, on projects with little discernible impact on the company's day-to-day work. Instead, its scientists are tracing migratory patterns of humans and studying the DNA of cocoa trees.

"Research is IBM's secret sauce, the special house dressing that allows all the divisions to stay competitive," said Richard Doherty, research director at Seaford, N.Y.-based technology analyst Envisioneering Group Inc.

Chief Executive Officer Sam Palmisano is following in the tracks of Lou Gerstner, who overhauled research strategy while he ran IBM from 1993 to 2002. Now scientists, not just sales staff, meet with customers of the biggest computer-services company, and it has more than 75 development centers worldwide.

IBM last week reported a 22 percent increase in second-quarter profit, beating analysts' estimates, as research spending rose 8.2 percent to $1.66 billion. The stock trailed only Wal-Mart Stores Inc. among members of the Dow this year, up 17 percent.

IBM's research budget, which includes the science projects and development of new products like server computers, grew by $1.4 billion from 2002 to 2007. In 2007, for the 15th straight year, IBM won the most U.S. patents of any company, 3,148. Research and development totaled 6.2 percent of sales.

In contrast, Xerox Corp., whose Palo Alto Research Center in California developed the graphical user interface, spent $912 million on research last year — about the same as it did in 2002 — or 5.3 percent of sales, according to its annual reports.

Hewlett-Packard Co., the world's largest personal-computer maker and No. 10 on the patent list, said in March it would fund fewer projects. It spent $3.6 billion on research in 2007, a little more than in 2002, or about 3.5 percent of sales.

Research spending for Alcatel-Lucent SA, which absorbed Bell Telephone Laboratories in 2006, was $4.3 billion in 2007, a 6.5 percent decrease from what the combined companies spent in 2002, or 17 percent of sales. Bell Laboratories developed the transistor.

"Their heyday seems to be over," Jonathan Eunice, an analyst with Nashua, N.H.-based researcher Illuminata Inc., said of Bell and Xerox PARC. "What IBM did was it figured out how to get value out of the R&D process."

At Thomas J. Watson Research Center in Yorktown Heights, N.Y., one of IBM's eight major labs, Mr. Pitman, an expert in biomolecular dynamics, and Ajay Royyuru, senior manager for IBM's computational biology program, led a three-dimensional tour through the light detector of the human eye on a June day.

On three 48-square-foot screens, Mr. Pitman projected a blue, green, red and white model of rhodopsin, the membrane protein responsible for dim-light vision. Mr. Pitman and a team of scientists created the simulation using IBM's Blue Gene supercomputer.

Working with the National Institutes of Health and university researchers, IBM showed for the first time that a significant amount of water resides in the protein structure during light detection in the eye. Understanding the makeup of the protein is critical to drug development in a market worth tens of billions of dollars.

"As we develop faster and faster supercomputers, it's not always clear how to use them most effectively, and this is one example of how use them to advance drug discovery," said Mr. Pitman, who has worked at IBM for 12 years.

Elsewhere in the lab, Ted van Kessel and Bob Sandstrom stood in the sunshine. Like children burning leaves with a magnifying glass, they focused 230 watts of sunlight through a lens onto a solar cell 1 centimeter square.

The cell converted the energy into 70 watts of usable electricity, about five times what a cell that size typically generates. The trick lies in cooling the tiny square as it absorbs enough heat to melt stainless steel.

IBM doesn't plan to make or sell solar panels. The company intends to license the technology to a solar-power company or form a partnership with a utility. Last year, it got about 10 percent of profit — almost $1 billion — from licensing its technology.