During the predawn hours of March 1, about 20 Pakistani intelligence agents and soldiers, backed by a team of CIA operatives, stormed a drab house in Rawalpindi. After a brief shootout, Khalid Sheikh Mohammed, the notorious al Qaeda mastermind of 9/11, emerged dazed, disheveled, and in handcuffs. Not much has been revealed since about how authorities hunted him down, but intelligence sources say the case turned on a months-long game of electronic cat-and-mouse between the terrorist and U.S. spy agencies. And it couldn't have been broken without the contributions of a mild-mannered, slightly eccentric, 78-year-old nuclear physicist named J. Robert "Bob" Beyster and his remarkable company, Science Applications International Corp.
Most people have never heard of either Beyster or SAIC -- and that's fine with Beyster, the company's reclusive founder and CEO. Privately held SAIC makes much of the supersecret technology that's at the core of the sleuthing done by the National Security Agency, CIA, and other spook services. Neither Beyster nor the company will discuss any role in Mohammed's capture. But it's known within intelligence circles that SAIC data-mining and sensor systems helped tease out crucial clues about Mohammed's activities from intercepted text messages that he sent to his al Qaeda operatives using as many as 20 different cell phones. Now, with Mohammed allegedly talking to his captors, intelligence officials say they're pumping new material through SAIC-designed systems virtually around the clock, homing in on other terrorists and their plots. "We're winning victories almost every day, and a lot of them are based on stuff made by SAIC," says one intelligence agency consultant.
For Beyster and his company, these are just the latest in a long line of unsung coups. Since 1969, when Beyster quit a comfortable defense-industry job at the age of 45, mortgaged his house, and founded SAIC in a tiny La Jolla, Calif., office with rented typewriters, the company has been quietly compiling a record of technological breakthroughs, management innovation, and financial performance that few corporations can match. After 33 straight years of profits and growth, SAIC is now the country's largest privately held infotech company, with 2002 revenues of $6.1 billion.
About a third of SAIC's business is systems integration for other companies, such as Pfizer (PFE) and BP (BP), but its heart and soul is spy tech. Intelligence agencies don't list or rank their contractors. Intelligence sources, however, say SAIC was the NSA's top supplier last year and in the top five at the CIA. In addition to the high-powered data-mining software that helped nail Mohammed, SAIC makes undersea thermal imaging sensors for tracking submarines. It produces software that spy satellites use to map the earth and feed target data to precision munitions, including those that have been pounding Iraq. It's also a leader in the booming homeland security business: It builds gear that uses gamma rays to peer inside cargo containers and truck trailers.
Adding to SAIC's covert aura, Beyster has hired an unusual number of former spies, law enforcement chiefs, and secret warriors. Some 5,000 employees -- roughly one-seventh of the workforce -- have security clearances. Beyster himself has one of the highest arrays of top-secret clearances of any civilian in the country. "We are a stealth company," says Keith Nightingale, a former Army special ops officer. "We're everywhere, but almost never seen."
Much of the work may be hidden, but it has never been more vital. SAIC is on the front lines of today's most momentous national security battles. It's not too much to say that the future safety of many Americans rests in the aged hands of a brilliant and quirky septuagenarian and his clandestine army of techno geeks.
In a way, Beyster seems to have been preparing all his life to take on today's myriad global threats. He grew up in Detroit, the son of a General Motors (GM) engineer and a mother who dreamed he'd become a lawyer. He served in the Navy in World War II; to Mom's dismay, an aptitude test he took as he was mustering out indicated, as Beyster puts it, "Whatever you do, don't be a lawyer." It did suggest that he might have a calling in science. Beyster went to the University of Michigan, and by 1950 he had earned a doctorate in nuclear physics.
With the Cold War in high gear, Beyster headed to Los Alamos National Laboratory to do top-secret research related to atomic weaponry. He continued to work on nuclear projects in the private sector at General Atomics, but after its parent company was bought by Gulf Oil, Beyster's group of scientists felt marginalized by the oilmen. "It was a terrible company to work in," he recalls. So he took a flier on founding his own firm and was soon back to doing what he loved best: classified work on subjects like the output of a nuclear bomb in the first fraction of a second after detonation.
From the start, SAIC was a different kind of animal. In 1974, Beyster embraced what was then still a novel idea for motivating employees: rewarding them with stock. He created a special arm of SAIC called Bull Inc. that effectively acts as a trading floor for the stock, setting a price for the shares based on SAIC's performance and that of peer companies. The internal market gives workers a way to buy and sell their shares. (The stock has routinely outperformed the S&P 500.) Today, SAIC is fully employee-owned; Beyster retains only a 1.3 percent stake. The program has been widely lauded by corporate governance advocates and has been a key tool for attracting and motivating SAIC's brainiac talent.
Beyster -- Dr. B, as he's known throughout the company -- also motivates by example: A lifelong workaholic, he still puts in 70-hour weeks. He may finally be ready to dial that back a bit: In April he asked his board to search for someone to take on the CEO post by next February. But he still intends to remain deeply involved as the company's main visionary, mapping out product strategy and new technological directions. For decades Beyster has been jotting down thoughts in dime-store spiral notepads everyone at SAIC calls "Beyster books." He's filled more than 1,000 of them; he keeps them in a company archive and often pulls out books from years past to hammer home some technical or managerial point. He keeps fit for the grind by running several miles almost every day, often trailed by other SAIC executives toting notebooks of their own to record the insights the boss is known to toss off as he jogs.
Throughout SAIC's development, Beyster has shown a keen instinct for spotting emerging technologies and a sure touch on acquisitions. He made one of the great steals of the Internet boom by buying Network Solutions, the Web domain name keeper, for $4.5 million in 1996 and selling it for $3.1 billion before the bubble popped. But Beyster's most powerful insight came in the early 1990s, when he foresaw that the mushrooming power of microchips meant that intelligence agencies would soon be wrestling with vast volumes of new data. SAIC's core expertise back then was in manipulating the huge amounts of information required to simulate the firing of atomic warheads. Beyster set out to adapt those technologies to deal with the coming tsunami of data. "He saw the explosion in data mining and knowledge management, even before those were terms people were using," says Duane Andrews, a former Pentagon chief intelligence officer who now heads SAIC's government contracting division.
Among the fruits of that vision are two of SAIC's most technically advanced products: TeraText and Latent Semantic Indexing (LSI). They're data-mining programs -- some of the most powerful in existence. Both are central to enabling intelligence agencies to sift the immense volumes of data they now collect. Beyster was right about the information explosion; the NSA alone intercepts millions of phone calls, faxes, e-mails, and other types of electronic communications every single hour.
TeraText is designed to help make sense of it all. Written texts -- books, magazines, intercepted messages -- in almost every language are digitized and loaded into a database linked to TeraText. The program can drill into that data at blinding speeds: It can process 2 billion documents every four seconds. It works by identifying patterns and connections between names, terms, and ideas that would take the human mind months to collate. For instance, an intelligence analyst might enter a request for all documents mentioning the name "Khalid," the word "sleeper," and the term "blind date," a possible code for a terror operation. The search could be tailored in innumerable ways -- by language, by time of day, and so on -- and would retrieve all records in which the terms appeared in, say, a single sentence. Finding such seemingly tenuous connections can produce the needle-in-the-haystack moment that unearths a terrorist plot.
LSI is even more esoteric. It looks for abstract relationships among intercepted texts and public documents, and can find even less distinct patterns. Say intelligence analysts learn that an al Qaeda subordinate refers to Osama bin Laden as "Blue Nose" and uses the code word "red" for the date of an attack. LSI will group all documents in any language that relate to "Blue Nose" and "red," even ones that don't contain those exact words. How can it do that? It's incredibly complicated. Suffice it to say that LSI processes language in much the same way the human mind does and contains a degree of artificial intelligence that allows it to make judgments about abstract connections. "That gives LSI a great power that, frankly, we've never seen before," says an intelligence agency consultant who uses the software.
However powerful, the current generation of SAIC data-mining software wasn't good enough to help prevent the 9/11 attacks; signals were missed. In an SAIC lab in Annapolis, Md., Steve Rizzi and his 150-person team are working on a highly classified program, called Trailblazer, designed to avoid a recurrence. Trailblazer, several intelligence sources say, may be the most important program for the entire future of U.S. intelligence efforts. It's currently in a development phase but will likely generate billions of dollars for contractors.
Rizzi, a cherubic 40-year-old who has been at SAIC since he was 21, can't say much about his work. But he does offer that the goal is software that can plow through millions more documents in fewer seconds and will be better at capturing signs of trouble. Rizzi seems inspired by the challenge. "We're all about solving hard problems," he says. "Terrorism is a hard problem."
Duane Andrews was a hot commodity when he left his post at the Pentagon at the end of the first Bush administration. He'd been part of the government's intelligence elite for more than 20 years; among other things, he'd headed the Department of Defense's satellite surveillance systems and overseen the CIA budget. Many of the big-name defense contractors were courting him, eager for his expertise and extensive contacts. Almost on a lark, Andrews went to see Beyster in Albuquerque, N.M. Beyster jotted in his ever-present notebook and talked to Andrews about the benefits of employee ownership. The big defense contractors were about to enter their post-Cold War contraction, but Beyster had a vision of where defense technology was headed that, to Andrews, seemed fresh -- inspired, even. "SAIC moved up from no place on my list to first place after one day of interviews," Andrews says. He accepted the job a week later.
The hire turned out to be a bonanza for Beyster. When Andrews took over SAIC's government division in 1993, it had about $900 million in annual revenue; in 2002, that figure was $4 billion. The growth didn't come from scoring a few giant deals, as can happen in government contracting; SAIC's biggest single government win is a $1.25 billion, seven-year deal to run the National Cancer Institute's main research center. But what SAIC lacks in contract size, it makes up for in quantity: it has 5,300 other government deals. To get them, it has become particularly effective at unseating larger incumbents when their pacts come up for renewal.
In the mid-1990s, SAIC was winning only about 30 percent of those bids against incumbents. Today it wins 67 percent of them. One key to that success grew from an idea culled from a Beyster book. In the early 1990s, Andrews and Beyster launched a program that dispatches corporate emissaries to its government clients to check on SAIC's people -- and to fix problems quickly. "We'll send in someone from another sector of the company who doesn't have an ax to grind," Beyster explains. "We'll move people around if somebody isn't working well or is not a good fit." (A former CIA analyst who has worked with SAIC puts it another way: "If you don't perform, you're outta there fast.")
Another boost has come from a program Beyster initiated in 1996. Called Lessons Learned, it's run by a senior vice president who answers only to the CEO. The VP's full-time job is to study, using a scientific method only a nuclear physicist could love, why SAIC wins or loses a bid. Under the program, SAIC has cataloged and data-mined thousands of its own bid proposals to see which phrases, descriptions, and other features have scored highest with government procurement clerks. The resulting information has been used to develop guidelines for everything from how to describe SAIC's plan for staffing a project to how to explain its software development process, right down to the specific words to use.
In Beyster parlance, the guidelines are called "fragments." They're disseminated to all managers on CDs for use again and again in preparing bids. "It's been very effective in improving the win rate," Andrews says. One recent victory: SAIC supplanted Raytheon (RTN) on a $155 million, five-year contract to help manage the U.S. Geological Survey's massive store of data from sensors and satellites.
SAIC has also benefited from what Andrews calls "punching up the barrel," meaning hiring heavyweights from agencies the company does business with -- people like himself, with deep Rolodexes. Their job is to court their contacts, including those at their former agencies. To some, that is tantamount to influence peddling, but it's the way defense contracting has always worked, and Andrews makes no apologies. He has landed some big fish lately: SAIC recently hired Donnie Marshall, a former head of the Drug Enforcement Administration, to run a group that's pitching business to the new Department of Homeland Security. Andrews expects that agency to be a bureaucratic maze, and Marshall earned a reputation during his rise at the DEA as a master navigator of byzantine organizations. He was the first chief in the history of the 9,000-member force to rise to the top through the ranks. David Tubbs is a former high-ranking FBI official who, after leaving the bureau, coordinated security at the 2002 Winter Olympics in Salt Lake City. SAIC got wind of the Greek government asking Tubbs to come to Athens to discuss security for the Games there in the summer of 2004. SAIC soon hired Tubbs. Last month it beat out Raytheon and other large companies for a $272 million contract to provide security for the Athens Games.
An organization full of former spies, soldiers, and crime busters naturally presents management challenges. Indeed, SAIC has some critics in the government who complain about a hyperaggressive cowboy culture and the company's constant raids on the managerial ranks of intelligence and defense agencies. Some Pentagon officers have dubbed the pair of buildings at SAIC's Washington-area site the Two Towers, after the strongholds of the evil wizards in the Lord of the Rings trilogy. "They're so cutthroat that they actually had two different parts of the company fighting for the same contract," says an Air Force intelligence analyst who has had dealings with the Two Towers. Even some SAIC insiders acknowledge that the company displays, as one executive puts it, a certain "Darwinian chaos."
But that, like everything else at SAIC, is by the deliberate design of Dr. B; he believes that organizing the company as a loose federation of semiautonomous entrepreneurs is the only way to generate the constant innovation that SAIC thrives on. And he brushes off assertions that SAIC can be a dog-eat-dog place. "That's totally wrong," he says. "I actually think there are people who get a 5th or 6th or 10th chance to find their niche."
In any event, Beyster's track record is hard to argue with. In its first year, his startup had $243,000 in sales; he has since increased that figure 25,000-fold. Besides, he doesn't have time for critics. He's far too busy dreaming up ways for SAIC to continue to push the frontiers of technology, and giving up the day-to-day responsibilities of the CEO job will free up more time for that. Sitting in an SAIC conference room, he's lean and gray but looks much younger than his 78 years, and his eyes light up as he talks about new SAIC data-mining projects he believes will revolutionize counterterrorism. His team has another contract to create software for a battalion of 70 miniature spy robots that operate autonomously and could swarm a building or battlefield -- something like the mechanical insects that hunt down Tom Cruise in Minority Report. Beyster believes that SAIC could be a $10 billion company before the end of the decade, and the work that he's doing still fills him with purpose. "I now think this company is an asset to the country," he says. "I'd like to see it continue." A few more Beyster books, it seems, remain to be written.
This site contains copyrighted material the use of which has not always been specifically authorized by the copyright owner. We are making such material available in our efforts to advance understanding of environmental, political, human rights, economic, democracy, scientific, and social justice issues, etc. We believe this constitutes a 'fair use' of any such copyrighted material as provided for in section 107 of the US Copyright Law. In accordance with Title 17 U.S.C. Section 107, the material on this site is distributed without profit to those who have expressed a prior interest in receiving the included information for research and educational purposes. For more information go to: http://www.law.cornell.edu/uscode/17/107.shtml. If you wish to use copyrighted material from this site for purposes of your own that go beyond 'fair use', you must obtain permission from the copyright owner.