Part IThe Story
It was raining in Boston. This is how Irwin Jacobs tells it, six decades on, to a room of college students at a Hillel center in San Diego: the subway to Harvard Square, then a bus the rest of the way home, then the door. Joan was standing in it. She had found a house. The couple had been looking around Boston for some time, and this one was contemporary, the kind of place she liked, and there was only one problem with it, which she delivered at the threshold while her husband was still wet. The house was in La Jolla.
The joke lands because of what it compresses. A year earlier, on leave from MIT after writing a textbook on digital communications, Jacobs had loaded Joan and their four sons into an old van and driven across the country to spend time at the Jet Propulsion Laboratory and Caltech. They fell for California. Then they drove home to Cambridge, and when the University of California's new campus at San Diego asked him to come help build its engineering program, he said no. The family had friends there, roots, careers. An assistant professor's salary at MIT, as he would later recall, ran about $5,500 a year, padded out with consulting. Moving three thousand miles to a university that barely existed, in a city that had not yet decided what it was, did not seem like the right move.
Then the rain, the bus, the door, the house. That was enough to flip the conversation. They packed the boys back into the van and drove west again. It took him, he says, "about six minutes or so" to feel at home.
He tells it lightly, but there is a second detail he offers in the same breath, almost as background: before the middle of the 1960s, Jewish families could not easily buy homes in La Jolla. That began to change, he says, with the growth of the university and institutions like the Salk Institute. So the house Joan described in the doorway sat in a town that, not long before, would not have wanted to sell it to them. The man who came through that door would go on to found the largest public company in San Diego, to put his and his wife's names on a medical center, an engineering school, a concert hall, a band shell on the bay, and to be dubbed, by the local paper in 2011, the city's "Philanthropist in Chief." In 1968 he put a down payment on a house in La Jolla Shores. Forty-five years later, asked to name his most prized possession, he named the house.
It is tempting to read the whole life as a story about a technology — code-division multiple access, the radio scheme that lets many people talk at once on the same slice of spectrum, each distinguished from the others by a code only the intended receiver knows. And it is that. But CDMA has a peculiar engineering problem at its center, one that its critics in the late 1980s believed would kill it, and the problem is a problem about distance and belonging. Phones close to the tower shout. Phones far away whisper. Left alone, the near drown out the far. Jacobs's company built its fortune on the fix.
By the Numbers
The Jacobs Ledger
7Linkabit alumni who founded Qualcomm on July 1, 1985
4,901,307Qualcomm's first CDMA patent number, filed 1986
$25MPrice M/A-COM paid for Linkabit, completed August 1980
2,621%Qualcomm's one-year stock gain after shedding manufacturing in 1999
$460M+Gifts from Joan and Irwin Jacobs to UC San Diego over 50 years
60+San Diego wireless startups descended from Linkabit, 1971–2000
The Counselor Was Wrong
New Bedford, Massachusetts, was once the richest city per capita in the United States. That was in the whaling days, when the money came in on ships that had been gone for years, and by the time Irwin Mark Jacobs was born there in October 1933 the money had long since gone out with the tide. It is a poor city now, Jacobs says, one with a low rate of college graduates, and he returns to it with scholarships, the way people return to a town that did not quite expect them to leave. His family ran a small restaurant. There was, in the Jacobs household and in the household of the girl he would marry, a little blue box for coins — the Depression, then the war, money was short, but "there was always that feeling that one should contribute whatever might be possible for you to do."
He was good at math, chemistry, physics. In 1950, a high school guidance counselor looked at this boy and offered a forecast that deserves to be preserved in amber.
— Irwin Jacobs, MIT Commencement Address, June 3, 2005My high school counselor advised me, and this was 1950, that there was no future in science, nor in engineering. And since I didn't really have a measure to evaluate that, I then took his advice.
So the son of a restaurant family enrolled in the School of Hotel Administration at Cornell. He might have run a very good hotel. What saved him from it was a roommate, an engineer, who spent a year and a half complaining about how hard it was to make grades in engineering — and somewhere in the complaining, Jacobs heard the thing he actually wanted. He transferred to electrical engineering. He was a co-op student, which meant working alongside practicing engineers, and one of them told him to go to graduate school. In his last term at Cornell he took a course on the theory and practice of vacuum tubes and built a couple himself, an artifact he would later invoke to a 2005 MIT audience as evidence of how fast the floor moves under a career. He graduated in 1956.
Two years earlier he had married Joan Klein. They had met at Cornell's Hillel, where Joan was the kind of student who made things happen — she helped start the bagel-and-lox breakfasts that Hillels across the country still hold — and in 1954, in New York City, the Cornell Hillel rabbi married them. She had gone to Cornell as a New York State resident under a special fee arrangement; he had gone on scholarships and fellowships. "We both felt we owed back a little bit," he would say, almost seventy years later, in the flat cadence of a man stating a known quantity.
MIT gave him a master's in 1957 and a doctorate in 1959, under Edward Arthurs, in a department where Claude Shannon, the father of information theory, was on the faculty. He stayed to teach. His office was next to Jack Wozencraft's — a Shannon student, later the founding chairman of the Naval Postgraduate School's command-and-control group — and the two became friends and then co-authors, building a book out of their class notes and their students' complaints. Principles of Communication Engineering appeared in 1965. It was described, later, as the first comprehensive textbook on digital communications, and it is still in use. One of the students in that era was Peter Denning, who would become a computer scientist of some consequence himself.
The book did something its authors had not planned. It made Jacobs famous among the people who needed what it described, and the requests for consulting began to arrive. That is the hinge on which the rest of the life swings: not an idea for a company but a reputation for understanding, and a growing pile of people who wanted to rent it.
Many Voices, One Channel
In San Diego, the requests multiplied. Local companies wanted to learn about digital communications, and Jacobs — the professor who had spent years telling students that communications theory would be useful in the real world — kept getting asked to prove it. On a flight back from NASA's Ames Research Center he mentioned the consulting load to two friends from UCLA, Andrew Viterbi and Leonard Kleinrock. Why not start a company, they said, and share the work.
Linkabit was formed in Los Angeles in mid-1968. Kleinrock soon left to help build something called the ARPANET, which is its own story. Viterbi stayed. The two had been engineering classmates at MIT and had reunited at an academic conference in 1959; Viterbi had gone the research route, helping design telemetry for Explorer I, the first successful American satellite, and had worked out the decoding algorithm that would carry his name into every cell phone and deep-space probe built afterward. If Jacobs was the translator, Viterbi was the source text. Their partnership would last through two companies and roughly forty years, and when the IEEE handed out its Medal of Honor in consecutive cycles, Viterbi got his in 2010 and Jacobs in 2013, with Morris Chang and John Hennessy in between, as though the profession were reading their names off a single plaque.
Linkabit grew, by Jacobs's account, 50 to 60 percent a year, profitably, mostly on government work. From 1969 to 1976 its people applied coding theory to make NASA's communications with the Voyager probes stronger and more reliable — the whisper from the edge of the solar system made legible. It designed and ran SATNET for DARPA. In 1974 it did something small companies did not do: it designed and contracted out the production of its own microprocessor, the Linkabit Microprocessor, described that year in a conference paper by a young engineer named Klein Gilhousen, and built a line of military satellite terminals around it. Jacobs would later explain the logic to a room of naval officers. Other companies were attacking the satellite problem the traditional way, in hardware. Linkabit built around a processor. It was a bet that the flexible, programmable approach would win — a bet that digital beats analog not because it is more elegant but because it can be changed.
The sale to M/A-COM was announced in November 1979 and closed in August 1980 for $25 million. For the first time, Joan and Irwin Jacobs had real money. Jacobs and Viterbi stayed on inside the acquirer; by 1983 Jacobs had risen to an executive vice presidency of M/A-COM, a public company headquartered back east. He kept the charitable giving going inside the new parent, and he kept a speech in his pocket for the day a shareholder stood up and accused him of spending their profits on charity. The speech was about retention. Losing an employee, he would argue, is the most expensive thing that can happen to a company; if civic giving made San Diego a place people wanted to stay, it paid for itself.
No one ever asked. He kept the answer ready for forty years.
In April 1985, Jacobs and Viterbi left. M/A-COM then sold Linkabit off in pieces: the VideoCipher television scrambling unit to General Instrument in 1986, the commercial satellite business in 1987 to form the nucleus of Hughes Network Systems, the government contracting remainder and the Linkabit name itself to the Titan Corporation in 1990. Eventually the name came to rest inside L3Harris. A coda worth noting for its irony: in January 1986, M/A-COM hosted the first-ever meeting of the Internet Engineering Task Force at its San Diego Linkabit facility — nine months after the two men who had built the place walked out.
What Linkabit left behind was not a product. It was people. The economist Joel West, who has traced the genealogy, counts more than 200 wireless-related startups formed in San Diego between 1971 and 2000; more than 60 of them were Linkabit spinoffs or spinoffs of spinoffs. That is a rate, West calculates, twice that of Fairchild Semiconductor, the legendary progenitor of Silicon Valley. ViaSat was founded by three Linkabit alumni in 1986. The best-known spinoff had been founded the year before.
A Company Without a Plan
On July 1, 1985, seven former Linkabit employees incorporated a company and called it Qualcomm, for "Quality Communications." Irwin Jacobs, Andrew Viterbi, Franklin Antonio, Adelia Coffman, Andrew Cohen, Klein Gilhousen, Harvey White. Jacobs led. What they did not have is the most quoted thing about the founding, and Jacobs has never been shy about it.
— Irwin Jacobs, MIT Technology Review, January 2011When we started Qualcomm, we didn't have a specific product in mind, and we didn't have a business plan. But we knew that wireless and digital would be of interest and that we would try to be innovative.
It began as a contract research shop, largely for government and defense work. In mid-1985 Hughes Aircraft hired it to research and test a satellite network proposal for the Federal Communications Commission, and in the course of that work Qualcomm's engineers kept returning to spread spectrum — the military technique of smearing a signal across a wide band of frequencies, so that it looks like noise to anyone without the key. In 1986 the company filed its first CDMA patent, No. 4,901,307. It would become, by one count, among the most frequently cited technical documents in history. The FCC project itself died in 1988, when the commission told all twelve vendors who had submitted proposals to merge into a single joint venture. Qualcomm, in other words, got the patent and lost the contract. This would turn out to be a pattern.
The money, in those years, came from trucks.
In 1988 Qualcomm merged with a company called Omninet and raised $3.5 million to build Omnitracs, a two-way satellite messaging and tracking system for long-haul trucking fleets. It is about the least glamorous product imaginable in a story that ends in every smartphone on Earth: a box in a truck cab that let a dispatcher in Green Bay know where a driver was on Interstate 80 and send him a text about his next load. It worked. Qualcomm went from eight employees in 1986 to 620 in 1991. By 1989 it had $32 million in revenue, and half of that came from a single customer, Schneider National. Jacobs did not take the Omnitracs profits home. He poured them into the cellular idea that most of the industry thought was a fantasy.
Here is the thing to understand about Omnitracs. It is not a detour. It is the method. The professor who consulted to fund Linkabit, and the Linkabit that did defense contracts to fund its own microprocessor, became the Qualcomm that tracked trucks to fund a revolution. At every stage the near thing paid for the far thing. The truck paid for the phone.
The Holy Wars
In 1989 Qualcomm submitted its CDMA scheme to the Cellular Telecommunications Industry Association as the digital standard for second-generation American cell networks. A few months later the CTIA rejected it in favor of time-division multiple access, TDMA, the more established approach championed by Ericsson. TDMA slices time: each phone gets a turn, a few milliseconds of channel, then waits. CDMA does not take turns. Everyone talks at once, all the time, across the whole band, and the receiver untangles them by code.
The objection was the near-far problem. In a CDMA cell, the phone standing next to the tower arrives at the receiver loud as a foghorn; the phone two miles out, behind a hill, arrives as a murmur. If they share the channel, the loud one buries the quiet one, and the scheme collapses under exactly the conditions — many users, uneven distances — that a commercial network has to handle. Serious engineers believed this made CDMA unworkable at scale. They were not fools. They were describing physics.
In 1989 Qualcomm filed three more patents. The first was for power control: the network constantly tells each phone how loud to transmit, turning down the near ones and turning up the far ones, hundreds of times a second, so that every voice arrives at the tower at roughly the same strength. The second was for "soft handoff," a method of passing a caller from one tower to the next without dropping the call — the phone is briefly heard by both, held by both, before the first lets go. The third was for a variable-rate encoder that used less bandwidth when a caller wasn't speaking. Silence, under this scheme, became a resource; the pause in your sentence made room for someone else.
It is hard to overstate how much of what followed is contained in those three filings. They are the engineering. They are also, read slantwise, a set of convictions about how a crowded system ought to behave: that the strong should be asked to speak more quietly so the distant can be heard; that transitions should be overlapping rather than abrupt; that what you don't say has value.
The CTIA said no, so Jacobs went around it. Once the FCC made clear that carriers could deploy standards the industry association had not blessed, Qualcomm took CDMA directly to the carriers, and the fight that followed is remembered in the industry as the Holy Wars of Wireless. Qualcomm's weapon was the demonstration. It ran tests in San Diego in 1989 and in New York City in 1990. That year Nynex Mobile Communications and Ameritech Mobile Communications became the first carriers to choose CDMA over TDMA. Motorola, previously in the TDMA camp, ran CDMA trials in Hong Kong and Los Angeles. AirTouch paid for a $2 million trial network in San Diego. In November 1991, fourteen carriers and manufacturers ran large-scale CDMA field tests together. In 1993, the CTIA reversed itself and adopted Qualcomm's CDMA as the IS-95A standard, to howls from companies that had already spent heavily on TDMA, and from Ericsson.
The money was always running out. Qualcomm lost money through much of the early 1990s on CDMA research, and in September 1991 it went public, raising $68 million. Jacobs was, in effect, selling a future to anyone who would buy it, and the press noticed. On September 6, 1996, the Wall Street Journal ran a front-page piece by Quentin Hardy under a headline that reads, today, like a period piece:
— Headline, The Wall Street Journal, September 6, 1996Jacobs's Patter: An Inventor's Promise Has Companies Taking Big Cellular Gamble — Qualcomm Boss's Innovation In Digital-Phone System Is Problematic, and Late — Are Claims Hope or Hype?
Patter. The word sits there like a raised eyebrow. By then, though, the gamble was mostly placed. In 1991 Qualcomm had agreed with South Korea's Electronics and Telecommunications Research Institute to develop CDMA jointly for that country's infrastructure; in May 1993 Korea made CDMA its national wireless standard, and commercial networks launched there in 1996. Hong Kong built the first commercial-scale CDMA network in 1995. On July 21, 1995, PrimeCo — a joint venture of Bell Atlantic, Nynex, US West and AirTouch — announced CDMA service across fifteen states. Qualcomm raised another $486 million that year by selling 11.5 million more shares, to pay for building the phones and base stations carriers needed. Revenue was $383 million in 1995 and $814 million in 1996. By 1997, CDMA held 57 percent of the American market; TDMA held 14.
By 2001, 35 percent of Qualcomm's revenue came from South Korea. A country that had wanted to leapfrog had been carried over the fence on a technology an American industry association once refused.
The Art of Taking Things Apart
The victory had a problem, and the problem was that no one else would build the hardware. If carriers were going to run CDMA networks, somebody had to make CDMA base stations and CDMA phones, and in the early 1990s that somebody was mostly Qualcomm, a company whose manufacturing experience consisted of a small ASIC team supporting Omnitracs. In May 1993 it agreed to supply 36,000 CDMA phones to US West. There were delays. There were manufacturing problems. In 1994 it partnered with Northern Telecom and formed a joint venture with Sony to borrow expertise it didn't have.
By 1997, Nokia, Samsung and Motorola were selling their own CDMA phones, and prices fell. In March of that year Qualcomm introduced its own handset, the Q phone, and Motorola promptly sued, alleging it copied the StarTAC; the suit settled in 2000. The infrastructure business was worse. The base station division, by one account, never sold another base station after its tenth sale and was losing $400 million a year.
Here the founder did the thing founders find hardest. He took his company apart.
In 1998 Qualcomm restructured and laid off 700 people. In 1999 it sold its infrastructure division to Ericsson — the same Ericsson that had backed TDMA, the same Ericsson Qualcomm had been fighting in a patent dispute since 1996 — as part of an out-of-court settlement. The timing was not leisurely. The International Telecommunication Union had warned it would exclude Qualcomm's CDMA from third-generation standards entirely unless the Ericsson dispute was resolved; the companies reached agreement one month before the ITU's deadline, cross-licensed their technology, and agreed to work together on 3G. In December 1999, Qualcomm sold its handset business to Kyocera, a Japanese licensee.
What remained was a company that designed chips and licensed patents. In 1999 Qualcomm's stock rose 2,621 percent, the fastest-growing stock on the market. By 2000 the company had 6,300 employees, $3.2 billion in revenue and $670 million in profit. A Fortune headline that May called it a $4-billion-a-year hotshot from "surf's-up San Diego" that wanted to be the next Intel.
There is a cost line in this ledger that rarely appears in the triumphant version. About 1,200 employees went to Ericsson with the infrastructure division. Many were unhappy working for a competitor and losing their Qualcomm stock options — options that would have been worth a fortune in the year their former company's shares rose twenty-six-fold. The dispute dragged on; by 2005, Qualcomm had paid $74 million in settlements. The people who built the base stations that proved CDMA worked were, in a sense, the ones nobody turned the volume up for.
The shape of the business that emerged from 1999 would define Qualcomm's next quarter-century, and also its quarrels. Because so many of its patents were essential to industry standards, it agreed to license them on "fair, reasonable, and non-discriminatory" terms, and charged, by one estimate, around 5 percent or $30 per device — five to ten times what other patent holders typically collected, according to Fortune. Qualcomm's position was simple: its patents cost more because they mattered more. Licensees, competitors and regulators on four continents disagreed, loudly. When European interests pushed WCDMA, a 3G standard that leaned less on Qualcomm's patents, Qualcomm at first refused to license its intellectual property for it, and later claimed 38 percent of WCDMA's essential patents; a research paper sponsored by European GSM interests put the figure at 19. That gap — 38 versus 19, two numbers for the same thing — is a fair summary of the next two decades of litigation: Broadcom, Nokia, the Korea Fair Trade Commission, China's regulators, Taiwan's, the European Commission, Apple, the Federal Trade Commission. In August 2020, the Ninth Circuit reversed a lower-court antitrust ruling and characterized Qualcomm's practices as "hypercompetitive" rather than "anticompetitive." It is a distinction a company could put on a coat of arms.
Most of that fighting happened after Jacobs left the chief executive's office. But the architecture was his: a company whose most valuable asset is its intellectual claim on a standard everyone must use. He had bet that the world would one day need a channel where everyone talks at once. Having won, he charged rent on the code.
Irwin Jacobs May Refer To
Search for his name and the encyclopedia stops you at a fork. Irwin Jacobs may refer to: Irwin M. Jacobs, born 1933, co-founder of Qualcomm; Irwin L. Jacobs, 1941 to 2019, Minneapolis-based investor. Two signals with the same name on the same channel. You need the code to tell them apart.
The second Irwin Jacobs was also born to a Jewish family and also came up in a small family business — his father's, the Northwestern Bag Company. He enrolled at the University of Minnesota and attended for one day, so that he could keep working with his father. As a teenager he was already putting together close-out merchandise deals. In 1975, at thirty-three, he bought the ailing Grain Belt Brewery for $4.1 million, spent eight months trying to stop it from losing nearly $200,000 a month, then liquidated it, sold the brand to G. Heileman and cleared $4 million. He read in the Wall Street Journal that W.T. Grant had gone bankrupt and bought its $276.3 million in consumer receivables for $44 million plus a cut of the first year's sales. In the 1980s he took big stakes in Kaiser Steel, Walt Disney, Castle & Cooke, pressed ITT's Rand Araskog to break up his conglomerate, and dismantled AMF. The press called him "Irv the Liquidator." A longtime lieutenant explained the philosophy in one line to the New York Times: "He firmly believes that you make your money when you make your buy."
The Qualcomm Jacobs made his money a decade after his buy, if you count the founding as the buy; for most of those years he was losing it on a radio scheme his industry had formally rejected. One Irwin bought companies to take them apart. The other built one and then, in 1999, took it apart himself — not to cash out but to save the part that mattered. In June 1996, three years before Qualcomm shed its factories, the raider told the Times that the era of men like him was ending because companies had learned to do it to themselves: "Look at how AT&T and all these other companies are breaking up without being forced to." He might have been describing San Diego.
The rhymes keep coming once you listen for them. Both men married young and stayed married for more than half a century. Both gave: the Minneapolis Jacobs donated $8 million to the Special Olympics in 1991 and once served as its chairman; his daughter Sheila has cerebral palsy. Both understood the value of the person on the ground — the raider called a January 1996 meeting with Watkins's top hundred salespeople, after a disastrous price hike on starter kits from $81 to $99, "the best learning experience of my career," and cut the kit to $29.95. "I want to be the nicest guy in town," he said, "but I don't want to be the nicest because I'm the stupidest." It is a line the other Irwin Jacobs could have said, though he never would have said it that way.
On the morning of April 10, 2019, Irwin L. Jacobs and his wife, Alexandra, married fifty-seven years, were found dead in their home on a 32-acre estate between Lake Minnetonka and Lake Tanager. She had been using a wheelchair and showing signs of dementia, and he was reported to be distraught over her condition. The medical examiner concluded he had shot her and then himself. The house, which he had bought at thirty for $340,000 and put on the market in 2014 for $22 million, was demolished in 2021 to make room for six new homes.
There is no lesson here, and it would be indecent to manufacture one. Two men with the same name, a few hundred miles of family history apart, each building a life out of deals and devotion. The encyclopedia lists them one above the other and asks you to choose.
Soft Handoff
The hardest moment in a cellular call is the one where you move. The tower that has been carrying your voice is fading; the next tower is coming up; the old way of handling this, the "hard" handoff, cut you loose from one before attaching you to the other, and sometimes in the gap you fell. Qualcomm's 1989 patent let both towers hear you at once for a moment. The connection overlaps. Nobody drops.
Jacobs was chief executive of Qualcomm until 2005. That year his son Paul — one of four boys who had ridden west in the old van, an engineer who had come up through the company — became CEO. Irwin stayed on as chairman, and in March 2009 announced that he was stepping down from that post too and that Paul would succeed him. He retired from the board in 2012. The New York Times, in June 2011, ran a story whose headline did the analysis on its own: "At Qualcomm, Rise of Founder's Son Defies Hazards of Succession."
The handoff was soft in the technical sense — years of overlap, the father present while the son took the load — and also in another sense. Paul did not simply extend his father's company. Where Irwin had built Qualcomm around CDMA patents, Paul redirected much of its new research toward what would become the Internet of Things, and the Snapdragon processor line, first shipped in November 2007, turned Qualcomm into the engine inside most Android phones. The father had argued for CDMA in the 4G era; Qualcomm backed a CDMA-based standard, UMB, that no carrier adopted, halted it in 2008, and shifted to LTE, which leaned less on its patents, and then bought its way back in. In January 2011, at seventy-seven and long since off the chief executive's desk, Jacobs bristled when an interviewer suggested the company was reinventing itself. "I wouldn't call it 'reinvention.'" He preferred evolve. The towers overlap.
In December 2013, Qualcomm announced Steven Mollenkopf would succeed Paul. The family's direct hold on the company ended. In late 2017, Broadcom offered $103 billion for Qualcomm, raised it to about $121 billion, and went hostile — a raid of the old Minneapolis kind, conducted at a scale the Liquidator never imagined — until President Trump blocked it by executive order on March 12, 2018, citing national security. The company Jacobs founded with no business plan had become a strategic asset of the United States.
The handoff kept going in other directions too. Gary Jacobs, the eldest son, became head of the board of a charter school, High Tech High. A granddaughter, Sara Jacobs, won California's 51st congressional district as a Democrat, with significant funding from her grandfather; Vox put it bluntly in 2020 under the headline "A tech billionaire spent millions to elect his granddaughter. It's working." And a few years ago, just before the pandemic, a couple of grandchildren came to Irwin and Joan with a request. Could they arrange a weekend, just the grandchildren and the grandparents, to talk about philanthropy — the old couple's and their own? The weekend happened at Rancho Valencia. The grandchildren set one condition, which Jacobs reports with evident delight: no parents.
The Little Blue Box
The first real gift came after the M/A-COM sale. Jacobs and Joan bumped into Dick Atkinson, then chancellor of UC San Diego, and said that if the university was interested, they would like to endow a faculty chair. UCSD was a public university; it was not supposed to be competing with private schools for donors and did not really have a fundraising operation. A few weeks went by. Nothing. Then Joan ran into Atkinson in the grocery store and asked why they hadn't heard back. His answer, as Jacobs tells it: "Were you really serious?"
They were. It would take the University of California some time to understand how serious. Over the next fifty years, Joan and Irwin Jacobs gave UC San Diego more than $460 million — the Irwin and Joan Jacobs School of Engineering, the 245-bed Jacobs Medical Center, chairs, fellowships, scholarships. By 2009 they had given $31 million to MIT; $62 million across the Technion, Cornell and KPBS. They pledged $120 million to the San Diego Symphony, and Copley Symphony Hall became the Joan and Irwin Jacobs Music Center. $20 million for a new downtown central library. $5 million in 2002 to renovate an old train station baggage building for the Museum of Contemporary Art San Diego. In April 2013 they gave $133 million toward Cornell Tech on Roosevelt Island, the applied-science campus New York's mayor Michael Bloomberg had dangled land and $100 million to attract, a competition everyone expected Stanford to win; Jacobs helped pair Cornell with the Technion, and they won it instead. In January 2022 the ACLU Foundation renamed its list of pending Supreme Court cases the Joan and Irwin Jacobs Supreme Court Docket, honoring a $20 million gift, the largest one-time endowment gift in its history. In August 2010 they signed the Giving Pledge; Jacobs had been at one of Bill Gates's early dinners where the idea was still being kicked around.
At the Salk Institute, where Jacobs chaired the board for a decade and oversaw a $330 million capital campaign that beat its goal by $30 million, the couple's most characteristic move was not a gift but a challenge. For every $2 million someone else contributed toward an endowed chair for a senior scientist, they would add $1 million, reaching the $3 million needed to fully endow it. Twenty chairs were offered. All twenty were filled. It is power control by another name: you don't shout over the room; you adjust the levels until the whole room can be heard.
His criteria, stated across many interviews, were never elaborate. Pick areas that interest you. Pick projects where you can make a difference. Look for good leadership — "somebody that has thought through that project," as he put it in 2024, someone who will show the excitement. On names, he was more ambivalent than the skyline suggests. Putting one's name on a building, he said, was "always a question," and the couple generally let the institution decide whether the name would help. The argument for it was the same argument as the Salk match: visibility recruits. Walk to the UCSD health campus on the east side of Interstate 5, he points out, and several of the buildings carry the names of former Qualcomm people.
Not everything worked. The most famous failure is in Balboa Park, where Jacobs funded an engineering study in 2010 and championed a plan to clear cars from the Plaza de Panama: a new bridge and bypass curving south off the Cabrillo Bridge, a three-level underground garage beneath what would become a large planted picnic ground behind the Organ Pavilion, the old plaza restored to pedestrians. Some environmentalists objected, preferred a different route, and sued. By the time the suits were cleared, costs had risen, the city council had capped them, and although much of the money was philanthropic the cap killed the project. Asked about it in 2024 by a foundation president who called the plan visionary, Jacobs laughed and named it without being asked twice. Then he changed the subject to New York, and Cornell Tech, where students had already started well over a hundred businesses.
If a single title could hang over the family's habit of acting on opportunities before they quite made sense — the van, the door, the chair, the plan-less company — it might be the one on a business-shelf book called Just Say Yes, with its subtitle about life and luck. Jacobs's own version was more conditional and more patient. He said yes to the house in La Jolla on the second try. He said yes to the chair and then waited for the chancellor to believe him. He said yes to the Jeff Koons sculpture he and Joan saw in a picture in Basel, put down a deposit, and paid it off over years — and then waited fifteen of them for it to be finished.
Everything Has an Interesting Story
He would rather talk about what is next. Ask him, at ninety, about legacy, and he talks about artificial intelligence and robotics, which he expects to greatly increase productivity over the coming decade; about the question of where the resulting profits go, which he frames as a question of medical care for all and housing; about the leisure hours people will need to fill and why that means building more concert halls and maker spaces, not fewer. Ask about the border and he talks about the haircuts — how they used to take the boys down to Tijuana for haircuts and to buy rolls, back when you could cross quickly — and then about why the country needs more immigrants, not fewer, "a lot of us are at best second generation." Ask what he would be in another life and he says an architect. San Diego "has not done the best job on its architecture," and the empty office towers left by remote work are, to him, an opportunity. "Change is not necessarily bad. Change gives you opportunity to do better things."
The podcast in which he said most of this was recorded in the spring of 2024, just before Joan died. She was ninety-one. They had been married seventy years — since the Cornell Hillel rabbi, since the bagels and lox, since before the van and the textbook and the rain. The hosts released the conversation anyway, with a note explaining the timing, and listening to it you hear a man still using the first-person plural in nearly every sentence. We came here. We bumped into Dick Atkinson. We opted, normally, to let the organization use our names.
On March 15, 2025, at UC San Diego's Chancellor's Medal Awards, the University of California's president, Michael Drake, presented its highest honor, the UC Presidential Medal, to Irwin Jacobs, and a second medal to Joan, in memoriam. He accepted hers. The medal had been established in 1997 by the then president of the University of California, Richard C. Atkinson — the same Dick Atkinson who, years before, stood in a San Diego grocery store and asked Joan Jacobs whether they had really been serious.
A year later, at ninety-two, Jacobs sat on a stage at the Melvin Garb Hillel Center at San Diego State and told students the story of the rainstorm and the door. He urged them to keep the bagel-and-lox breakfasts going, the ones Joan had helped invent. He offered a consumer tip: "Lox from Costco is very good by the way." The room roared.
He has a way of saying, when one story leads to another, that everything has an interesting story. The party hat does. Walk into the Jacobs Medical Center at UC San Diego by its west side door and you come upon it: an enormous, gleaming, candy-colored Koons party hat, the one he and Joan first saw in a photograph in Basel and waited fifteen years to receive, and then waited almost another year for, because the pandemic arrived and the German technicians who had to install it could only get into the country by crossing through Tijuana. They could have kept it at home; collectors kept calling with offers to buy it back. They put it at the hospital door instead. And there, most days, a mother and father come out of the maternity suites carrying a person who did not exist a few days before, someone who has just arrived in San Diego from elsewhere, and they stop, and turn, and take a picture by the hat.
Part IIThe Playbook
What follows is an attempt to extract the operating logic from Irwin Jacobs's six decades as professor, founder, chief executive and donor. The principles are drawn from the record above — the textbook that became a consulting business, the truck terminal that paid for a cellular revolution, the 1999 divestiture, the challenge grants — and they are meant for people building things now. Some are uncomfortable. Jacobs's success included a licensing model that regulators on four continents fought for twenty years and a restructuring that cost 1,200 employees their options. A playbook that hid that would be patter.
The principles
- 1.Let the near thing pay for the far thing.
- 2.Found on capability, not on a plan.
- 3.Patent the answer to the objection.
- 4.When the gatekeeper says no, go to the buyer.
- 5.Demonstrate in public, at scale, repeatedly.
- 6.Find the market that wants to leapfrog.
- 7.Build what you must; sell what you shouldn't own.
- 8.Price your essentiality, and budget for the war.
- 9.Bet on the programmable version.
- 10.Make the institution a seedbed.
- 11.Give like an operator: match, don't shout.
- 12.Hand off with overlap.
Principle 1
Let the near thing pay for the far thing
Every stage of Jacobs's career was financed by a less glamorous adjacent business. Consulting paid for Linkabit. Defense and NASA contracts paid for a proprietary microprocessor. Omnitracs — satellite messaging for trucking fleets, half its 1989 revenue from a single customer, Schneider National — paid for CDMA research the cellular industry had formally rejected. The pattern is not a compromise of the vision; it is the mechanism by which the vision survived the decade it took to be proven right.
The discipline here is that the near business must be genuinely good, not a side hustle. Omnitracs grew Qualcomm from eight employees to 620 in five years. It was a real product with real customers, which meant real cash flow and real operating experience — the company learned to ship hardware to people who would complain when it broke. A cash business that is also a training ground is worth more than venture money that buys only time.
Tactic: Identify the most boring product your core capability could sell profitably this year, build it well, and route its margin explicitly into the long bet.
Principle 2
Found on capability, not on a plan
"When we started Qualcomm, we didn't have a specific product in mind, and we didn't have a business plan." This is usually quoted as a charming confession. It is better read as a theory of company formation. Seven people who had spent years together at Linkabit, who shared a vocabulary and a set of technical instincts, incorporated around a conviction — that wireless and digital would matter — and then went looking for the problem. Within months a Hughes Aircraft contract had pointed them at spread spectrum, and the first CDMA patent followed in 1986.
What made this work was that the team had already been tested. Linkabit had grown 50 to 60 percent a year profitably; its people had shipped to NASA and DARPA. A planless company of strangers is a gamble. A planless company of colleagues with a decade of shared execution is an option on whatever they notice first. Jacobs's earlier career followed the same arc: the textbook created a reputation, the reputation created consulting demand, and the demand created a company. He never set out to start Linkabit.
Tactic: Before writing the plan, ask whether the founding team has already delivered something hard together; if not, do a paid contract project as a team first.
Principle 3
Patent the answer to the objection
The case against CDMA in 1989 was specific: the near-far problem would make it unusable in crowded commercial networks. Qualcomm's response was not rhetoric. It was three patents filed that year — power control, soft handoff, variable-rate encoding — each of which addressed a concrete weakness skeptics had named. The objection became the roadmap, and the roadmap became the moat.
This is the most transferable idea in the story. When serious people tell you why your idea will fail, they are handing you the list of inventions you will need. Solve them, and you own not just the idea but the conditions under which it works — which is far harder for a fast follower to route around.
Q
The Three Patents of 1989
Each filing answered a specific reason the industry believed CDMA could not scale.
| The objection | The invention | What it implied |
|---|---|---|
| Phones near the tower drown out distant phones. | Power control: the network adjusts each phone's transmit strength. | Capacity Every caller arrives at equal volume. |
| Calls drop when moving between cells. | Soft handoff: two towers hold the call briefly at once. | Continuity Overlap instead of a gap. |
| Bandwidth is wasted on silence. | Variable-rate encoder: less data when a caller isn't speaking. | Efficiency Pauses become shared capacity. |
Tactic: Write down the three strongest technical objections to your product from credible skeptics, and assign each one an owner and a filing date.
Principle 4
When the gatekeeper says no, go to the buyer
The CTIA rejected CDMA in 1989. Many companies would have treated that as a verdict. Jacobs treated it as a procedural fact and looked for the actual decision-makers. Once the FCC confirmed carriers could deploy standards the association hadn't approved, Qualcomm pitched carriers directly. By 1990 Nynex and Ameritech had chosen CDMA; by 1993 the CTIA had reversed itself.
The lesson is about where authority actually lives. Industry associations, standards bodies and analyst consensus aggregate the preferences of incumbents, and incumbents are rarely the ones who benefit from a step change. The people with the capacity problem — carriers running out of spectrum — had a reason to listen. Jacobs found them and made the association follow its own members.
Tactic: Map who pays for the pain your product solves, separately from who certifies or standardizes your category, and sell to the first group until the second has no choice.
Principle 5
Demonstrate in public, at scale, repeatedly
Qualcomm did not argue CDMA into existence; it ran it. San Diego in 1989. New York City in 1990. Motorola's trials in Hong Kong and Los Angeles. A $2 million trial network for AirTouch. Fourteen carriers and manufacturers in large-scale field tests in November 1991. Each demonstration converted a theoretical debate — the "Holy Wars" — into an empirical one, and empirical debates are winnable by the side with the better system.
There was a cost: the Wall Street Journal was still calling it "patter" in 1996, three years after the standard was adopted. Demonstrations don't end skepticism; they shrink it to a size the business can survive. Note too that the demonstrations were often paid for or co-run by the prospective customer. A test the buyer funds is a test the buyer has already half-believed.
T
From Rejection to Majority
How CDMA moved from a refused proposal to the dominant U.S. standard in eight years.
1986
Qualcomm files its first CDMA patent, No. 4,901,307.
1989
CTIA rejects CDMA for TDMA; Qualcomm files three patents answering the near-far problem; first San Diego demonstration.
1990
New York City demonstration; Nynex and Ameritech choose CDMA.
1991
Agreement with Korea's ETRI; IPO raises $68 million; 14-company field tests in November.
1993
CTIA adopts CDMA as IS-95A; Korea makes it the national standard.
1995
First commercial network in Hong Kong; PrimeCo commits across 15 states; $486 million follow-on offering.
1997
CDMA holds 57 percent of the U.S. market; TDMA holds 14.
Tactic: Replace your next pitch deck with a live deployment in the customer's own environment, ideally one they co-fund.
Principle 6
Find the market that wants to leapfrog
In 1991, while American adoption was still contested, Qualcomm agreed with South Korea's ETRI to develop CDMA jointly. In May 1993 Korea adopted it as the national standard; by 2001, 35 percent of Qualcomm's revenue came from South Korea. A country building its wireless industry from a comparatively clean slate had no sunk investment in TDMA to protect and every incentive to adopt the higher-capacity system.
Jacobs returned to this theme repeatedly. In 2011 he described India, where restricting mobile service to GSM had kept prices high and usage low, as a case of policy holding back a population that technology could have served. The insight is that the most enthusiastic customer for a disruptive standard is often not the largest market but the one with the least to unlearn.
Tactic: Rank prospective markets by how little legacy infrastructure they would have to abandon, and pursue the top of that list as aggressively as the biggest.
Principle 7
Build what you must; sell what you shouldn't own
Qualcomm entered manufacturing because no one else would build CDMA phones and base stations, and without them the standard was a paper victory. That was correct. Staying in manufacturing once Nokia, Samsung and Motorola were building CDMA phones and the base station unit was losing $400 million a year would have been ruinous. In 1999 Jacobs sold infrastructure to Ericsson and handsets to Kyocera. The stock rose 2,621 percent that year.
The comparison with the other Irwin Jacobs, the Minneapolis raider, is instructive. Both men took companies apart. One did it from outside, buying cheaply and liquidating. The other did it from inside, to concentrate the company on its irreplaceable asset. The second is harder, and it has a human cost that should be named: about 1,200 employees moved to a competitor and lost options, producing $74 million in settlements by 2005.
J
Two Irwin Jacobses, Two Ways to Break a Company
The same name, opposite theories of value.
| Question | Irwin L. Jacobs (Minneapolis) | Irwin M. Jacobs (San Diego) |
|---|---|---|
| Where is value made? | At purchase. You make your money when you make your buy. | In the decade of losses before the standard wins. |
| Why break a company up? | Extract To unlock value others mismanaged. | Concentrate To keep only what no one else can do. |
| Representative deal | Grain Belt Brewery, 1975: bought for $4.1 million, liquidated within a year. | 1999: infrastructure to Ericsson, handsets to Kyocera. |
| Hostile posture | Raider pressing ITT and others to split. | Founder of a company later targeted by a $121 billion hostile bid. |
Tactic: For every business unit you run, write one sentence on why only you can do it; sell or spin out the units where that sentence is weak, and design retention terms for the people who go with them.
Principle 8
Price your essentiality, and budget for the war
As the sole early investor in CDMA, Qualcomm owned much of what the standard required, and it priced accordingly: roughly 5 percent or $30 per device, by one estimate five to ten times what other patent holders typically charged. Qualcomm's argument — its patents cost more because they matter more — was coherent, and it built one of the most profitable licensing businesses in technology.
It also guaranteed perpetual conflict: Ericsson in the 1990s; Broadcom, Nokia and the "Project Stockholm" complaint in the 2000s; regulators in Korea, China, Taiwan and Europe; Apple and the FTC in the 2010s. Some fights Qualcomm won, as with the Ninth Circuit's 2020 reversal. Some it settled at great cost, including $891 million to Broadcom and $975 million to China. The principle is not that essential-patent pricing is wrong. It is that if you choose it, litigation is a recurring operating expense, not an anomaly.
Tactic: If your pricing depends on being indispensable, fund a standing legal and regulatory budget from day one and treat it as cost of goods, not contingency.
Principle 9
Bet on the programmable version
At Linkabit, Jacobs explained, competitors attacked military satellite terminals with traditional hardware. Linkabit built around a processor — eventually its own, the Linkabit Microprocessor of 1974. CDMA itself was a programmable bet: a digital scheme whose capacity could be improved by smarter coding and signal processing rather than by more spectrum. Qualcomm's later shape — chip designer and licensor, not manufacturer — is the same bet at the level of the company.
Programmable systems lose early benchmarks and win long careers, because they can be changed when the world changes. Jacobs's MIT commencement theme, delivered in 2005, was precisely that graduates should prepare for change they could not anticipate. He had built two companies on that premise.
Tactic: When choosing between a fixed-function solution that wins today's benchmark and a flexible one that can be updated, choose flexibility unless the market will close within two years.
Principle 10
Make the institution a seedbed
Linkabit's most important product was its alumni: more than 60 of San Diego's 200-plus wireless startups between 1971 and 2000 descended from it, twice the rate of Fairchild Semiconductor. Qualcomm and ViaSat were both founded by Linkabit people. Jacobs then extended the logic into philanthropy — the engineering school, scholarships, fellowships at KPBS that trained journalists who went on to other stations, Cornell Tech, whose students have started more than a hundred businesses.
This is ecosystem building without the jargon. Jacobs's own explanation of San Diego was that so many people had come from elsewhere that "we had to make friends," and that newcomers without old relationships cooperate more readily. A seedbed is a place where people who don't belong yet can become each other's network.
Tactic: Measure your organization partly by how many credible companies or institutions your former people start, and make leaving on good terms easy.
Principle 11
Give like an operator: match, don't shout
The Jacobses' signature philanthropic instrument was the challenge grant. At the Salk Institute, they added $1 million for every $2 million others gave toward a $3 million endowed chair; all twenty chairs were filled. Their stated filters were operational: areas they cared about, projects where a gift would make a difference, and good leadership — someone who had thought the project through. Names went on buildings when the institution thought it would help recruit other donors.
Like power control in a CDMA cell, the match adjusts the volume so that more voices are heard, rather than letting the loudest donor drown out the rest. The Balboa Park failure is the counterexample worth studying: a philanthropic plan that collided with public process, litigation and a city cost cap. Even the best-funded gift cannot skip consent.
Tactic: Structure your next significant gift as a match with a defined count of units to fill, and publicly name the leader you are betting on.
Principle 12
Hand off with overlap
Jacobs gave up the CEO title in 2005, the chairmanship in 2009, the board seat in 2012. For years he and his successor occupied the same structure, the way two towers briefly hold a single call. The Times noted in 2011 that the founder's son had defied the usual hazards of succession; the son, in turn, redirected the company toward Snapdragon and connected devices in ways the father might not have chosen.
The family extended the same principle to giving: the grandchildren's weekend at Rancho Valencia, with the grandparents present and the parents excluded, at the grandchildren's request. The older generation explained what it had done and why, then listened to ideas it had never considered. Overlap is not control. It is presence during the transfer.
Tactic: Set a multi-year transition calendar for your own role with explicit dates when each authority passes, and stay available but silent on decisions you have already handed off.
Free playbook
Get The Business Model Playbook
58 business models, one visual page each: how the money flows, the metrics that matter, and who runs it. Free when you join the Faster Than Normal email.
Free. No spam. Unsubscribe anytime.
Part IIIQuotes and Maxims
In their words
— Irwin Jacobs, MIT Commencement Address, June 3, 2005The opportunity for philanthropy never goes away.
— Irwin Jacobs, La Jolla Light, October 2013, on the character traits he most admiresI think it's important that you value the way others feel, that you try to find areas of mutual interest — but even where you have disagreement, you can understand that disagreement and still find ways to work together.
— Irwin Jacobs, MIT Technology Review, January 2011We listen to our experienced customers and then do a lot of R&D so that we can make them aware of what is doable. And then we'll do what is always done: try to run real fast and to innovate.
— Irwin Jacobs, Stop & Talk podcast, 2024So many of us came here from elsewhere. And so we're new to the area. We didn't have all these built in long term relationships. We had to make friends. We had to work with one another and build together.
— Irwin Jacobs, Stop & Talk podcast, 2024Change is not necessarily bad. Change gives you opportunity to do better things.
Maxims
- Distrust the counselor. A confident forecast from someone with no measure to evaluate it is an opinion, not a constraint.
- Let the truck pay for the phone. Finance the long bet with a real, boring, profitable business.
- Found on a team, not a plan. People who have shipped together can find the product after they incorporate.
- The objection is the roadmap. Each serious reason your idea will fail names an invention you need to own.
- Go around the gatekeeper to the customer. Standards bodies follow buyers who have already decided.
- Run it in public. Demonstrations shrink skepticism to a size you can survive.
- Sell what you shouldn't own. Keep only what no one else can do, and treat the people you let go as a cost you pay honestly.
- If you charge for being essential, budget for the war. Litigation is an operating expense of indispensability.
- Match, don't shout. The best gifts turn up other people's voices.
- Hand off with overlap. Stay present during the transfer, then let go of the call.
