CHAPTER 9
I bought my first house in the Seattle suburb of Bellevue, a four-bedroom, split-level contemporary that backed onto the woods and came complete with field mice. It had big picture windows and a spacious deck overlooking Lake Sammamish. I bolted my couch together and set up the big-screen, then unpacked my Laser-Discs and audiocassettes. My sister moved in with me for a while, and it felt as though I was home again. I’d traded in my Chevy Monza for a little black Mazda RX-7, which was fast and nimble and got me to home-cooked meals at my parents’ house in half an hour. Life was good.
With our new downtown Bellevue office to myself for a few days, I fired up our spanking new 2020, DEC’s smallest mainframe. For Microsoft, that purchase was both a rite of passage and a key to boosting our productivity: no more sharing time with junior high schools, no disruptions when some other operation monopolized or crashed the machine. Our new location resembled our old one in Albuquerque, on the eighth floor of a ten-story building owned by Old National Bank. We’d expanded from ten offices to maybe fifteen, with a good-size foyer for a receptionist. But we still had to walk single file in the corridors, sidling by boxes stacked with incoming hardware. My office was next to Bill’s, and we shared a secretary.
In April 1979, our BASIC interpreter became the first microprocessor software product to surpass a million dollars in sales. With more than three hundred thousand users in the United States and abroad, it was installed on more machines than any other single program. But we didn’t pause to celebrate. A jumble of 8-bit computers, many of them prototypes, crowded a group of tables near the programmers’ offices. I’d need to get BASIC ported onto each of them to consolidate Microsoft’s dominance in high-level languages.
If a hardware company used a BASIC that wasn’t ours, we’d disassemble it to see if they’d reverse-engineered our copyrighted code. If our suspicions were confirmed, a stern letter or two usually sufficed. If the code came from another company, we’d press the point that our BASIC was light years better. And it was, because we’d never stopped striving to add features and improve it.
Convinced that our future lay in the 16-bit world, I began work on a stand-alone 8086 BASIC with Bob O’Rear, an air force veteran who became my de facto deputy for development. We were still working on faith, since the first 16-bit microcomputer had yet to appear. That May I took a call from Tim Patterson, a young designer at a local hardware shop called Seattle Computer Products (SCP). He’d built a prototype computer with the 8086 chip mounted on a processor board and was hunting for software to test it. I told him, “Bring it on up. We’ve got something that might work.” Tim was an engineer after my own heart, someone who’d roll up his sleeves and dive into the knottiest problems. After a week of tinkering, both hardware and software passed a run-through. It was a useful collaboration that would lead to a more important alliance down the road.
That early 8086 initiative was just one example of our trying to stay ahead of an ever-accelerating game. We constantly feared that someone might be gaining on us. In those early years in Seattle, I had a disturbing, recurrent dream: Bill and I on the flight deck of a B-17, struggling to get hold of the plane while turbulence buffeted us all over the sky. We never crashed, but we never gained complete control, either. And there was no bailing out. We were strapped in for the duration.
THE JAPANESE MARKET was exploding, thanks in large part to Kazuhiko (Kay) Nishi, our flamboyant agent in East Asia. Kay published a chain of glossy computer magazines that worked hand in glove with his nonstop salesmanship for Microsoft. In August 1979, after he snagged a big contract with NEC, Bill and I went to Japan to help drum up more business. It was my first trip outside North America and everything was new to me, from our futon mats with wooden headrests to the multicolored plates of sushi and boiling pots of shabu-shabu.
We traveled first to Kobe, where Kay’s parents owned a girls’ school with an outdoor swimming pool. There were two diving platforms, one three meters high (plenty for me) and another at ten meters. A bunch of the girls watched between classes as Bill climbed to the top of the high dive. He jumped, feet first, and they screamed. He must have hit the water at a slight angle—when he pulled himself out, the whole front of his body was bright red. It must have stung, but it didn’t stop him. Bill kept jumping, and the girls screamed each time.
We took a bullet train to Tokyo, where we noted a development that had yet to catch on in the United States: the computer “superstore.” After checking in at the Hotel Okura, I ordered a hamburger with mustard from room service. I took a large bite, and instantly my sinuses began to burn. As I gasped, I saw Bill laughing at me. Even the mustard was different here.
As we headed out for our first meeting, the elevator stopped and a couple squeezed in: a long-haired guy with Coke-bottle eyeglasses and a Japanese woman with curly black hair. Could it be John Lennon and Yoko Ono? I leaned against the side of the elevator, trying to look casual, and inadvertently pressed an intermediate floor button. We stopped, the door opened, and the guy with the glasses said, “Nobody ’ome.” Now I knew it was John Lennon. I desperately wanted to say something but my brain froze. After we reached the lobby and the couple walked off, I said to Bill, “Did you see that? That was John Lennon and Yoko Ono.”
“Really?”
“Yes, look, there they go!”
And Bill said, “Oh, yeah, you might be right.” He wasn’t a Rolling Stone reader like me. He knew bits and pieces of popular culture, but he was thinking about the software business first, second, and third.
Kay Nishi was an unusually Westernized and entrepreneurial Japanese, a high-octane maverick who flew his own helicopter to business meetings. He was a big spender, piling up huge debt. (On a later trip, Bill was not amused after Kay excitedly took him by a Tokyo train station to show him “the dinosaur”: a $1 million, life-size, concrete brontosaurus built to promote a new joint PC format, with Microsoft on the hook for part of the bill.) Kay worked against the grain of Japan’s conservative business culture but opened many doors for us, including Matsushita Electric, now Panasonic. A junior technical person met us in reception to escort us to the top floor, where the chief technology officer awaited. As our elevator ascended, our guide looked more and more uncomfortable. I said, “Have you ever been to the top floor before?”
“Oh, no, never,” he said.
“Have you ever met with Mr. —— before?”
“Oh, no, never.” The poor man was sweating bullets.
The meeting room contained two large tables facing each other about six feet apart. There were a dozen people on Matsushita’s side; the chief took the middle chair, while the others flanked him in descending order of status. They smoked like chimneys and drank rocket-fuel coffee. Bill and I sat at the other table with Kay, who would fill us in later about any byplay among the Japanese.
“So, Mr. Gates,” the chief said, “how do we know that you’ll deliver on time?” They grilled us for four hours, standard procedure in Japan; they wanted to be sure that we could make good on our promises. Bill was confident and assertive. He’d go into his rocking mode and say, “Well, we’ve done this with Apple and this with Tandy. …” No one seemed to care that he was twenty-three years old. The more technical or speculative questions came to me: “So tell us, Mr. Allen, how do you see the future of the personal computer industry?” (After we got to know each other better, I’d become Allen-san.)
We toured the plant until five o’clock, at which point the air conditioning shut down and the company song was piped in over a sound system. Kay motioned for us to stand along with our hosts, who sang together, full out. When it was over, I asked Kay if they were going home now. “Oh, no,” he said. “They’ll work until eight o’clock and then go out to eat with their buddies, and then they’ll start again really early tomorrow.” I was thinking: These guys are working a lot harder than the average American. How in hell can we keep up? And for the most part, we couldn’t. The migration of consumer electronics manufacturing out of the United States was already well under way.
Each night we’d get taken out by Japanese executives on expense accounts. They chose European restaurants, a big treat for them, until we finally pleaded for some Japanese food. Toward the end of our stay, one executive said, “That was a great meeting. I’d like to invite you to something special, a geisha house or a really great dinner. You choose.” Bill looked at me and I knew without asking that he’d vote for the geishas. “A nice dinner sounds great,” I said. The man reserved a private room at one of the top spots for the four of us. There were endless courses of spectacular sashimi and cooked dishes, and the service was outstanding. As our host took the check, Kay got really quiet and began shaking his head. I sensed that something extraordinary had just taken place. As we left the restaurant, I said, “Kay, how much did that dinner cost?”
He thought for a second and said, “Six thousand dollars.”
“Six thousand for four people? How is that possible?”
“Best fish,” Kay said. “Big room.” Someone had gone to the immense Tokyo fish market and selected the top specimens. Quality and privacy came at a premium in Tokyo.
Before returning home, we took in Alien at a downtown movie theater. I’d seen it in Seattle, where I gasped like everyone else when the alien popped out of John Hurt’s chest. But in Tokyo, no one made a sound except Bill and me. Afterward I asked Kay if the audience had liked it. “Such a monster,” he said, shaking his head. “Such a terrible monster.” Kay looked nauseated—he’d been immersed in the film like everyone else, but they did not react. They held it all in.
In Japan we saw firsthand that our ambition to become the software language company had real potential. With China still closed and Korea not yet a player, to dominate Japan was to rule Asia. Back in Bellevue, the bullpen table grew cluttered with more 8-bit Japanese hardware. It was a preview of where personal computer design was headed—how a company like NEC, for example, was implementing color graphics that went miles beyond the Commodore PET’s.
The Japanese market was fiercely competitive. Late one night, we surprised a bespectacled engineer who’d sneaked into our office to snap Polaroid pictures of the competition. Another time, some Ricoh reps came by to ask what we had available. We ran down our list of every language on our shelf and one or two that weren’t ready yet. The reps kept nodding, and at the end they said, “We’ll take them all.” When their prototype machine malfunctioned and we failed to meet our delivery date, the head rep was distraught. “Mr. Allen, I promised to deliver,” he said, almost sobbing. He camped out at our office for days to help me get the software running. His honor was on the line.
Many of the Japanese machines were unconventional, with strange key placements, and a thought began to gnaw at me. Coming off my experience at MITS, I believed that we could build an 8-bit system superior to anyone else’s, including Apple’s, and customize it to run our software. Kay was pushing us to join forces with a Japanese company that would manufacture under the Microsoft name. He wanted to approach Sony, which was known for televisions and audio speakers but had no track record in computers. As Kay saw it, a Microsoft/Sony computer would be completely new and different, a true multimedia machine with state-of-the-art audio and video, the sort of thing I’d been talking about for years. Vern Raburn, the president of our consumer products division, was in favor. But Bill was adamant about staying out of hardware. “We’d be in conflict with our customers,” he told Kay. More than fifty companies were licensing our 8080 BASIC alone by that point, and the last thing Bill wanted was to turn those clients into competitors.
Our growing confidence made it easier to reject a mid-seven-figure purchase offer that summer from H. Ross Perot, the Dallas billionaire. It just felt way too soon for us to cash out. “Our conclusion is that at present we wish to remain independent,” Bill wrote to Mort Meyerson, Perot’s number two. “We see the potential to double the size of our organization and earn over $2 million per year before taxes.” Bill was on the mark: Microsoft’s year-end revenues would total $2.4 million in 1979, and our staff would more than double, to twenty-eight.
IN JUNE 1979, we made our first trip to New York City for the National Computer Conference. We took a two-bedroom suite at the top of the Plaza Hotel, the perfect spot for launching bottle rockets over Central Park. Kay Nishi came up with a request: He had friends in from Japan with no place to stay. Could they bunk in with us? Sure, we said—we didn’t want to be rude. A few minutes later, Kay showed up with half a dozen businessmen, all very polite, from Fujitsu, Toshiba, and NEC. I called the front desk and said, “How many rollaway beds do you have available?”
“I think we have six, sir.”
“OK, bring them all up.”
Soon there came a knock on the door. Six chuckling bellhops lined the corridor with six rollaways, a less than typical request for a high-priced suite. The beds filled the living room until you could hardly inch past them. The next morning, I had to fight my way through a forest of socks hanging in the bathroom, which the Japanese had left out to dry. But our hospitality paid off. One of our guests snapped open his briefcase, filled to the brim with U.S. currency. He was so eager to buy our BASIC interpreter that he’d brought cash for a down payment on the license, over ten thousand dollars. Bill wrote out a receipt on his business card.
The annual event was where suits from firms like IBM and DEC pitched their latest mainframes and minicomputers. Microcomputer companies were the new kids on the block, shunted to a small annex in a hotel by the main arena. Eddie Currie, Ed Roberts’s old number two at MITS, had moved to Lifeboat Associates, a software distribution company in New York, and he invited us to share his ten-foot-square exhibition space. We’d brought Tim Patterson along to help us debut our BASIC-86 on Tim’s prototype machine. No one else had a 16-bit BASIC, and ours would shortly be in the market. I was feeling pretty good until I stopped by the booth of a Massachusetts outfit called Personal Software. They had an Apple II running something I’d never seen before, on any class of computer: an interactive accounting spreadsheet. They called it VisiCalc.
Though the booth wasn’t drawing much of a crowd, it did grab the attention of an electronics analyst who later became the venture capitalist behind Compaq Computer. Ben Rosen understood that he was looking at the first “killer app,” an application that would dominate and redefine its category. As Rosen wrote the following month in the Morgan Stanley Electronics Letter:
Today, virtually the only user of personal computers who is satisfied with the state of the software art is the hobbyist. And he does all of his programming himself. But for the professional, the home computer user, the small businessman, and the educator, there is precious little software available that is practical, useful, universal, and reliable.
Enter VisiCalc … a new concept in software. … Though hard to describe in words, VisiCalc comes alive visually. In minutes, people who have never used a computer are writing and using programs. Although you are operating in plain English, the program is being executed in machine language. But as far as you’re concerned, the entire procedure is software transparent. You simply write on this so-called electronic blackboard what you would like it to do—and it does it.
Rosen described a dividend discount valuation model that had taken him twenty hours to program in BASIC; he created a more flexible version of the same thing with VisiCalc in fifteen minutes. “Who knows?” he concluded. “VisiCalc could some day become the software tail that wags (and sells) the personal computer dog.”
That was our philosophy, too; we believed that software was more valuable than hardware. But we hadn’t counted on someone outflanking us with a whole new approach. To that point, business programs had been written almost exclusively for higher-end microcomputers like Tandy’s TRS-80 Model II, machines marketed to small businesses that did their own data processing. Apple computers were viewed as toys for educational programs and games. But once VisiCalc enabled nonprogrammers to do financial modeling on the Apple II, all that was about to change.
At Microsoft we’d had good excuses for putting off a move into applications software. The field was competitive and highly fragmented, and Bill and I had decided that we wouldn’t enter a market unless we knew we could be number one. And with our programmers straining to fulfill our language contracts, it was hard to see how we could plunge into a whole new sector. Still, I’d had pangs as I watched WordMaster evolve into WordStar, the first widely accepted application of its kind. I knew in my gut that word processing would become a major revenue source. Were we missing the boat?
VisiCalc was another wake-up call. We’d licensed Applesoft BASIC on a fixed-fee basis, so we had nothing to gain from a spike in Apple II sales. Worse yet, our other languages ran exclusively on CP/M, which was incompatible with the microprocessor used by the Apple II: the MOS Technology 6502, Intel’s cut-rate competitor. With VisiCalc boosting its sales geometrically, Apple would be positioned to carve out a big slice of a growing market, one Microsoft couldn’t penetrate. And we didn’t need to read Ben Rosen to realize that people could use the new spreadsheet program without our software.
From the start, we’d built Microsoft around the premise that our products would be universal. Wherever the general-purpose microcomputer market went, we’d be there. But as personal computing matured from an enthusiast subculture into a mass medium, I came to see that languages would soon be outweighed by applications. Our mission could be at risk unless we built our own spreadsheet, and our own word processor and database, as well. The Altair had taught us how quickly fortunes in the tech world could rise and fall.
As Microsoft’s technical leader, I faced a more immediate bind: How could we get our existing products onto the Apple II platform? In theory, we could develop new compilers for the Apple in FORTRAN, COBOL, Pascal, and the rest. But the job would require years of coding by several programmers. It would leave us understaffed in our core business of porting BASIC to new 8080 machines, not to mention the 8086 computers just around the corner. Morever, the Apple work would saddle us with a new catalog of assembly code to debug and enhance, a costly, labor-intensive proposition. All told, the expense and distraction of full-scale development could cripple our still-small company. It was a dilemma that begged for an original solution.
A few months after seeing VisiCalc, heading to lunch in the back of Steve Wood’s pickup truck, I got one of those ideas that fortuitously flash into my head, a mix of inference and extrapolation. Instead of rewriting our entire software catalog, why not turn the Apple II into a compatible system? If we designed an 8080 circuit board to plug into the Apple, the machine could run CP/M from a floppy disk and all our languages on top of it. By importing a CP/M-friendly CPU, we’d avoid a massive recoding project and get into the Apple II market at least six months sooner.
In effect, I’d turned a software problem into a hardware problem—an elegant shortcut, a sort of Hail Mary pass. At first Bill wondered if it might be a distraction. But he came to agree that this was one Microsoft hardware effort that might be worth the trouble.
There are two phases to any invention. The first is the moment of inspiration. The second is the execution, which is less exciting but more than challenging in its own right. I had no idea whether my idea was actually doable. I called Tim Patterson and said, “Can you design this thing?”
And Tim said, “I think it’s possible.” A few weeks later, he came back with a circuit board containing a Z-80 chip, a cheaper equivalent to Intel’s 8080. It was simple enough to undo the plastic snaps and pop the lid off an Apple II, then slide the card into an expansion slot wired to the CPU. The native MOS Technology 6502 still ran the Apple’s peripherals (display, keyboard, printer) but otherwise went into a state of suspended animation. The Z-80 SoftCard, as we called it, took over most of the actual processing. We’d turned the Apple II into something that Steve Jobs wouldn’t have imagined: a CP/M computer.
Tim understood the Z-80 well, and the card’s general design was more than adequate, but getting two processors to coexist was a nightmare. The thing would work fine for a while, but then the native CPU would crash and take our SoftCard with it. In March 1980, we rolled out the prototype at the West Coast Computer Faire in San Francisco, fretting that it would go down at an inopportune moment. I can recall Steve Jobs passing by with a scowl. He had to be irritated that we’d barged into his Apple II walled garden and thrown the gate open to the whole CP/M software community.
To eradicate the SoftCard’s gremlins, I brought in an Apple-savvy engineer named Don Burtis and paid him $8,000 for a ground-up redesign. He quickly found the defect in the hardware’s architecture. On April 2, 1980, we issued a press release entitled “Cornucopia for Apple Computer Owners”:
A product that will allow the more than 75,000 Apple computer owners to use a vast array of new software, including business packages, was announced today by Microsoft Consumer Products. …
“Most of the existing 8080/Z80 programs require a five thousand dollar or more computer,” says Paul Allen, Microsoft Vice President and Z-80 SoftCard creator. “After hearing about the Z-80 SoftCard, several business people have told us they plan to bring home their word processing, accounting or statistical programs to run on their home Apple computers at night. That makes Apple computers tax deductible.”
We bundled the SoftCard with diskettes for CP/M and our BASIC interpreter, and priced it at $349. It started shipping that fall to strong demand. As we thought it might, VisiCalc helped to drive Apple’s sales through the roof; Jobs had nearly a year’s head start before the spreadsheet was developed for other microprocessors, and he exploited his lead well. The Apple II went from 35,000 units in 1979 to 210,000 in 1981, lagging only the Atari 400/800 and the TRS-80. It became a hit on college campuses and made a notable dent in the small business market.
My invention allowed Microsoft to share in that success. We sold approximately 25,000 SoftCards in 1981 alone, worth about $8 million in sales, and continued our strong run into 1983 before imitators cut into our margins. For Apple II owners who’d been limited to a thin catalog of native applications, the SoftCard gave them two computers in one. Suddenly they had access to tens of thousands of CP/M-compatible programs written in BASIC, FORTRAN, or COBOL. On the flip side, the SoftCard represented a huge windfall for Peachtree Software, creator of the popular Peachtree Accounting, which with no development costs had a new market handed to it. And of course, our new product was a boon to Gary Kildall and Digital Research. More copies of CP/M would be sold for use in the Apple II, a hitherto incompatible machine, than for any other computer.
For Microsoft, the SoftCard provided a point of entry into the Apple environment. It gave us a new and substantial customer base for our Disk BASIC and other languages. Moreover, the SoftCard turned computer-pricing strategy on its head. In the old world, everyone from IBM to MITS had bundled software as a throw-in with the machine. Now we were bundling a cheaply made piece of hardware to help us sell BASIC and our expensive suite of software. The SoftCard was the razor; our languages were the blades.
The SoftCard lent Microsoft a needed revenue boost in an awful recession year. Perhaps most important, it gave us comfort in abandoning the 8-bit development world and turning our energies to software for the 8086 chip, a shift that would prove critical in landing our big contract with IBM less than a year later. As Bill noted in a 1993 interview for the Smithsonian:
[The] question was, “Should we spread those products over to other 8-bit chips, like the 6502 that runs in [the Apple II]? Or, should we immediately move up and do 16-bit software?” And I said, “No, we are going to do 16-bit software.” Everybody was a little bit disappointed because it meant that we wouldn’t be able to sell onto these machines. That is when Paul invented the idea of the SoftCard, so that we could actually take our Intel software and run it on this machine, and, at the same time, go ahead and devote our resources to being way ahead of everybody else in developing software for the 8086.
I had already been instituting the move to 16-bit software, but Bill wasn’t wrong about the SoftCard’s importance. Under the circumstances, I felt that our 64–36 partnership split was out of whack. Bill had set a precedent by claiming extra equity for his work on Altair BASIC, another exceptional contribution. Now it was time, I thought, to augment my share. A modest adjustment in the ratio seemed only right.
But when I made my case, Bill would have none of it. “I don’t ever want to talk about this again,” he said. “Do not bring it up.”
In that moment, something died for me. I’d thought that our partnership was based on fairness, but now I saw that Bill’s self-interest overrode all other considerations. My partner was out to grab as much of the pie as possible and hold on to it, and that was something I could not accept. I didn’t have it out with Bill at the time. I sucked it up and thought, OK … but one day I’m out of here.
MICROSOFT NOW COMMANDED the CP/M 8-bit market in programming language software, and the SoftCard gave us a secure beachhead with Apple. But as we grew, our need for more help became glaring. Neither Bill nor I had a lot of experience as managers, and both of us had other areas of responsibility—Bill in sales, I in software development. Steve Wood had filled in admirably as general manager, but he, too, was a programmer by background. Bill came to see that we needed someone to help him run the business side of things, just as I ran technology. He chose Steve Ballmer, a Harvard classmate who’d worked in marketing at Procter & Gamble and was now studying at Stanford’s business school. Bill sold him hard to me: “Steve’s a supersmart guy, and he’s got loads of energy. He’ll help us build the business, and I really trust him.”
I’d run into Steve a few times at Harvard, where he and Bill were close. The first time we met face-to-face, I thought, This guy looks like an operative for the NKVD. He had piercing blue eyes and a genuine toughness. (Though as I got to know him better, I found a gentler side as well.) Steve was someone who wouldn’t back down easily, a necessity for working well with Bill. In April 1980, shortly before leaving town on a business trip, I agreed that we should offer him up to 5 percent of the company, because Bill felt certain that Steve wouldn’t leave Stanford unless he got equity.
A few days later, after returning from my trip, I got a copy of Bill’s letter to Steve. (Someone apparently found it in the office’s Datapoint word processing system, and it had made the rounds.) Programmers like Gordon Letwin were furious that Bill was giving a piece of the company to a person without a technical background. I was angry for another reason: Bill had offered Steve 8.75 percent of the company, considerably more than what I’d agreed to.
It was bad enough that Bill had chosen to disregard me on a partnership issue we’d specifically discussed. It was worse that he waited till I was away to send the letter. I wrote him to set out what I had learned, and concluded: “As a result of discovering these facts I am no longer interested in employing Mr. Ballmer, and I consider the above points a major breach of faith on your part.”
Bill knew that he’d been caught and couldn’t bluster his way out of it. Unable to meet my eyes, he said, “Look, we’ve got to have Steve. I’ll make up the extra points from my share.” I said OK, and that’s what he did.