Microsoft Volume I: The Greatest Business Negotiation Ever

In one sentence: Two geeks in their early twenties used a contract that looked like a concession and was actually a lock to turn the most valuable company on earth into their own distribution channel. In 1980 Microsoft sold IBM the rights to DOS for a one-time $430,000, no royalties — and in exchange kept the right to relicense DOS to anyone it wanted. IBM was running its usual playbook (we show up, the market is ours) and never noticed it was handing away pricing power over the entire PC ecosystem. Within thirteen years Microsoft passed IBM in market cap (January 1993); it took until 2015 to pass IBM in revenue. Today it's the most valuable company in the world. Ben calls it "the greatest deal in at least computer industry history, if not all business history, full stop"; David puts it even more simply — "the single best business deal negotiation of all time." This episode is the Microsoft Volume I that Acquired spent ten years working up the nerve to tackle (1975–1995, garage to Windows 95), recorded in April 2024 — a time capsule with no antitrust hearings, no internet wave, no Satya Nadella. All of that is saved for Volume II.

The Company on One Page

Founder Profile

Bill Gates

Born into power and expectation. His father, Bill Gates Sr., served in World War II, went to the University of Washington law school on the GI Bill, and became a founding partner of Preston, Gates & Ellis — today's K&L Gates, one of the largest law firms in the world. His mother, Mary Maxwell Gates, never worked a full-time corporate job in her life and still sat on the boards of First Interstate Bank, Pacific Northwest Bell (part of the AT&T breakup), the University of Washington Board of Regents, and the national United Way — David's verdict: "she became one of the most powerful business people in the Pacific Northwest, period." Bill Sr. later joined Costco's board. At nine and ten, Bill sat at a dinner table that regularly hosted CEOs, senators, and governors — David compares him to Paul Atreides in Dune, bred from birth to be a business mind. At 13, the question he and his best friend discussed wasn't whether he'd be a CEO, but which industry he'd run.

Competitiveness was baked in. A childhood friend recalled: "Everything Bill did, he did competitively and never simply to relax" — a sharp contrast with the "shy computer nerd" image that stuck to him later. Paul Allen described him this way: "You could tell three things about Bill Gates pretty quickly. He was really smart, he was really competitive, and he wanted to show you how smart he was. And he was really, really persistent."

At Lakeside School, seventh grade, 13-year-old Bill got his first hands-on computing experience through a teletype the Mothers Club had crowdfunded — wired by phone line to a GE-owned DEC PDP-10 downtown. This was 1968; the microprocessor didn't exist yet, Silicon Valley was still Lockheed's territory, and "computer" meant either a room-sized machine or a person (think of the Black women "computers" in Hidden Figures). Bill threw himself into programming and quickly became one of the best in the group, co-founding the "Lakeside Programmers Group" with the two-years-older Paul Allen.

He was a businessman from the start, not just a geek. Debugging for local timesharing startup C-Cubed, he and his friends were paid in computer time, not cash — picking up Fortran, Lisp, and PDP machine code along the way. Later, writing payroll billing software for a Portland timesharing firm, he negotiated (with his lawyer father's help) a revenue royalty instead of hourly pay — earning at least $10,000, more than the median US household income of the day.

The 1971 conversation about Moore's Law is the origin moment of the entire Microsoft story. Paul had observed that semiconductors were improving exponentially; Bill's reaction:

"Paul and I had talked about the microprocessor, and it was really his insight that because of semiconductor improvements, things would just keep getting better. I said to him, oh, exponential phenomena are pretty rare, pretty dramatic. Are you serious about this? Because this means, in effect, that we can think of computing as free. It was a gross exaggeration, but it was probably the easiest way to understand what it means to cut costs like that. And Paul was quite convinced of it."

This conversation planted the seed of the famous vision — "a computer on every desk and in every home" (the second half, "running Microsoft software," would quietly get dropped once the DOJ started sniffing around). In spring 1974, after Intel announced the 8080, Bill (who'd entered Harvard in fall 1973 intending to be a world-class mathematician, got a B in Math 55, realized he'd never be the smartest guy in the room, and switched to applied math) read the news in Electronics Magazine: "All at once we were looking at the heart of a real computer, and the price was under $200. We attacked the manual. I told Paul, DEC can't sell any more PDP-8s now." That same year at Harvard he met the guy down the hall, Steve Ballmer.

When Micro-Soft was founded as a partnership in April 1975, Bill took 60% (later 64%) and Paul 40% (later 36%) — Bill's reasoning was blunt: "you took a job and were doing this on the side, I was all in." The "I'm all in, so I deserve more" logic would resurface five years later when Steve Ballmer joined.

What really cements Bill Gates's reputation as a businessman is his unconventional pricing instinct. Selling BASIC to Apple for just $31,000 (eight years of access) wasn't a failure to extract more money — it was a calculated bet: price Microsoft BASIC cheaply enough that "nobody would take the risk of a clone that might have one or two things wrong with it," and it becomes the de facto standard. Once that happens, "all of our competitors will just wither away." He called this "success reinforces success" — the same underlying logic he'd run again, on a much bigger stage, in the IBM negotiation five years later.

In 1980 he moved the company from Albuquerque back to his hometown of Seattle — not nostalgia, but three cold-blooded business reasons (escaping Silicon Valley's rumor mill, avoiding its job-hopping culture, and tapping the steady engineering pipeline of the University of Washington). That same year he engineered two consequential moves: persuading Steve Ballmer to drop out of Stanford Business School and join on a handshake deal for roughly 8.5% equity, and — after IBM's team walked away from a botched meeting with Gary Kildall — deciding to buy QDOS and sign the contract David Rosenthal would later call "the greatest business negotiation of all time."

By the 1986 IPO, after dilution from Ballmer, from VC firm TVI, and from the option pool, Bill still held 49% of the company — nearly unthinkable in today's venture-backed era. This wasn't a company where the founder happened to still hold the CEO title; this was Bill's company. In a 1998 conversation with Warren Buffett at the University of Washington, he made a striking admission: "I think the multiples of technology stocks should be quite a bit lower than the multiples of stocks like Coke and Gillette because we are subject to complete changes in the rules. I know very well that in the next 10 years, if Microsoft is still a leader, we will have had to weather at least three crises." He said this as the world's richest man, right as antitrust scrutiny was starting to close in.

Paul Allen

If Bill was the company's will, Paul was its earliest technical vision. When they met at Lakeside, Paul was already a tenth-grader — two years older than 13-year-old Bill. That gap nearly vanishes in the rest of the story, but at the start, he was the one who knew more.

In the pivotal 1971 conversation, it was Paul, not Bill, who first raised the idea that semiconductors would keep improving exponentially — Bill's own memory of the moment was disbelief ("exponential phenomena are pretty rare"), while Paul's read was closer to plain observation: this thing was already happening and looked like it would keep happening. He turned that instinct into concrete engineering moves twice. During Traf-O-Data, with the Intel 8008 chip not yet available, he wrote a full instruction-set emulator on a PDP-10 at Washington State, working purely off the manual. He repeated the trick for the Altair: Bill wrote the BASIC interpreter at Harvard entirely against an 8080 emulator Paul had built — neither of them had ever touched the actual chip. David draws the line live in the episode: this is exactly what NVIDIA's Jensen Huang would do decades later with new-generation GPUs — build against a simulator before the silicon exists, then ship blind.

The Albuquerque trip was Paul's alone to carry: on the flight he hand-wrote a bootloader in octal (not assembly — raw octal machine instructions) because they'd forgotten one was needed. Landing, he hooked it up to MITS's prototype Altair — the first run failed, they changed nothing, tried again, and the machine spat out "4" (they'd typed "print 2+2"). Neither he nor Ed Roberts could quite believe it worked. Paul then joined MITS as VP of Software — a department of one — while Bill stayed at Harvard, joining him in Albuquerque only once the school year ended.

When Micro-Soft was founded, Paul took 40% (later 36%) without objection — Bill's framing was direct: you still have a job, this is your side project; I'm all in. That temperamental deference (rather than a fight over the math) shaped his early role at the company: he was more drawn to hardware possibilities, and was often pulled back by Bill's insistence that "we are a software company" — one of the few visible fault lines between the two founders in this account.

In 1982 or 1983, Paul was diagnosed with Hodgkin's lymphoma. He took a leave, then left full-time work entirely, remaining only an intermittent board member from then on — the sharpest turn in this episode's narrative, and the point where the "two guys building it together" story effectively ends within the 1975–1995 window this episode covers. Even so, at the 1986 IPO he still held 28% of the company, the second-largest stake after Bill's. He never returned to day-to-day operations, and the episode — a time capsule that stops around 1995 — doesn't mention him again.

Steve Ballmer (the unofficial third architect)

Ballmer wasn't a co-founder, but nearly every former employee interviewed for this episode credits him with half of everything. In 1980 he went from Bill's hallway neighbor at Harvard to Microsoft employee — at the time enrolled at Stanford's business school, talked into dropping out by Bill. Bill offered roughly 8.5% equity (a number even the hosts can't pin down exactly — "8.5%, 8.75%, something like that"), a grant too large for a typical early hire and closer to a founder-level stake — Ben's read: "This is really a reflection that the way that Bill thought about Steve was as a founder."

Ballmer's own phrase, "riding the bear," captures Microsoft's mid-1980s position vis-à-vis IBM precisely: "You just had to try to stay on the bear's back and the bear would twist, turn, and try to throw you off. But we were going to stay on the bear because the bear was the biggest, the most important. You just had to be with the bear. Otherwise, you would be under the bear." Coming from a future CEO, it's a sign of how deeply he was embedded in strategy from the earliest days — not just "the sales guy."

What truly defines his historical role is the enterprise market. Once the Microsoft-IBM "divorce" became official at the end of 1990, it was Ballmer alone who took on the job of getting PCs into the Fortune 500 — at a company that, at the time, had no idea how to sell software to businesses. A direct sales force, channel-partner networks, independent-software-vendor relationships, customer service, the executive briefing center on the Redmond campus — all the enterprise-software infrastructure that seems obvious today, he built from nothing. In 1988 he also led a key talent raid, poaching IBM's director of software strategy, Mike Maples, to run Microsoft's applications business — a move that itself signaled the coming split between the two companies.

The Playbook

Each entry: origin story → insight → effect.

1. Capital efficiency isn't a virtue — it's control itself

• Story: the partnership-era cap table was Bill 60%/Paul 40% (later 64/36); Ballmer joined in 1980 for roughly 8.5%; the 1981 C-corp conversion brought in VC firm TVI for $1M and 5% ($20M post-money, against $17M in revenue that year — an absurdly cheap price by today's standards); by the 1986 IPO, after three rounds of dilution, Bill still held 49%, Paul 28%, Ballmer 7.5%.
• Insight: Ben's read — early software had a brief window where zero marginal cost plus a trivially low fixed cost to start (essentially just two people's time) made it possible, for the first time, to build the business without raising money. That window "would never be true again" — today even the most speculative new market has a minimum viable fixed cost in the billions.
• Effect: Bill was the true owner of this company, not a founder-CEO with a large but diluted stake. That's why every subsequent all-in strategic pivot (betting on Windows, walking into a fight with IBM and staying in it) could happen so decisively — it was his company, and he owed the board no justification.

2. The IBM deal: a concession on the surface, a lock underneath

• Story: the November 1980 contract had IBM pay Microsoft a one-time $430,000 ($75K testing/consulting + $45K DOS license + $310K for a batch of 16-bit language interpreters/compilers) — no royalties, no matter how many machines IBM sold. On its face, Microsoft had sold its most valuable asset cheap.
• Insight: the real leverage was buried in the other half of the deal — Microsoft kept the right to relicense DOS to "anyone, on any terms." IBM's assumption (as in every market it had entered before) was that once it showed up, the market was theirs; it never anticipated an army of clone makers, Compaq leading the charge, replicating the entire IBM PC out of off-the-shelf parts.
• Effect: Ben's summary cuts to the point — "Microsoft used IBM to generate demand for their software, and then they used every other PC manufacturer to capture the value that all that demand created." Once the clone wave took off, Microsoft collected a per-machine royalty on every one, and revenue rocketed from $25M in 1982 to $98M in FY1984 — nearly 4x in 18 months. David's verdict: "I think this might be the single best business deal negotiation of all time."

3. The MITS lesson: you don't fix piracy with enforcement, you fix it with deal structure

• Story: the 1975 exclusive license with MITS ($30/copy, capped at $180,000 total) nearly sank the company — customers "pirated" BASIC en masse (whether software was even copyrightable was legally unsettled at the time; it took a 1980 amendment to Title 17 of the US Copyright Act to define a "computer program" as protectable work), and MITS itself, facing a conflict of interest (sublicensing meant arming its own competitors), dragged its feet on getting BASIC into other companies' hands. 1975 revenue was just $16K; 1976 was $22K — less than the pair had made in high school.
• Insight: Bill's key realization — BASIC shouldn't be something the customer buys separately; it should be baked into the price of the hardware itself. Once it's bundled into the purchase, piracy disappears, because there's no longer a separate buying decision to route around.
• Effect: this lesson directly shaped the structure of the IBM contract — DOS was designed from the start to ship preinstalled and folded into the machine's price, not left as an optional purchase for the end user to decide on.

4. Don't maximize value on the first deal — maximize on becoming the standard

• Story: once the MITS arbitration was won, Microsoft licensed BASIC to nearly every hardware maker it could find — Apple got eight years of access for just $31,000 (Apple's own BASIC, written by Woz, was 95% done but never got floating point; Jobs pushed him to finish it, Woz refused, and Apple went and licensed Microsoft's BASIC instead — the first-ever deal between the two companies).
• Insight: Bill wasn't maximizing per-deal profit; he wanted it to be unthinkable for anyone selling a microcomputer not to include Microsoft BASIC. He called this "success reinforces success" — once a de facto standard forms, user and developer path dependence crushes any later entrant automatically.
• Effect: Microsoft BASIC became the industry's de facto lingua franca by the late 1970s, laying the trust foundation for the IBM partnership that followed — there was exactly one company in the industry doing this, and it had already proven itself.

5. Always aim at the next platform, never the current one

• Story: Microsoft's own spreadsheet, Multiplan, chased broad compatibility across as many machines as possible; upstart Lotus made what looked at the time like a crazy call — build 1-2-3 exclusively for the IBM PC, and make it the best. 1-2-3 became the most successful software of its era, at times outearning and outvaluing Microsoft itself (including the year Microsoft went public). Multiplan was left in the dust.
• Insight: the lesson Microsoft's applications team drew from the postmortem — never leave yourself exposed to the next generation of technology without a plan. The target is always the next platform, not whichever one currently dominates, even if you don't own it yet. That rule directly produced Excel: rather than slug it out with 1-2-3 on the command-line battlefield, bet that the GUI was the next platform and build the first — and best — graphical spreadsheet for it.
• Effect: Excel for Mac (1985) became history's first graphical spreadsheet, and its first-mover advantage crushed 1-2-3 once the GUI transition completed — though that payoff didn't fully arrive until Windows 3.0 (1990), five patient years later.

6. Hedge your bets, don't bet on conviction

• Story: from 1983 to 1990, Microsoft ran three parallel tracks: publicly declaring "OS/2 is the future" and building it in earnest with IBM (the company's most prestigious project at the time); quietly maintaining a sub-30-person, unglamorous Windows team on the side; and simultaneously going deep with Apple on Mac applications. The three tracks pulled against each other hard enough to split the company's culture in two — "applications" and "systems" might as well have been different companies.
• Insight: Ben's distillation — this wasn't a lack of conviction, it was the opposite. They had total conviction that PC software would win, and deliberately stayed agnostic about which specific path would win. David adds: doing this requires actually owning the company — you don't have to justify every step to a board as consistent with a "stated strategy."
• Effect: when OS/2 flopped on release in 1987, Windows 3.0 (1990) — riding on an installed base of 386/486 hardware that had finally become common enough — took off, doubling Windows's penetration within six months. Because the Windows team had existed for five years and accumulated real experience, the company could pivot its entire strategic weight almost without losing time.

7. Copying isn't shameful — but it only wins alongside a platform shift

• Story: Microsoft's applications were almost never first to market — spreadsheets, word processors, all copied products others had already validated (Steve Jobs's famous line: "Microsoft has no taste"). But David adds a crucial nuance: copying alone doesn't win — Multiplan copying VisiCalc never beat 1-2-3; only once the GUI platform shift arrived did Excel actually pull ahead.
• Insight: copy, then wait for a platform shift, is a risk-adjusted-optimal strategy — you skip the cost of educating an unproven market and the churn risk of early adopters, and only need to out-execute during the transition window. What this takes isn't inspiration; it's patience and cash reserves to survive until the window opens.
• Effect: "Microsoft's first and second versions of something usually aren't great, but the third version usually is" became an accepted internal law — they treated this iteration cadence itself as a strategic asset, not a flaw.

8. Software is never finished

• Story: Microsoft never had a "ship it and we're done" mentality — DOS, Windows, and Office were all continuously patched and extended even between major releases. This was fundamentally different from the mindset of hardware companies of the era (ship the box, install the software, the job's done).
• Insight: David's read — this follows necessarily from the software business model itself. You can always ship another floppy disk, another network update, at nearly zero marginal cost; "finished" is a concept that only makes sense in the world of physical products. This line of thinking runs straight through to today's default of continuous cloud shipping (and stands in contrast to Apple's still-annual release cadence, versus Microsoft's early embrace of "always shipping").
• Effect: continuous iteration became a competitive advantage in its own right — the fact that the third version is always better than the first only pays off inside an organizational culture that assumes there will be a fourth and a fifth.

9. OEM distribution: be Visa, not Amex

• Story: Ben's analogy — Apple designed, built, and controlled everything itself, like Amex, and was structurally capped in scale; Microsoft made only software and handed hardware manufacturing and distribution entirely to OEMs like Compaq, HP, Dell, and Gateway, like Visa handing card issuance and customer acquisition to banks. The team managing OEM relationships for Windows stayed around 20 people for years.
• Insight: an open network's growth rate isn't bound by how many people you can hire — it's bound by how many independently-scaling nodes you can activate. Every OEM expanded Windows's installed base on its own timeline, with its own capital, through its own channels; Microsoft did nothing, and the scale economics compounded automatically.
• Effect: this approach let Microsoft drive the largest hardware ecosystem in the world with a tiny team — PC unit shipments grew at a 98% compound annual rate from 1975 to 1986 (from 4,000 units a year to 9 million), and Microsoft rode nearly the entire growth curve without ever taking on the capital expenditure or inventory risk of making hardware itself.

10. Go international early, and turn "patch work" into an organizational capability

• Story: in 1978, a five-person company barely established in its own home market signed an exclusive Japan distribution deal over a single phone call with enthusiast Kay Nishi; by 1979, half of the company's revenue came from Japan. Every product release from then on forced the team to build multi-language localization in parallel.
• Insight: treating internationalization as a capability to build (rather than "something to think about once we're bigger at home") forces product architecture and team process to adapt to globalization early — string extraction, multi-language layout, cross-timezone coordination. Once these muscles exist, every future product inherits them for free, while competitors have to build them from scratch.
• Effect: Microsoft's international revenue share stayed near half for decades; when Windows 95 launched simultaneously in eight languages worldwide, it turned an OS release into a global cultural event (lighting up the CN Tower and the Tower of London) — the payoff of over a decade of accumulated organizational capability, not something improvised at the last minute.

11. A talent magnet: an extreme culture only earns extreme commitment at the right moment in history

• Story: Brad Silverberg, who led Windows 3.1 and later Windows 95, recalled the secret: "We laid out principles for product and then pushed responsibility down." No stack of specs, no three rounds of cross-functional review for every feature — understand the principles, and use your own judgment. He adds that everyone felt personally responsible for the product, and it showed.
• Insight: the team that shipped Windows 95 was only 360 people — an unthinkably small group by today's standards, building the operating system that defined an era. Veterans recall that everything else in life got neglected during that stretch, and nobody regretted it. An organization can only convert that level of commitment into product quality if everyone genuinely believes they're part of something historic.
• Effect: this cultural gravity made Microsoft the top destination in the 1980s–90s for ambitious people on every axis — technical, sales, marketing. Whatever kind of ambition you had, this was where you wanted to be.

12. The first principle of a platform: let the ecosystem earn more than you do

• Story: a Bill Gates line repeated inside Microsoft to this day — you want the ecosystem around you generating more revenue than you're taking for yourself. Compaq, Lotus, Intuit, and later Netscape all succeeded on top of the Microsoft platform — a role IBM never actually played for anyone.
• Insight: the fundamental difference between a product company and a platform company is that a product company captures all the value it creates, while a platform company deliberately hands most of the value it enables to the ecosystem, taking rent only at the point of access. If that rent-collection point is chosen well (DOS, for every microcomputer), the value given away eventually comes back at a far larger scale.
• Effect: in this episode's Power table, Cornered Resource is judged the strongest of the seven, and it comes down to one word: "DOS, full stop." Not because DOS was technically superior — it was born from Tim Patterson's one-man "quick and dirty" project — but because IBM's distribution turned it into an unavoidable point of access for the entire industry, leaving Microsoft to simply defend the position and collect the rent.

Moat Analysis (the 7 Powers framework)

This is a rare case of Acquired running through all seven powers at speed rather than digging into each one — the hosts say outright that this might be one of the most defensible business models in history and that all seven powers plausibly apply, so each gets "about 45 seconds," with the full Bear/Bull case and pantheon ranking explicitly deferred to Volume II (antitrust, the internet wave, and the enterprise era). The table below is this episode's compressed verdict.

Bear Case

This episode's Bear Case is deliberately left thin — the hosts say plainly, "we know what happened after this," handing the full verdict to Volume II. What can be pulled from this episode as foreshadowing:

• Antitrust: the quiet dropping of the second half of the "computer on every desk" vision ("running Microsoft software") once the DOJ started sniffing around, and the 1988–94 per-processor licensing practice (tying DOS/Windows royalties to every processor shipped, regardless of whether Microsoft's OS was installed) would later become a key piece of antitrust evidence — planted here purely as foreshadowing, with the details saved for the next episode.
• The fragility of a single-OS bet: the Bull case is "the party continues" — Microsoft keeps shipping great operating systems and stays the most important company in the world. The Bear case is "something else becomes the important thing" — betting an entire company on a deeply entrenched operating system isn't, in the long run, a strategy for the future; it's a way of extending the last war's victory.
• The internet: entirely absent from this episode except as repeated foreshadowing — "the Internet Tidal Wave memo," "we'll talk about the internet next episode" — the single biggest piece of the puzzle saved for Volume II.
• IBM's mirror lesson: the episode repeatedly warns that passing someone in market cap doesn't mean you've replaced their business — Microsoft passed IBM in market cap in January 1993, but it took until 2015 for Microsoft to pass IBM in revenue. Enterprise customer migration moves far slower than market sentiment, a lesson that will eventually apply to Microsoft itself.

Bull Case

By the point where this episode ends (1995), the structural advantages are all in place and still compounding: installed-base scale economics, the DOS-to-Windows relicensing network, an enterprise growth curve Ballmer had only just cracked open, and an organizational habit of always betting on the next platform. David's verdict — "I think this might be the single best business deal negotiation of all time" — is itself this episode's compressed Bull Case: the compounding effect of the contract that built this moat was, as of 1995, only just beginning to pay off.

Deep Cuts (Microsoft itself)

• What the Lakeside Mothers Club actually bought wasn't a computer — it was computer time: in 1968 the school's parents club pooled money for a teletype wired by phone line to a GE-owned DEC PDP-10 downtown — 13-year-old Bill's first hands-on brush with computing, at a moment when "computer" meant either a room-sized machine or a person (think of the Black women "computers" in Hidden Figures).
• C-Cubed and Steve Russell: the Lakeside Programmers Group traded bug-hunting for computer time at timesharing startup C-Cubed, picking up Fortran, Lisp, and PDP machine language along the way. One of the executives who mentored them, Steve Russell, was the same man who wrote Spacewar on the first PDP-1 at MIT — history's first video game, and the direct inspiration for Nolan Bushnell's Atari.
• The emulator trick, run twice: during Traf-O-Data, with the Intel 8008 chip not yet shipped, Paul wrote a complete instruction-set emulator on a PDP-10 straight from the manual; the Altair project repeated the move, with Bill writing the BASIC interpreter at Harvard entirely against an 8080 emulator Paul built — neither of them had ever touched a real chip. David draws the parallel live: this is exactly what NVIDIA's Jensen Huang would do decades later with new-generation GPUs — simulate the software before the silicon exists, then ship blind.
• MITS was a near-bankrupt model-rocket company: the Altair 8800's maker, MITS, was in the model-rocketry business; founder Ed Roberts got into it while stationed with the Air Force in Albuquerque. The Altair launch was a last-ditch save — a Popular Electronics cover secured through a personal connection, and a $397 price point (about $2,300 in 2024 dollars) made possible by a sweetheart deal Ed personally negotiated with Intel (getting 8080 chips at $75 apiece against Intel's $360 list price, roughly a fifth). Four thousand pre-orders in the first month or two brought in $1.5 million in cash — pulling the company back from the brink.
• Hand-written octal on the plane: flying to Albuquerque for the demo, Paul realized they'd never written a bootloader for the BASIC interpreter, and hand-coded one in raw octal (not assembly — the machine instructions themselves) on the flight. The first demo failed; they changed nothing, tried again, and it worked — the machine printed "4" for "print 2+2," and neither he nor Ed Roberts could quite believe it.
• The BASIC Woz almost finished: Apple's own BASIC, written by Steve Wozniak, was about 95% done, missing only floating-point math. Jobs pushed him to finish it; Woz refused; Apple went and licensed Microsoft's BASIC instead ($31,000 for eight years) — the first-ever deal between the two companies.
• Gary Kildall's NDA snafu: IBM's secret team originally wanted an operating system from Gary Kildall's Digital Research; the day they flew to California to negotiate, Gary himself was out flying his own plane (or on a business trip, accounts differ), and his wife Dorothy refused to sign IBM's NDA. IBM left believing the company couldn't deliver and turned to Microsoft instead — a candidate for the biggest business blunder in history, and the accident that rerouted the entire trajectory of PC operating systems.
• The origin of QDOS/86-DOS: at Seattle Computer Products, a local motherboard maker, programmer Tim Patterson wrote his own "Quick and Dirty Operating System" (QDOS) to test the company's new 16-bit board. Microsoft first paid $25,000 for the right to adapt and sell it to one anonymous OEM, then paid another $50,000 for full, unlimited relicensing rights — $75,000 total. Tim himself later joined Microsoft and turned QDOS into real DOS.
• Compaq's clean-room reverse engineering: three former Texas Instruments engineers founded Compaq and reverse-engineered the IBM PC's BIOS using a clean-room process — one engineer documented the original BIOS's interface behavior without writing implementation code, and a second engineer, seeing only that spec, independently reimplemented it. Compaq did $111M in revenue its first year and became the world's largest PC maker by 1994.
• The other half of the Xerox PARC story: nearly every foundational invention of modern computing — the graphical interface, the mouse, the desktop metaphor, object-oriented programming, Ethernet, laser printing — came out of Xerox PARC, whose 1973 prototype, the Alto, was really a minicomputer in Mac's clothing costing tens of thousands of dollars each. The popular version of the story has Steve Jobs touring PARC and stealing the idea for the Mac; but that same year (1980), Microsoft also hired Charles Simonyi away from PARC, bringing the same body of knowledge straight into the teams building Word and MultiPlan — David's verdict: "Apple was not the only person who dined lavishly at the feast."
• The real leverage buried in the contract: the DOS contract, a one-time $430,000 ($75K testing/consulting + $45K DOS + $310K language compilers), zero royalties, with Microsoft keeping the right to relicense DOS to anyone — David: "I think this might be the single best business deal negotiation of all time."
• "Come be IBM's next CEO" (unverified): per the unauthorized biography Hard Drive, a month after Microsoft passed IBM in market cap (January 1993), IBM's board was in disarray and fired its CEO; board member Tom Murphy (of Capital Cities media fame) reportedly flew out to Redmond to personally ask Bill Gates to become IBM's next CEO. Bill declined. The hosts flag this source explicitly as unverified — "spiritually, you could believe that happened."
• Dave Marquardt's football-stadium pitch: TVI's Dave Marquardt spent a full year getting screened by Steve Ballmer before finally getting a meeting with Bill — at a University of Washington vs. Arizona football game, where Bill ignored the game entirely and grilled Dave on his understanding of the company's strategy. Against Don Valentine's famous rule of never investing outside bicycling distance from Sand Hill Road, Dave flew to Seattle most weekends — his own explanation: "I was young, I was single, I had nothing better to do, and it was really fun and intellectually interesting."
• Windows 95, by the numbers: 1 million copies sold in the first week, 7 million in the first month; 75 million Windows users already on 3.1 before the launch; a simultaneous eight-language worldwide launch that lit up the CN Tower and the Tower of London; Coca-Cola's own marketing team reached out to Microsoft to ask how it had marketed the launch so well.

Era & Industry Trivia (tangents worth keeping)

• IBM used to be the entire definition of "tech company": the hosts quote Ben Thompson's essay at length: "Fifty years ago, what is a tech company was an easy question to answer. IBM was the tech company, and everybody else was IBM's customers." IBM owned hardware (System 360), software, and services all at once, until antitrust pressure forced it to unbundle its pricing in 1968 — cracking a door that Microsoft would walk through more than a decade later.
• DEC and the classic low-end disruption of the "minicomputer": Ken Olson's DEC never competed with IBM's mainframes head-on; instead it built minicomputers — smaller, weaker, and an order of magnitude cheaper (the PDP-1 launched at $120,000, against mainframes costing tens of millions) — a textbook case of low-end disruption, and Ken Olson himself was young Bill Gates's hero.
• BASIC was the Python of its day: high-level, forgiving, readable even by non-programmers, yet powerful enough to underpin everything from education to real business applications — the hosts' comparison: "you can say it's not elegant, but it really was the lingua franca of the era."
• Software piracy is a legal concept that had to be invented: when Bill wrote his famous "Open Letter to Hobbyists" in 1975–76 accusing users of piracy, whether software was even copyrightable was legally undecided in the US — it took a 1980 amendment to Title 17 of the Copyright Act to define a "computer program" as protectable creative work. Without that change, the entire business model of software licensing (rather than bundled with hardware) doesn't work.
• The "1977 Trinity": Byte Magazine's name for the TRS-80 (Tandy/RadioShack), Commodore PET, and Apple II, all launched the same year as fully assembled machines — a huge leap from the Altair's kit form — the year the microcomputer market broke out of the hobbyist niche.
• Lotus once outearned Microsoft: 1-2-3 gave Lotus higher revenue and market value than Microsoft for several years in the late 1980s, including the year Microsoft went public — a fact the hosts admit they had no idea about before researching this episode.
• Jensen Huang is only seven years younger than Bill Gates: yet NVIDIA (founded 1993) and Microsoft (founded 1975) feel nearly a generation apart — a reminder that Bill Gates himself isn't as old as the institution Microsoft has become.
• IBM was the most valuable company in the world in 1980: at a $34 billion market cap, more than any oil company — Ben guessed $150 billion on the spot, off by more than an order of magnitude, a useful gut-check on how much the scale of "most valuable company" has changed in a few decades.
• The mirror of "nobody got fired for buying IBM": the old enterprise-procurement adage got flipped by David into "don't get fired for buying Microsoft" — the exact moment the brand baton passed was also the moment the power baton passed.
• Microsoft Research was born out of running out of low-hanging fruit: when Nathan Myhrvold founded Microsoft Research in 1991, the logic was direct — everything Microsoft had done for over a decade was porting tasks already proven on mainframes and minicomputers onto the personal computer. Once that job was essentially done, someone needed to be dedicated to inventing what came next.

Cross-domain Notes

This episode has no strong intersection with the PH (geopolitics) discourse network, and none is forced here — the 1975–1995 history of PC software is a story that happens almost entirely inside the United States, driven by patent and copyright law and corporate contracts, with no energy, currency-system, or great-power dimension for the PH domain's national-level concerns to attach to.

The one thing worth flagging is methodological rather than narrative: the "linchpin/point-of-integration" power logic this episode surfaces — whoever controls the node in an ecosystem that nobody can route around holds the pricing and bargaining power, DOS to the IBM PC ecosystem being the case in point — is the same underlying mechanism that recurs in the TSMC and NVIDIA episodes' "chokepoint" analysis, just operating on software instead of physical manufacturing. If the PH domain, or this domain's own semiconductor/platform thread, keeps expanding, that "who controls the ecosystem's point of access" framework could serve as a cross-episode or cross-domain connector — without needing to be force-fit onto a specific geopolitical narrative here.

Pages Worth Creating

• Entities: bill-gates (founder page: born into a power family, Lakeside to Harvard, the 1971 Moore's Law epiphany, the IBM negotiation, 49% ownership at IPO), paul-allen (co-founder page: the emulator trick run twice, hand-writing the Albuquerque bootloader in octal, fading from the story after his 1982–83 Hodgkin's diagnosis)
• Episodes: acquired-microsoft-2 (Volume II: every fuse lit in this episode — antitrust, the internet wave, the enterprise era, the bloodline from Windows NT to Azure — finally goes off)
• Concepts: 7 Powers 护城河框架 (Hamilton Helmer's framework; this episode is a rare case of running through all seven at speed rather than digging deep, with Cornered Resource landing as the strongest call), Counter-Positioning（反向定位） (existing page; Microsoft's counter-position against IBM — "we don't need to make money on hardware" — sits alongside GEICO and Walmart as a textbook case), 学习曲线（Learning Curve） (existing page; Bill Gates's "success reinforces success" flywheel shares the same underlying intuition as learning-curve pricing)

Source · acquired