IBM's Pattern of Dismissal: The Decisions That Created Trillion-Dollar Competitors

The Inventor in the Kitchen

Chester Carlson was a patent attorney who hated copying documents by hand. In the 1930s, he began researching what he called "electrophotography" — a dry copying process that didn't require the messy chemicals of existing methods.

Carlson toiled away in crude makeshift labs, sometimes using his kitchen, other times a rented backroom. By the early 1940s, he had a working concept. He just needed funding to commercialize it.

Between 1939 and 1944, Carlson knocked on the doors of more than 20 companies. He approached every major office equipment manufacturer: IBM, General Electric, RCA, Kodak. The response was universal.

Not interested.

The IBM Meeting That Almost Happened

In March 1941, IBM's director of market research actually requested a demonstration of Carlson's invention. This could have been the moment that changed everything.

But Carlson faced a critical problem: he had no working model to show. Despite years of effort and substantial financial investment, his prototype was far from functional. He could explain the concept, but he couldn't produce usable copies.

IBM's persistent requests to see the model left Carlson frustrated and evasive. He knew his invention wasn't ready to demonstrate. The opportunity slipped away.

IBM's official position became clear: they felt that carbon paper was a cheaper alternative. Why invest in an unproven technology when the existing solution worked?

The Decision Logic

From IBM's perspective in 1941-1944, the rejection made sense:

• Carbon paper worked. It was cheap, familiar, and required no new equipment.
• Carlson couldn't demonstrate a functioning product.
• The market for "automatic copying" was unproven.
• IBM's core business was tabulating machines and punch cards, not office consumables.

What Actually Happened

In 1944, the Battelle Memorial Institute — a non-profit research foundation — took a chance on Carlson. In 1947, a small photographic paper company called Haloid licensed the technology.

Haloid became Haloid Xerox in 1958, then Xerox Corporation in 1961.

The Xerox 914 copier, introduced in 1959, became one of the most successful products in American business history. It revolutionized office work and made Xerox a dominant force in business technology.

IBM didn't enter the copier market until 30 years later, in 1970. When they did, Xerox immediately sued them for patent infringement.

The Lesson Ignored

This should have been IBM's wake-up call. A technology they dismissed as unnecessary — because the existing solution seemed "good enough" — created an entirely new industry and a powerful competitor.

But IBM didn't learn this lesson. They would repeat the same pattern, almost verbatim, with the next revolution.

The Birth of DEC

In 1957, Kenneth Olsen founded Digital Equipment Corporation (DEC) with $70,000 in venture capital. His vision was simple: computers didn't have to be room-sized mainframes that cost millions of dollars.

While IBM was selling mainframes to Fortune 500 companies, DEC began building smaller, cheaper machines for scientists, engineers, and researchers who couldn't afford — or didn't need — IBM's massive systems.

IBM's Dismissal

IBM so disregarded minicomputers that they wouldn't introduce one until 1976 — 19 years after DEC was founded.

The logic was familiar: minicomputers were "too small to do serious computing." IBM's customers were large enterprises that needed real processing power. Why would anyone want a smaller, less capable machine?

Meanwhile, between 1968 and 1972, ninety-two companies introduced minicomputers. A new industry was being born, and IBM wasn't in it.

The PDP-8: The Computer That Changed Everything

In 1965, DEC released the PDP-8 for just $18,000 — making it the least expensive computer on the market, dramatically undercutting IBM and other mainframe makers.

The PDP-8 found customers IBM never imagined: research labs, universities, small manufacturers, process control systems. It proved that "smaller" didn't mean "less useful" — it meant "accessible to more people."

Too Little, Too Late

When IBM finally entered the minicomputer market with the Series/1 in 1976, DEC had already established a near-monopoly. By the early 1980s, DEC was the second-largest computer company in the world, directly challenging IBM for market leadership.

The VAX minicomputer was cheap enough that a low-level executive could approve the expenditure. An IBM mainframe purchase required approval at the top executive level. IBM had ceded an entire market segment because they couldn't see value in machines that weren't mainframes.

The Irony

Here's where the story gets interesting: Ken Olsen, who built DEC by exploiting IBM's blind spot, had his own blind spot.

In 1977, at a World Future Society convention, Olsen reportedly said: "There is no reason for any individual to have a computer in their home."

(Olsen later clarified he was talking about centralized "home control" computers, not personal computers. But the quote captured a real failure of imagination that affected DEC's strategy.)

The company that rose by seeing what IBM missed... missed the next revolution themselves. DEC never successfully pivoted to personal computers and was acquired by Compaq in 1998 for $9.6 billion. Compaq was then acquired by HP.

The pattern of dismissal isn't unique to IBM. It's a disease that infects market leaders.

The Apple II Arrives

In 1977, two young men named Steve Jobs and Steve Wozniak introduced the Apple II — the first personal computer as we know it today. It had a keyboard, a display, and came fully assembled. Regular people could buy it and use it at home.

IBM's response echoed their minicomputer dismissal: personal computers were "too small to do serious computing" and therefore "unimportant to their business."

The Rejection Pattern

Before Apple became Apple, Jobs and Wozniak tried to sell their invention to established companies. HP rejected them. Commodore rejected them.

IBM never even considered personal computers a market worth entering. Their customers were corporations, not hobbyists.

The Wake-Up Call

By mid-1980, IBM executives had reason to be worried. Personal computers were no longer toys. VisiCalc, the first spreadsheet program, was turning Apple IIs into business tools. Customers were asking about personal computers. Some were buying them with or without IBM's blessing.

It became very clear that IBM needed to sell a personal computer. Their customers expected it. The question was whether IBM could move fast enough to matter.

The Secret Project in Boca Raton

In 1980, William Lowe, director of IBM's General Systems Division lab in Boca Raton, Florida, pitched IBM CEO Frank Cary on an audacious idea: build a personal computer in one year, sell it for $1,500.

Cary gave Lowe a month to develop a prototype — and just one year to get a product to market. For a company where major projects typically took three to five years, this was revolutionary.

Lowe was promoted shortly after, and development fell to Don Estridge — a 43-year-old engineer who would become known as the "father of the IBM PC." The project was code-named "Chess."

The Three Fateful Decisions

To meet the impossible deadline, Estridge's team defied IBM's rigid corporate culture. They made three decisions that would change the industry — and ultimately destroy IBM's position in the market they created:

• Decision 1: Off-the-shelf parts. Instead of designing custom components (IBM's usual approach), they used commercially available parts. The CPU was Intel's 8088 — a chip anyone could buy.
• Decision 2: Open architecture. They published the PC's technical specifications, inviting third-party developers to create software and accessories.
• Decision 3: Licensed operating system. Rather than developing software in-house, they licensed the operating system from Microsoft — a small company run by a 24-year-old named Bill Gates.

Why These Decisions Made Sense (In the Moment)

Each decision was rational given the constraints:

• Custom chips would take years. Intel's 8088 was ready now.
• Proprietary architecture would limit software availability. Open specs meant more developers would build for the platform.
• Building an operating system from scratch would blow the timeline. Microsoft's DOS was (almost) ready.

The Launch

On August 12, 1981, Don Estridge unveiled the IBM PC at New York's Waldorf Hotel. It was a massive success. IBM's brand legitimized personal computing for corporate customers who had dismissed Apple as a toy.

But the three decisions that enabled the rapid launch had already set IBM up for failure.

The Tragedy of Don Estridge

Don Estridge, the visionary who led Project Chess, died on August 2, 1985, in the crash of Delta Flight 191 near Dallas. He was 48.

By then, the clone makers were already eating IBM's lunch. The market Estridge created was slipping away from the company that created it.

The Phone Call

On July 21, 1980, Jack Sams of IBM called a young software entrepreneur in Seattle. "Don't get too excited," Sams told him, "and don't think anything big is about to happen."

The entrepreneur was Bill Gates. He was 24 years old. When Sams and his IBM colleagues arrived at Microsoft's office, Sams thought Gates was the office boy — he had the physique of a 12-year-old.

IBM needed an operating system for their new personal computer. Gates ran Microsoft, which was known for its version of the BASIC programming language. They didn't make operating systems.

The Gary Kildall Saga

Gates told IBM they should contact Gary Kildall at Digital Research. Kildall had created CP/M — the dominant operating system for microcomputers at the time. He was 13 years older than Gates, held a PhD in Computer Science, and ran an established company.

What happened next is one of computing's greatest "what if" stories — and the accounts diverge dramatically.

August 21, 1980: The Meeting That Changed Everything

On Thursday, August 21, 1980, the IBM delegation arrived at Digital Research's headquarters in Pacific Grove, California.

IBM's version: Gary Kildall wasn't there — he was out flying one of his private planes. His wife and business partner, Dorothy McEwen, refused to sign IBM's non-disclosure agreement. After hours of haggling, the frustrated IBM executives left without even starting discussions.

Kildall's version: He returned from a business trip that afternoon, signed the NDA, and met with IBM throughout the day. They reached an understanding to license CP/M. He even changed his flight to travel with the IBM representatives back to Florida. They had a handshake deal.

The Real Issue

Beyond the NDA drama, there were substantive business disagreements:

IBM wanted to pay a flat fee for CP/M rights. Kildall wanted royalties per copy sold.

Kildall had signed "favored nations" contracts with his existing customers. He couldn't give IBM better terms without offering the same deal to everyone else.

Most critically: CP/M-86 (the version that would run on Intel's 8086/8088 chips) couldn't be delivered in IBM's timeline.

Whatever the reason, IBM declared negotiations had failed. They returned to Gates.

Gates Makes His Move

Gates didn't have an operating system. But he knew someone who did.

Seattle Computer Products had developed an operating system called QDOS — "Quick and Dirty Operating System." It was written by Tim Paterson and was designed to be CP/M compatible.

With IBM's secret backing, Microsoft acquired QDOS. Paul Allen negotiated a licensing agreement: a flat fee of $10,000, plus $15,000 per company Microsoft licensed the OS to, with SCP granting non-exclusive rights. Seattle Computer Products didn't know who was really funding the deal — if they had, the price would have been far higher.

In July 1981, Microsoft bought QDOS outright for $50,000. Gates renamed it MS-DOS.

The Clause That Created Microsoft

On November 6, 1980, Microsoft and IBM signed the final contract. Gates agreed to sell DOS to IBM for a modest $50,000 — far less than what CP/M would have cost.

But here's what IBM missed: it was a non-exclusive royalty deal.

IBM's rights were non-exclusive. Gates kept ownership of the DOS program. Microsoft could sell MS-DOS to anyone they wanted — including every PC clone maker who would soon flood the market.

Steve Ballmer, a Microsoft employee, supported Gates during the negotiations. They both understood what IBM's lawyers apparently didn't: the operating system would be more valuable than the hardware.

Bill Gates' Role: More Nuanced Than the Myth

Interestingly, there's no controversy about one aspect: Gates initially tried to help Kildall. He sent the IBM team to Digital Research. When negotiations broke down, he called Kildall to urge him to restart talks.

It was only after Kildall couldn't reach terms with IBM that Gates — convinced by Paul Allen and Kay Nishi — decided Microsoft should offer an operating system to save the project.

Gates didn't steal the opportunity. He was handed it.

The Clone Invasion

The open architecture that enabled rapid development also enabled competition. In June 1982 — less than a year after the IBM PC launched — Columbia Data Products released the first functional IBM PC clone.

In 1983, Compaq released the Compaq Portable, advertised as the first "100% IBM compatible" computer. Unlike IBM, Compaq was a nimble startup that could move fast.

IBM had created an industry standard it didn't control. The clone makers could build equivalent hardware cheaper and often innovate faster than IBM's bureaucracy allowed.

Why IBM Did It Anyway

IBM's decisions made sense in the context of a 12-month deadline. Custom chips would take years. An exclusive Intel deal would require more negotiation time. Proprietary architecture would limit software availability.

But these were short-term optimizations that sacrificed long-term strategic position. IBM won the battle (shipping the PC on time) and lost the war (controlling the PC industry).

A Pattern Repeats

IBM's failure to fully embrace the Internet and the rise of mobile computing proved particularly costly. IBM was slow to recognize the transformative potential of the World Wide Web.

While IBM eventually embraced the internet — Steve Mills, head of IBM's software group, led discussions about how to "Webify" the company's enterprise tools, resulting in WebSphere in 1998 — they missed crucial early opportunities.

The Opportunities Lost

IBM missed the chance to establish dominance in:

• Search: Google was founded in 1998 by two Stanford students. IBM had massive computing resources and enterprise relationships. They never seriously pursued search.
• E-commerce: Amazon launched in 1994. IBM was selling to the very enterprises that would become e-commerce giants. They could have been the infrastructure. They weren't.
• Cloud services: This would become their most costly miss of all.

Gerstner's Vision

Lou Gerstner is widely credited with saving IBM. When he arrived as CEO in 1993, the company was losing billions and many expected it to be broken up. Instead, Gerstner kept IBM together and repositioned it around services.

When Gerstner left the company in 2002, he left with a clear vision: cloud and on-demand computing would define the future.

He was right.

What IBM's Leaders Actually Believed

But IBM leadership after Gerstner chose to protect existing revenue streams instead of building cloud infrastructure. Former IBMers describe the internal mindset:

• "AWS is losing money."
• "Enterprises won't trust Amazon."
• "We'll enter cloud when the market matures."
• "Our enterprise brand guarantees victory."

The Stock Buyback Strategy

From 2003 to 2013, IBM spent over $130 billion on stock buybacks.

Stock buybacks boost short-term earnings per share — which was directly tied to executive compensation. The strategy optimized for quarterly results while competitors built the future.

During the same period, AWS was learning through iteration, investing in massive infrastructure, building developer loyalty, and accepting early losses to shape a new computing paradigm.

The Softlayer Acquisition

IBM did recognize cloud computing eventually. In 2013, they acquired Softlayer — the #2 cloud provider behind Amazon.

But instead of investing aggressively in Softlayer's growth, IBM continued to prioritize buybacks. They had bought the horse but wouldn't let it run.

The Lost Decade

IBM entered what observers called a "lost decade" from roughly 2012 to 2020. While Amazon, Microsoft, and Google built dominant cloud platforms, IBM fell further behind.

Today, AWS generates more annual revenue than IBM's entire company. Microsoft's Azure has become the enterprise cloud leader. Google Cloud is rapidly growing.

IBM defended the past. AWS built the future. By the time IBM took cloud seriously, the market was already gone.

The Ambitious Promise

In 2012, IBM partnered with MD Anderson Cancer Center in Houston with an audacious mission: use Watson AI to help eradicate cancer.

The project was called the Oncology Expert Advisor (OEA). Watson would use natural language processing to summarize patients' electronic health records, search medical databases, and provide treatment recommendations to oncologists.

It was supposed to demonstrate that Watson could revolutionize healthcare.

The Reality

Five years and $62 million later, MD Anderson let its contract with IBM expire. Watson was never used on actual patients at MD Anderson.

What Went Wrong

The Core Problem

IBM failed in developing Watson for healthcare because:

• The scope was too big. Instead of achievable goals — like predicting cancer probability based on symptoms — they aimed to "eradicate cancer" with an AI module.
• They didn't keep it simple. There were easier ways for AI to address medical problems, but IBM preferred to develop something that sounded fascinating rather than building a practical solution.

The Pattern Continues

Watson for Oncology revealed a fundamental mismatch between IBM's marketing promises and the reality of AI capabilities. They over-promised and under-delivered — a pattern that damaged IBM's credibility in the AI space.

Today, OpenAI, Google, and others lead AI development. Watson, despite IBM's massive early investment, is not a major player.

The Acquisition

In 1995, IBM bought Lotus Development Corporation for $3.5 billion. It was IBM's largest acquisition at the time. Lotus Notes was seen as a powerful collaboration platform that could compete with Microsoft.

At its peak, Lotus Notes had an estimated 10 million applications running on the platform.

The Mismatch

Betsy Kosheff, who did PR for Lotus when it was sold to IBM, later observed: "IBM had no business doing software innovation. That point was very obvious right from the acquisition. It's not their fault — IBM is just not designed that way."

When Lotus was sold to IBM, they were in a head-to-head battle with Microsoft Exchange. But Microsoft had critical advantages: they owned the operating system and had majority market share with Office applications.

IBM intended Lotus Notes to sell more hardware. They never reached out to the experienced community of developers, deployers, and roll-out experts who had actually made Notes successful in corporations.

The Slow Decline

From those 10 million apps at peak, the first seven years under IBM saw a reduction to about 2 million — mostly through attrition and easy migrations to Microsoft SharePoint.

One unique "advantage" of Lotus Notes became its problem: servers and clients ran unattended for years. Customers simply stopped paying IBM for upgrades and support. Most also froze development of Notes apps, drying out associated revenue for IBM Global Services and IBM Partners.

The Sale

In December 2018, IBM sold Notes/Domino and several other software products to HCL Technologies for $1.8 billion.

IBM had spent $3.5 billion to acquire Lotus. Twenty-three years later, they sold the remnants for about half that amount — without adjusting for inflation.

The sale was partly driven by IBM's need for cash: they had just announced a $34 billion acquisition of Red Hat. But it marked the end of IBM's attempt to compete in collaboration software.

The Recurring Elements

Across 80 years and 12+ major decisions, the same patterns appear:

The Cognitive Biases

These patterns reflect well-documented cognitive biases:

• Status quo bias: Carbon paper works. Mainframes work. Our current business model works. Why change?
• Confirmation bias: Looking for evidence that new technologies won't matter while ignoring signs that they will.
• Innovator's dilemma: Successful companies can't disrupt themselves because doing so threatens existing revenue streams.
• Anchoring: IBM's identity as a mainframe company anchored their thinking even as computing evolved.

The Organizational Factors

Beyond individual biases, IBM's structure made these errors more likely:

• Bureaucratic inertia: Large organizations move slowly. The 12-month PC deadline was exceptional precisely because IBM normally took years.
• Siloed divisions: Divisions competed with each other rather than collaborating on new opportunities.
• Risk aversion: A culture that penalized failure more than it rewarded bold bets.
• Compensation structures: Executives rewarded for quarterly EPS (boosted by buybacks) rather than long-term market position.

Combined, Microsoft, Apple, Amazon, and Alphabet are worth over $10 trillion. IBM — the company whose decisions enabled or failed to prevent these competitors — is worth about 2% of that.