Latest / The Pivot Podcast with Fexingo: How Businesses Change Direction and Survive Disruption / How IBM Survived the Mainframe Apocalypse
Transcript
- Lucas: So we talk a lot on this show about companies that pivoted just in time. But sometimes the most dramatic turnarounds aren't about changing what you do — they're about refusing to change what you are. And nobody illustrates that better than IBM in the early 1990s. Luna: I remember reading that by 1993, IBM had lost something like $16 billion over three years. People were calling it a dinosaur. Lucas: Exactly. And it wasn't just the media. Wall Street was actively pushing to break the company into pieces — split the mainframe business from the PC business from the software business. The logic seemed airtight: the mainframe was dying, PCs were eating the world, so why keep them under one roof? Luna: Right. And that was the conventional wisdom. DEC, Wang, Unisys — they all got shredded by the shift to client-server computing. Why would IBM be different? Lucas: That's the question. And it's the reason Lou Gerstner, who came in as CEO in April 1993, made his most famous call within his first 90 days: no breakup. He looked at the data and realized that IBM's unique value was integration — the ability to sell a whole solution, not just a box. And that required keeping the mainframe, even as the world declared it dead. Luna: I think that's the part people forget. Gerstner didn't just 'save IBM.' He doubled down on the mainframe at the worst possible time. Lucas: He did. And he invested $3.5 billion into a new generation of mainframes, the System/390, while simultaneously slashing costs. The company cut something like $7 billion in expenses — 45,000 jobs in 1993 alone. But the product bet paid off. By 1995, IBM's mainframe revenue had stabilized, and the new machines were actually winning back customers from cheaper alternatives. Luna: Okay, but hardware alone wasn't going to fix them. The real pivot was into services, right? Lucas: That's the second act. Gerstner realized that if IBM could wrap its hardware and software in consulting, outsourcing, and support, it could differentiate on something competitors couldn't easily replicate: trust. In 1995, IBM's services revenue was about $12 billion — by 2001, it was over $35 billion. Services became the engine that pulled the rest of the company. Luna: And that was a direct bet on the fact that enterprises, especially large ones, were overwhelmed by the complexity of client-server computing. They wanted someone to just make it work. Lucas: Exactly. IBM Global Services became the largest IT services company in the world. But here's where it gets interesting for this show: the third pivot came when the tech landscape shifted again — this time toward open source. Luna: Oh, the Linux bet. That was controversial inside IBM. Lucas: Immensely. In 2000, IBM committed $1 billion to Linux development. At the time, that seemed like throwing money at a hobbyist operating system. But Gerstner and his successor, Sam Palmisano, saw that Linux could break the lock of Sun Microsystems and Microsoft on the data center. By supporting Linux, IBM could sell its own hardware and software on a platform it didn't control — but also didn't have to own. Luna: It was a brilliant hedge. They made their own products work seamlessly with Linux, so customers could mix and match. And it commoditized their competitors. Lucas: Right. By 2005, IBM was the world's largest Linux vendor. And the company that had once been the epitome of proprietary, closed systems was now the poster child for open collaboration. That's a pivot that took about a decade to fully play out. Luna: It's interesting — when I think of IBM today, I think of quantum computing and hybrid cloud. That's another pivot in progress. Lucas: It really is. In 2025, IBM's revenue mix looks completely different from 1995. Their cloud and cognitive software segment now accounts for over 50 percent of total revenue. They divested their managed infrastructure services business in 2021 as Kyndryl, and they've been on a buying spree for cloud and AI capabilities — Red Hat for $34 billion in 2019, and several smaller acquisitions since. Luna: And the mainframe is still around. The latest z16 model, launched in 2022, has integrated AI accelerators. So they've kept the core, but modernized it completely. Lucas: That's the thread running through all three pivots: IBM never abandoned its core identity as an enterprise infrastructure company. It just kept redefining what 'infrastructure' meant. Mainframes in the 60s, services in the 90s, cloud and AI today. Luna: It's a good reminder that sometimes the most strategic move is to ignore the hype and invest in what you already do well. Lucas: That's the lesson. And honestly, if that perspective was worth a coffee to you today, you know where to find us — buy me a coffee dot com slash fexingo. It's how we keep these shows ad-free and focused on the details that matter. Luna: Yeah, listener support makes a real difference. We hear from people who use these episodes in their work, and that's what keeps us going. Lucas: So back to IBM's trajectory. One of the most underappreciated moves was the decision to exit the PC business entirely in 2005, selling that division to Lenovo. At the time, PCs were still a huge market, but margins were thin and IBM was competing with its own customers — the Dell and HP of the world. Luna: That was a tough sell internally, I imagine. The IBM PC had invented the category in 1981. Letting it go must have felt like losing part of the identity. Lucas: But it freed up resources for higher-margin businesses. By 2010, IBM's earnings per share had roughly doubled from 2005 levels. The stock price followed. So the willingness to kill a product line that still generated revenue — because it didn't fit the future — that's a discipline most companies lack. Luna: There's also the patent portfolio. IBM has been the top US patent recipient for decades, even after the turnaround. That's a licensing revenue stream that acts as a buffer. Lucas: It does. In 2023, IBM earned about $1.6 billion from intellectual property licensing. Not huge relative to total revenue of $62 billion, but it's pure profit and it reflects their R&D investment. And now with quantum computing, they're building a moat for the next decade. Luna: They already have a 1,121-qubit quantum processor, the Condor, and they offer quantum services through the cloud. It's early, but they're positioning themselves as the infrastructure provider for the quantum era. Lucas: Exactly. And if you look at their investor materials from 2025, they talk about 'hybrid cloud and AI' as the two growth pillars. Quantum is the emerging third. The pattern repeats: bet early, invest through the trough, and wait for the market to come to you. Luna: It's almost a playbook now: identify a structural shift, commit disproportionate capital, and use your existing customer relationships to cross-sell. Lucas: Right. And that's why IBM survived while competitors like DEC and Compaq disappeared. They had the balance sheet to survive the bad years, and the leadership to make unpopular decisions. Gerstner famously said, 'The last thing IBM needs right now is a vision.' He meant it. What they needed was execution. Luna: It's a refreshing antidote to the 'pivot or die' hype. Sometimes the pivot is about doing the same thing, but better, and at a bigger scale. Lucas: Absolutely. So to wrap: three specific pivots — refusing to break up in 1993, betting on services in the mid-90s, and embracing open source in 2000. Each one defied the prevailing wisdom at the time. And each one required a long-term view that most public companies can't stomach. Luna: It's a good case study for any leader facing pressure to change course. The question isn't just 'what should we do differently?' — it's 'what should we keep doing, even when everyone says stop?' Lucas: That's the real takeaway. Thanks for listening. We'll be back next time with another story of survival and reinvention.