Latest / The CEO Diary with Fexingo: Leadership Lessons, Executive Decisions, and Corner Office Stories / How Intel Built Its Silicon Foundry Empire
Transcript
- Lucas: Let’s talk about the invisible infrastructure of modern computing. Specifically, Intel’s decision to open its fabs to the rest of the world. For decades, Intel was vertically integrated. They designed their own chips, built them in their own factories, and sold them under their own brand. It was a closed loop that worked brilliantly until it didn’t. Luna: Right, because when TSMC started pulling ahead on process nodes, that vertical advantage became a liability. So Intel flipped the script and said, essentially, we’ll build your chips too. Lucas: Exactly. This is the IFS, or Intel Foundry Services division. The logic is sound on paper: if you have excess capacity and leading-edge process technology, why not monetize it by making Apple, Qualcomm, or even AMD use your lines? It turns a cost center into a profit center. Luna: But does the market actually want to rely on a former competitor who might also be designing competing products next quarter? Lucas: That is the central tension. In 2026, with geopolitical tensions affecting supply chains, many companies are looking for non-Chinese foundry options. Intel offers scale and location stability in Arizona and Ohio that TSMC cannot easily replicate. But trust is hard to build when you’re playing both sides of the table. Luna: So they’re betting that the need for diversified supply chains outweighs the fear of giving tech to a rival. Lucas: Precisely. And the numbers support the ambition. Intel has committed over one hundred billion dollars in capital expenditure over the last few years to expand these facilities. That’s not just maintenance; that’s an aggressive expansion into advanced packaging and logic manufacturing. Luna: One hundred billion is a staggering amount of cash. How do they justify that kind of spend when their core CPU business is facing margin pressure? Lucas: It’s a classic moat-building exercise. Process leadership is expensive. You need the best engineers, the most expensive lithography equipment, and years of R&D. If Intel stops investing, it falls behind again. If it keeps investing and wins the foundry war, it becomes the toll booth for all advanced electronics. Luna: Which sounds less like a chip company and more like a utility provider. Is that a good thing for valuation multiples? Lucas: Utilities get stable, predictable cash flows. Foundries can offer similar stability if you secure long-term contracts with major clients. But the risk is higher. One bad process node, one yield issue, and clients walk. It’s high stakes. Luna: And what about the internal conflict? Intel still designs its own processors. Won’t those clients worry about IP leakage? Lucas: They absolutely will. That’s why Intel has built strict firewalls within its organization. The design team doesn’t see the foundry client’s data, and vice versa. It’s modeled after how banks handle wealth management versus investment banking. Segregation is key. Luna: Sounds like corporate governance as much as engineering. Does that bureaucracy slow down innovation? Lucas: It can. Large organizations are naturally slow. But Intel is trying to change that by treating IFS as a separate entity internally. They’re hiring salespeople who speak customer language, not just engineer-speak. It’s a cultural shift as much as a business model shift. Luna: Changing culture at Intel must be like turning an aircraft carrier. Where do you even start? Lucas: You start with incentives. Previously, engineers were rewarded for delivering Intel’s internal roadmap. Now, they’re measured on customer satisfaction and yield rates for external clients. Different metrics drive different behaviors. Luna: So they’re literally paying people to care about someone else’s product. That’s a brave move. Lucas: It’s necessary. The era of monopoly profits is over. Intel has to earn every dollar from external customers now. And that requires a level of service quality that legacy semiconductor firms rarely offered before. Luna: I’m curious about the geographic aspect. Why are so many new fabs being built in the US right now? Lucas: Government incentives. The CHIPS Act provided billions in subsidies for domestic manufacturing. This isn’t just about Intel; it’s about national security. The US wants to ensure it has the capacity to produce critical semiconductors without relying solely on Asia. Luna: So political policy is driving industrial strategy. That changes the risk profile significantly. Lucas: It lowers financial risk but increases regulatory scrutiny. Intel is now a strategic asset in a way it wasn’t thirty years ago. That comes with expectations from Washington that don’t exist for private companies in other sectors. Luna: Do you think this model works for everyone? Or is it only viable for giants with deep pockets? Lucas: Only giants. The barrier to entry is too high. A smaller player can’t compete on price or volume. But for Intel, it’s the only path forward. They have the scale to absorb the downturns and the technology to attract top-tier clients. Luna: What’s the biggest hurdle they face in the next two years? Lucas: Execution. Building the fabs is one thing. Running them efficiently while managing internal product launches is another. Any slip in timing could alienate potential partners. Trust takes years to build and seconds to break. Luna: And if they miss, what happens? Do they just go back to being a pure design house? Lucas: Unlikely. Once you invest that kind of capital, you’re locked in. The sunk costs are too high. They have to make IFS work, regardless of the friction. It’s a point of no return. Luna: It reminds me of how Netflix shifted from DVDs to streaming. A complete overhaul of their core identity. Lucas: Similar energy. Both companies had to cannibalize their old models to survive. Intel is risking its reputation for design leadership to gain dominance in manufacturing. It’s a bold gamble. Luna: Are there any early signs that clients are actually buying in? Lucas: Yes. Several mid-sized tech firms and automotive chipmakers have signed multi-year agreements. They aren’t all going to be household names, but they provide steady revenue streams. It’s building block by building block. Luna: Steady revenue beats volatile spikes. That makes sense for a manufacturing business. Lucas: Absolutely. Manufacturing thrives on utilization. Empty fabs are money burning holes in the ground. Full fabs are printing cash. Intel is chasing full utilization. Luna: This feels like a lesson in resilience. Adapting your strengths to meet market reality. Lucas: It is. Intel realized its greatest strength wasn’t just its chips, but its ability to make them. By leveraging that capability externally, they’re creating a new pillar of growth. Luna: Makes you wonder if other legacy tech giants will follow suit. Maybe Adobe or Cisco? Lucas: Adobe doesn’t have fabs. But companies with hardware divisions, like HP or Dell, might look at this model. Outsourcing manufacturing used to be a sign of weakness. Now, it might be a sign of sophistication. Luna: Or maybe just survival. The margins in pure hardware are thinning out. Lucas: Thin, but sticky. People always need chips. The question is who controls the bottleneck. Intel is trying to become the owner of that bottleneck. Luna: Control over supply is power. Especially when global logistics are so fragile. Lucas: Exactly. In a world of disruptions, owning the factory is the ultimate insurance policy. Luna: Before we wrap up, I want to mention something important. Our conversations here are free, but the time and expertise that go into producing them aren’t. If you’ve found value in how Lucas and I break down these business shifts, consider supporting us directly. Lucas: We keep this network ad-free because we believe you deserve clean, focused analysis without interruption. Your support helps us continue digging into cases like Intel’s foundry pivot without compromise. If you’d like to help, visit buy me a coffee dot com slash fexingo. Luna: Every contribution goes straight back into research and production. Thank you for keeping this show alive. Lucas: Back to Intel. As we look ahead, the real test will be whether they can deliver leading-edge nodes consistently for external clients. Can they match TSMC’s reliability while maintaining their own design pipeline? Luna: If they can pull that off, they won’t just survive—they’ll redefine the industry's structure. Lucas: And if they fail, they remain a giant stuck in the middle. High risk, high reward. That’s the nature of transformation. Luna: It certainly is. What do you think is the one metric investors should watch most closely over the next quarter? Lucas: Foundry revenue growth rate. Not total revenue, but specifically the IFS segment. If that number is accelerating, it means the market is trusting Intel again. Luna: Trust, measured in dollars. That’s a powerful indicator. Lucas: It is. And it’s a reminder that in business, reputation is the hardest currency to earn and the easiest to lose.