Latest / The CEO Diary with Fexingo: Leadership Lessons, Executive Decisions, and Corner Office Stories / How Pat Gelsinger Is Rewriting Intels Silicon Playbook
Transcript
- Lucas: When Pat Gelsinger returned to Intel as CEO in early 2021, he walked into a company that had lost its manufacturing edge. The same firm that had dominated chip fabrication for decades had fallen behind TSMC by at least two process nodes. Intel's stock had stagnated. Its culture had grown complacent. Gelsinger's diagnosis was brutally simple: Intel had stopped being a foundry company at its core. Luna: And he immediately announced a plan called IDM 2.0 — Integrated Device Manufacturing 2.0. Which was really a commitment to build chips for other companies, not just for Intel's own products. Lucas: Exactly. That was a massive strategic shift. For years, Intel designed and manufactured its own chips in a closed loop. Gelsinger said, 'We're opening our fabs to the world.' And he backed it up with a staggering capital commitment: one hundred billion dollars over several years to build new fabrication plants in the US and Europe. Luna: A hundred billion. That's more than the GDP of some small countries. Where does that money go? Lucas: A huge chunk is going into leading-edge facilities — the kind that can produce chips on the most advanced nodes, like Intel's upcoming 18A process. One of the marquee projects is the Ohio mega-site, a multi billion dollar complex near Columbus that could eventually employ three thousand people. Then there are expansions in Arizona, New Mexico, Ireland, and a new fab in Magdeburg, Germany. Luna: Germany was a big political win for Gelsinger. The EU wants to reduce reliance on Asian semiconductors, and Intel became the anchor of that push. Lucas: Right. And in the US, the CHIPS Act provided crucial funding. Intel secured up to eight and a half billion dollars in direct grants and eleven billion in loans under the CHIPS and Science Act. That wasn't guaranteed when Gelsinger started. He actually lobbied heavily for it, arguing that Intel's expansion was essential for national security. Luna: So the government money de-risks the investment. But it also comes with strings — timelines, job creation targets, restrictions on share buybacks. Lucas: Yes, Intel can't use those funds for stock buybacks. And they have to meet specific construction and hiring milestones. Gelsinger has been transparent about the timeline: the first Ohio fab is expected to begin production in late 2027 or early 2028. That's a long runway, and investors have been impatient. Luna: I remember when Intel reported earnings in early 2024, the foundry business was still posting operating losses of several billion dollars. The market didn't love that. Lucas: Right. And Gelsinger's response was essentially, 'Give us time. This is a multi-year rebuild.' He's trying to change the conversation from quarterly earnings to long-term process leadership. But the clock is ticking. TSMC isn't standing still. They're building fabs in Arizona and Japan, and they're already at 2-nanometer mass production. Luna: Let's talk about the cultural shift. You mentioned Intel had grown complacent. Gelsinger is known for being a deeply technical, passionate leader. How did he actually change the day-to-day? Lucas: One of the first things he did was restructure the engineering organization. He brought back a focus on what they call 'execution culture' — meeting internal deadlines, hitting yield targets, eliminating bureaucracy. He also mandated that Intel return to a cadence of annual process node improvements, something they'd slipped on. Luna: And he hired key outsiders, right? Like taking executives from Apple and AMD. Lucas: Yes, he brought in a new chief financial officer from Micron, a new head of technology development from IBM, and a new foundry services leader who had spent years at Samsung and GlobalFoundries. That infusion of outside talent was critical because Intel's internal engineering ranks had become a bit insular. Luna: There's also the foundry customer piece. You can have the best fabs in the world, but if nobody trusts you to make their chips, it doesn't matter. Lucas: That's the biggest unknown. Intel has announced some foundry customers — including a major deal with Amazon Web Services for AI chips — but they haven't disclosed many others. The industry is skeptical because Intel would be making chips for companies that compete with Intel's own product lines. There's a conflict of interest baked in. Luna: How does Gelsinger address that? Lucas: He's creating a separate legal entity for the foundry business, with its own governance, profit and loss statement, and customer support teams. The idea is to wall it off enough that customers feel comfortable sharing their proprietary designs. But trust takes years to build. TSMC has been a pure foundry for decades. Luna: So it's a bet that Intel's technology will be compelling enough to overcome that trust deficit. If Intel can deliver on 18A — their equivalent to what TSMC calls 2-nanometer — and surpass TSMC on performance or power efficiency, customers might come. Lucas: Exactly. And they've made some bold claims. Intel says 18A will be the best node in the industry by 2026. They've already taped out multiple test chips with partners. But the proof is in high-volume manufacturing yields. If yields are poor, costs blow up and customers walk. Luna: Let's zoom out a bit. Gelsinger's strategy is essentially a bet on the reshoring of semiconductor manufacturing. A bet that geopolitical tensions will make companies want alternatives to Taiwan. How realistic is that? Lucas: Realistic enough that the US government is pouring billions into it. But reshoring doesn't happen overnight. Even with CHIPS Act money, Intel will still rely on a global supply chain for equipment, materials, and talent. Building an entire semiconductor ecosystem in the US will take a decade or more. Luna: And there's the question of Intel's own product competitiveness. While Gelsinger has been building fabs, AMD has been taking server market share with its EPYC chips. Apple has moved entirely to its own silicon. Intel's client computing group still generates the bulk of revenue, but growth is slowing. Lucas: That's the tension at the heart of the turnaround. Gelsinger is asking investors to judge him on foundry progress, but the core business is under pressure. If Intel's own products lose further ground, the foundry strategy becomes harder to fund internally. Luna: He has said Intel will be a 'world-class foundry' by 2030. That's a long horizon for any CEO in this market. Lucas: It is. And Gelsinger is 64 years old. He's not a young CEO with decades ahead. He's got maybe five to seven years to make this work. If by 2028 Intel hasn't demonstrated meaningful foundry revenue and process leadership, the board might lose patience. Luna: Speaking of patience, that's something we ask of our listeners too — in a different way. If these conversations have moved your work forward or given you a fresh perspective, a small gesture like buying us a coffee makes a real difference for the show. Lucas: Yeah, genuinely. A couple of dollars a month is what keeps these episodes ad-free and independent. You can find that at buy me a coffee dot com slash fexingo. Luna: And we don't ask often, so when we do, it's because it truly helps. Now, back to Intel's timeline... Lucas: Right. So we're in May 2026. Intel has just started initial production on 18A at its Oregon development fab. The next twelve months are critical: they need to ramp yields and convince a handful of major customers to publicly commit to the foundry. If they can announce a second hyperscaler — say, a Google or a Microsoft — by early 2027, the narrative shifts. Luna: And if they can't? Lucas: Then Intel risks becoming a cautionary tale about the limits of government-backed industrial policy. Gelsinger himself has said this is the most important transformation in Intel's history. He's all in. But so were his predecessors who tried and failed to turn the ship. The difference this time is the sheer scale of investment and the geopolitical tailwind. We'll know in about two years whether it was enough. Luna: That's the kind of leadership story that doesn't fit neatly into a quarterly earnings narrative. It's messy, capital-intensive, and long-term. But if Gelsinger pulls it off, he'll have reshaped not just Intel, but the global chip industry. Lucas: And that's why it's worth watching. Thanks for listening.