Pat Gelsinger was the chief executive of Intel from February 2021 through December 2024, when the Intel board removed him over execution issues with the 18A timeline, the foundry cash burn, and the broader cycle of optimistic guidance that kept missing milestones. Eighteen months after his departure, the foundry strategy he initiated has produced the largest single foundry contract of the modern era (Google's 3-million-unit TPU order on 18A) and Intel stock is up 196 percent year to date. The honest read on the Gelsinger era is that the strategy was largely correct and the execution was largely wrong, which is why the comeback narrative requires understanding both halves.
Here is the background, the major decisions, and the structural setup he left for the Tan-era execution.
Background and the 40-Year Intel Career
Pat Gelsinger joined Intel in 1979 at the age of 18 as a junior engineer, with no college degree at the time of hire. He completed his bachelor's in electrical engineering at Santa Clara University while working at Intel, followed by a master's at Stanford a few years later. The early career arc at Intel ran through the 486 processor program where he was the lead architect by his late twenties, then through senior engineering and operational leadership roles across the 1990s and early 2000s.
Gelsinger rose to become Intel's first chief technology officer in 2001 and then senior vice president, but in 2009 he left Intel after being passed over for the CEO role. He moved to EMC as president of the Information Infrastructure Products business, and in 2012 he became chief executive of VMware. The VMware tenure ran nine years and produced significant revenue growth from roughly $4.6 billion to over $13 billion per the VMware historical annual filings, with VMware becoming the dominant enterprise virtualization platform during the cloud transition window.
The Intel board recruited him back in early 2021 to run the company at a moment when Intel had been losing manufacturing leadership to TSMC for several years, per the contemporaneous coverage by Reuters of the Gelsinger CEO appointment. The brief he was given was effectively to fix the foundry execution problem at all costs.
The IDM 2.0 Strategy and the $100 Billion Commitment
The first major strategic decision Gelsinger made was to commit Intel to the Integrated Device Manufacturer 2.0 (IDM 2.0) strategy, with the framework first laid out in his March 2021 Intel webcast presentation archived at the Intel investor relations site. The thesis was that Intel needed to do three things in parallel. It needed to keep designing its own chips (the existing CPU business). It needed to use external foundries (TSMC for some products) where the in-house manufacturing was not competitive. And critically, it needed to build out the in-house foundry business to manufacture chips for third-party fabless customers.
The capital commitment behind IDM 2.0 was the largest in Intel's history. Gelsinger announced over $100 billion in committed capital expenditure to build four new fabs in Arizona, Ohio, Germany, and Ireland between 2021 and 2025, with the broader IDM 2.0 framework documented in the Intel IDM 2.0 announcement archive. The Arizona Fab 52 (the facility now producing the 18A node) was the centerpiece of that buildout. The funding mix combined Intel cash flow, the CHIPS Act federal subsidies, and private capital partnerships including a Brookfield infrastructure investment.
The IDM 2.0 strategy has been broadly validated by the Tan-era execution. The Google TPU order is a direct consequence of having Fab 52 operational on the 18A node, which is a direct consequence of the capital expenditure Gelsinger committed in 2021 and defended through three years of cash burn. The strategic call was right.
Where the Execution Went Wrong
The execution went wrong in three places. The first was on the timeline. Gelsinger's original 18A milestone schedule had high-volume manufacturing targeted for late 2024, which slipped to mid-2025 and then to late 2025. Each slip eroded the credibility of the foundry pivot with potential customers who needed confidence in the production schedule before committing volume.
The second was on guidance. Gelsinger's quarterly earnings call commentary across 2022 and 2023 consistently anchored to optimistic foundry customer pipeline framing that did not show up in actual contract announcements. The pattern of forward-looking statements that did not convert to disclosed contracts eroded analyst trust in the long-term thesis, which compressed the multiple at exactly the moment the company needed market support for the foundry capex.
The third was on the customer relationship layer. The foundry business is fundamentally a relationship-driven, multi-year sales motion that requires deep technical engagement with potential customers over extended timelines. Gelsinger's technical background was an asset for product development but did not translate to the customer rebuild that the foundry business specifically needed. The Google TPU deal closed under Tan because Tan personally led the negotiation.
By late 2024, the board concluded that the IDM 2.0 strategy was correct but the leadership model needed to change. Gelsinger's departure was announced in December 2024 and the Tan appointment followed in March 2025.
What He Got Right That Made the Comeback Possible
The honest assessment is that three Gelsinger-era decisions made the current Intel rerating possible. The first was keeping the 18A research and development funded at full intensity through the multi-year cash burn period, when the financial pressure from public investors and from internal stakeholders made cutting back the obvious move. The 18A process node is now the structural anchor of the Intel foundry business, and it would not exist in its current form without sustained R&D commitment from 2021 through 2024.
The second was keeping the Arizona Fab 52 construction on schedule despite repeated cost overruns and supply chain disruptions. Fab 52 is now the highest-volume 18A facility globally, and Google's TPU order specifically requires Fab 52 capacity at scale. The facility would have been delayed by 18 to 24 months under more conservative capital allocation.
The third was the leadership team development that gave Tan a deep bench of operational executives when he stepped in. The senior leadership across the foundry business unit, the technology development organization, and the customer engagement teams were largely Gelsinger-era hires. Tan did not have to rebuild the technical organization, only the customer-facing and capital allocation layers.
The structural setup Gelsinger left for Tan is the reason the rerating could happen on a 15-month timeline rather than a five-year timeline. The CNBC coverage of the leadership transition at the time of his departure framed the question of if the IDM 2.0 thesis would survive his exit, and the answer 15 months later is clearly yes.
What His Career Says About the Limits of Technologist-CEOs at Scale
The Gelsinger story is a useful case study on the limits of the technologist-CEO model at a company the size and complexity of modern Intel. Technical depth helps when product execution is the binding constraint. Customer relationship depth helps when sales execution is the binding constraint. Intel under Gelsinger needed both, and the technical depth was insufficient on its own.
The broader implication for technology investors is that the right CEO profile for a given company depends on the binding operational constraint at that moment. A pure technologist works when the company is winning on product but struggling on operations. A pure capital allocator works when the company is winning on operations but losing on cost. A customer-relationship operator works when the company has the product and the operations but is losing on go-to-market execution. Intel's foundry pivot was the third case, and the board diagnosed that correctly even though it took until late 2024 to act.
Gelsinger has since taken a board role at a private semiconductor equipment vendor and remains active in industry commentary, with broader Intel comeback context covered in the Phemex NVIDIA stock 2026 coverage. His post-Intel public statements have been gracious about the strategic continuity at the company, and he has explicitly credited Tan for executing the customer-relationship layer that was missing during his own tenure.
Frequently Asked Questions
Why did the Intel board ultimately ask Pat Gelsinger to leave?
The board concluded that the foundry pivot strategy was correct but that the execution was not delivering against the timeline or the customer commitments, and that the credibility damage from repeated optimistic guidance was compounding. The change was about leadership style, not strategic direction.
Is the current Intel rerating fair credit to Gelsinger or to Tan?
Both. Gelsinger funded and defended the 18A R&D, kept Fab 52 on schedule, and built out the operational bench. Tan rebuilt the customer relationship layer, closed the Google deal, and ran the capital allocation discipline that the prior era lacked. The rerating required both halves, which is why the comeback narrative needs both names.
What is Pat Gelsinger doing now?
He has taken a board role at a private semiconductor equipment vendor and remains active in industry commentary. His post-Intel statements have been measured and have explicitly credited the Tan-era execution layer.
Could a different Gelsinger-style executive have closed the Google deal?
Probably not on the same timeline. The Google negotiation specifically required executive-level relationship rebuild that Tan was uniquely positioned to lead given his three decades at Walden. A different technologist-CEO would likely have needed two to three years longer to develop the same customer relationship depth.
Bottom Line
Pat Gelsinger committed Intel to the IDM 2.0 strategy, defended the 18A R&D through three years of cash burn, and kept Arizona Fab 52 on schedule despite repeated execution pressure. Those three decisions are the structural reason the current Intel foundry rerating is possible at all. The execution gaps that led to his departure (timeline slips, optimistic guidance, customer relationship rebuild) were real, and the board correctly diagnosed that a different leadership profile was needed.
The Google TPU win on 18A is the validation that the strategic call was right and that the new execution layer can deliver against it. The honest read on the Intel comeback story is that it required both the Gelsinger-era capital commitments and the Tan-era customer rebuild. Neither half on its own would have produced the current rerating.
This article is for informational purposes only and does not constitute financial or investment advice. Cryptocurrency trading involves substantial risk. Always conduct your own research before making trading decisions.
