AI / LLMUS
← Sticker Wall
JOURNEY · DETAIL
Intel: The Turnaround Journey from Layoffs and Restructuring to Foundry Recovery Under Hock Tan
Founded: Robert Noyce, Gordon Moore, Andy Grove · Intel Corporation
JOURNEY
Key Fields
FIELD STAMPSIndustryAI / LLM
RegionUS
ScaleGiant
ChannelOther
Origin
Founded in 1968 by Noyce and Moore, Intel started in memory chips before pivoting to microprocessors, launching the world's first commercial microprocessor, the 4004, in 1971. In the 1990s, the Wintel alliance allowed it to dominate the PC market. However, starting in the 2010s, the company faced a series of missteps: refusing to supply chips for the iPhone, missing the mobile internet wave; abandoning early GPU investments, missing the AI compute surge; and suffering years of delays in the 10nm process, allowing TSMC to overtake it. By 2024, the company's market cap had shrunk significantly, and it had laid off tens of thousands of employees. In March 2025, the board brought in semiconductor veteran Hock Tan as CEO to attempt a restart of the 57-year-old company at its most critical juncture.
Milestones
1968
Founding and Pivot Turning Point
In 1968, Noyce and Moore left Fairchild Semiconductor to found Intel, initially focusing on memory chips, and launched the 4004 in 1971. In the 1980s, facing a price war from Japanese manufacturers, Grove and Moore made the famous decision in 1985 to abandon the memory business and pivot entirely to microprocessors. This 'burning the ships' moment was the company's first existential turning point, laying the foundation for three decades of CPU dominance. This phase lasted from 1968 to 1985.
1993
PC Dominance Growth
The Pentium processor launched in 1993, and combined with the 'Intel Inside' marketing campaign, Intel and Microsoft formed the Wintel alliance to rule the PC era. By around 2000, the company held approximately 80% of the PC processor market. Revenue climbed from $3.9 billion in 1990 to $33.7 billion in 2000, and in 2005, its market cap was the highest among global semiconductor firms. Andy Grove's 'Only the Paranoid Survive' management philosophy became a Silicon Valley textbook. This phase lasted from 1993 to 2005.
2006
Consecutive Missteps Failure
Around 2006, Intel refused to supply low-power chips for the Apple iPhone, citing low margins and volume, which caused it to miss the entire mobile internet era as the ARM architecture rose via smartphones. Simultaneously, the company abandoned early GPU and discrete graphics projects, failing to see the potential of parallel computing, which paved the way for NVIDIA to dominate AI compute later. Management later admitted these were two of their most painful strategic errors. This phase lasted from 2006 to 2016.
2016
Process Stagnation Failure
Intel's 10nm process, originally scheduled for mass production in 2016, was repeatedly delayed until 2019, forcing the 14nm process to be extended for years. Competitor AMD used TSMC's 7nm process to launch Ryzen processors and capture significant market share. In 2020, Apple announced it would drop Intel chips for its own M1, and by 2021, the data center market share was being eroded by AMD EPYC. The core competitive advantage of process leadership was lost. This phase lasted from 2016 to 2021.
2021
Foundry Gamble Inflection Point
In 2021, Pat Gelsinger returned as CEO, announcing the IDM 2.0 strategy and establishing the Intel Foundry Services division. He invested tens of billions of dollars over four years to build plants in Arizona and Ohio, aiming for five process nodes in four years. However, the foundry business suffered massive losses. After a disastrous Q2 2024 earnings report, the company announced 15,000 layoffs and suspended dividends. In December 2024, Gelsinger was forced into retirement, with the stock price down about 60% from its 2020 high, leaving the company in its worst crisis in 50 years. This phase lasted from 2021 to 2024.
2025
Hock Tan's Restructuring Turning Point
After taking over as CEO in March 2025, Hock Tan implemented sweeping reforms: cutting management layers, laying off tens of thousands, selling non-core assets, refocusing on engineering culture, and securing a 10% equity stake from the U.S. federal government and a $2 billion investment from SoftBank. Q1 2026 results showed revenue and EPS beating expectations, with strong growth in data center AI server CPUs and 18A process yields ahead of schedule. At Computex in June, he announced a three-pronged strategy of products, foundry, and custom chips. This phase lasted from 2025 to 2026.
Turning Points
- 1985: Abandoned the memory business to pivot to microprocessors, completing the first existential transformation.
- Refusal to supply iPhone chips, handing the entire mobile internet era to the ARM camp.
- 10nm process delays led to the loss of process hegemony, allowing TSMC and AMD to overtake Intel.
- 2021: Gelsinger launched the IDM 2.0 foundry gamble, but massive losses ultimately led to his downfall.
- 2025: Hock Tan took over, implementing layoffs, restructuring, and securing government investment, with 2026 financial reports providing the first evidence of recovery.
Failures & Pitfalls
- Refusing Apple's iPhone chip order around 2006, missing out on a decade of mobile computing growth.
- Abandoning early GPU parallel computing investments, allowing NVIDIA to monopolize the AI compute market via CUDA.
- 10nm process delays from 2016 to 2019, exhausting user and investor trust with 'toothpaste-tube' style incremental upgrades.
- Foundry business suffering multi-billion dollar quarterly losses, with the 'five nodes in four years' plan draining cash flow.
- Announcing 15,000 layoffs and suspending dividends in 2024, leading to Gelsinger's departure after the board lost confidence in the transformation progress.
关键成功要素
- Treating process leadership as the only moat; once delayed, the entire moat collapses.
- The opportunity cost of twice rejecting new platforms was far higher than the short-term profits saved.
- The core of Hock Tan's restructuring is cutting layers, pruning businesses, and concentrating resources on winnable battlefields.
- Aligning with national strategy (government equity, domestic manufacturing subsidies) to secure a critical window for transformation.
- Cultural reconstruction is harder than a roadmap; restoring an engineering-led environment is Hock Tan's top priority.
Lessons
- The biggest risk for an incumbent is not the competitor, but arrogance regarding the next platform; Intel twice used pricing logic to reject the future.
- The vertical integration model is a barrier when leading in process technology, but a double burden when falling behind.
- Foundry transformation requires long-term customer trust; you cannot buy TSMC's 30-year customer relationships just by building factories.
- There are no shortcuts for a giant's turnaround; Hock Tan's choice to admit problems first and then focus through layoffs actually regained market patience.
- A company left behind by the era can still secure a second chance through its server CPU base and government support.
Core Data
- 2000 Revenue：$33.7 billion (Public data, independent verification not performed)
- 2024 Layoff Scale：Approximately 15,000 employees (Public data, independent verification not performed)
- U.S. Government Equity Stake：Approximately 10% (Public data, independent verification not performed)
- SoftBank Investment Amount：$2 billion (Public data, independent verification not performed)
- 2021-2024 Stock Price Decline：Approximately 60% (from 2020 high) (Public data, independent verification not performed)
- Q1 2026 Financial Results：Revenue and EPS both exceeded market expectations (Company financial report)
Competitors / Peers
In the PC and server CPU markets, Intel faces AMD, which made a comeback using TSMC's process, with EPYC continuously eating into data center share. In the AI compute market, Intel has been completely crushed by NVIDIA's GPU and CUDA ecosystem, with a market cap gap that once exceeded 20x. In the foundry market, it faces TSMC's ~60% market share and pressure from Samsung Electronics, with Intel Foundry still lacking major flagship external clients. In mobile and edge computing, it has failed to shake the ARM architecture camp. Fighting on four fronts without an advantage in any is the most difficult structural dilemma of its transformation.