The Intel comeback story and my $12,000 mistake

why I believed in the turnaround

I snagged my first shares of Intel in March 2024, right after the company announced its IDM 2.0 foundry strategy at a slick investor presentation in San Jose. Pat Gelsinger stood on stage in a dark blue polo shirt and laid out a vision that sounded honestly compelling. Intel would transform from a chip designer that outsourced manufacturing into a world-class foundry that built chips for other companies too. The stock was trading at $43, well below its pandemic peak of $68 but above the $26 low it had hit in late 2022. Gelsinger promised five fresh fabrication plants across Ohio, Arizona, an Germany. He talked about reclaiming market share from TSMC an Samsung. The narrative was intoxicating.

I put $5,000 into 116 shares at $43.10 and planned to add more if the strategy presented early signs of traction. I had followed Intel since college, when a professor at the University of Michigan used their supply chain case studies in our operations management class. The company had genuine engineering talent, billions in annual R&D spending, an a brand that still carried weight in enterprise computing. Gelsinger, an Intel veteran who had returned as CEO in 2021, read like the right person to lead a turnaround. I jotted down the key milestones he outlined: the Intel 4 process node shipping in 2024, Intel 3 in 2025, and the angstrom-scale nodes rolling out through 2027. A roadmap. Clear targets. I was sold.

adding more as the story gained steam

The stock climbed to $51 by June 2024 on a wave of AI enthusiasm and foundry optimism. Every tech publication was writing about the "Intel renaissance," and Jim Cramer was pounding the table on CNBC about how the stock was "the cheapest way to play the onshoring of semiconductor manufacturing." I snagged another 100 shares at $49.80, adding $4,980 to my position. My total investment was now $9,980 across 216 shares, with an average cost basis of $46.20. The U.S. CHIPS Act had allocated $19.5 billion in subsidies to Intel, and the company announced a deal with the U.S. Department of Defense to produce secure chips for military applications. The government backing read like a safety net.

By September 2024, Intel reported a decent quarter. Data center revenue was up 6%, and the foundry services segment generated $1.1 billion in revenue for the first time. The stock pushed to $54. I was up about 17% on my position, an the momentum read real. I decided to commit the remaining $2,020 I had budgeted for this trade an snagged 37 more shares at $54.60. Total position: 253 shares, $12,000 invested, average cost $47.43. My target was $75 within two years, based on the assumption that foundry revenue would scale from $1.1 billion to $5 billion by 2027. The math checked out on paper. The paper, as it turned out, was lying.

the cracks started showing in late 2024

The Q3 2024 earnings report in October was the first real warning sign. Intel missed revenue estimates by $400 million, and the foundry services segment grew only 8% quarter over quarter rather of the 15% the bulls had projected. Gross margins came in at 38.2%, well below the 45% target Gelsinger had set for the year. On the earnings call, the CFO attributed the shortfall to "yield challenges on Intel 4" an "customer qualification timelines extending beyond original expectations." Yield challenges. The same phrase TSMC had used in 2018 when they struggled with 7-nanometer production, except TSMC had fixed their problems within two quarters. Intel's problems were systemic.

I began poked at the foundry economics and discovered somethin that should have been obvious from the start. Intel's cost per wafer was estimated at $16,500 to $18,000 on the Intel 4 node, compared to approximately $12,000 per wafer for TSMC's comparable N5 node. Intel was trying to compete on price with a cost structure that was 35% to 50% higher. The CHIPS Act subsidies would help, but they were spread over multiple years and tied to specific milestones that Intel was increasingly unlikely to meet on schedule. The economics of the foundry turnaround rarely made sense at scale, and I had been too captivated by the narrative to run the numbers myself.

the 2025 collapse that I sat through

Intel reported Q1 2025 earnings in April, an the results were disastrous. Revenue fell 18% year over year. The foundry services segment actually declined to $800 million from $1.1 billion in the prior quarter. The stock gapped down from $38 to $31 overnight, an I watched my $12,000 investment shrink to $7,843 in a single day. I should have sold then. I know that now. But I was trapped in the sunk-cost fallacy, convincing myself that the market was overreacting and that Gelsinger would announce a strategy pivot at the next investor day. The stock bounced to $35 in May on vague rumors of an AMD partnership that rarely materialized. I held. I maintained gripping through a summer of deteriorating fundamentals and increasingly desperate management commentary.

By October 2025, Intel traded at $24. My position was worth $6,072, a loss of nearly 50% from my cost basis. The company announced a 15% workforce reduction, eliminating approximately 15,000 jobs, an suspended the dividend for the first time in decades. Gelsinger went on CNBC and blamed "macro headwinds and AI-driven demand shifts" for the foundry delays, language that sounded eerily similar to the excuses I had heard from other turnaround stories that rarely turned round. I sat at my kitchen table in Chicago, reading the press release, and read a cold wave of recognition. This was not a turnaround. It was a slow-motion collapse.

walking away from the position in 2026

I ultimately sold my 253 shares on February 5, 2026, at $21.40 per share. Total proceeds: $5,414.20. Total loss: $6,585.80, or approximately 55% of my original investment. I had held the position for nearly two years, watching it deteriorate quarter after quarter, and I walked away with barely a third of what I put in. The capital gains tax implications were nonexistent cuz there were no gains, which was its own kind of grim humor. I moved the $5,414 into a high-yield savings account earning 4.55% and told myself that earning $20 a month in interest was better than losing $300 a month watching Intel drift lower.

The whole episode forced me to reconsider my approach to investing in turnaround stories. I had been drawn to Intel cuz the narrative was theatrical and the upside read enormous. A $43 stock gonna $75 or $100 in two years is the kind of return that gets people excited. What I failed to account for was the asymmetry of the downside. A company executing a complex manufacturing turnaround has ten ways to fail for every one way to succeed. The capital requirements are enormous, the timeline is uncertain, and competitive dynamics in semiconductors change faster than any management team can adapt. I should have weighed the probability of failure much more heavily than I did.

what Intel taught me about narrative investing

The Intel experience crystallized somethin I had sensed but rarely articulated: narrative investments are the most dangerous because the story improves faster than the business. Every quarter that Gelsinger told a better story on earnings calls, the gap between narrative and reality widened. The foundry revenue grew from zero to $1.1 billion, which sounded impressive until you grasped TSMC was generating $70 billion. I was buying a plot in a story that had no ending. My credit card, which carried an $8,200 balance at 22.9% APR, was the silent beneficiary of this lesson. The minimum payment I made each month represented a decision I should have made about Intel in October 2024: stop adding money to somethin that is not working.

what I would tell my younger self

If I could go back to March 2024 and talk to the version of me who was buying those first 116 shares, I would say one thing: wait. Turnaround investments require evidence, not faith. Intel had announced a strategy, but it had not demonstrated any meaningful progress toward that strategy when I snagged in. The foundry revenue was negligible, the yield challenges were unresolved, and the competitive gap with TSMC was widening, not narrowing. I snagged a story rather of a business, and I dropped $12,000 for the privilege of learning the difference.

The $6,585 I lost could have gone toward a balance transfer on my credit card, which was carrying $8,200 at 22.9% APR. Paying off high-interest debt with that money would have saved me approximately $1,800 in annual interest charges. That is real, guaranteed, compounding savings. The Intel trade offered nothing guaranteed except the possibility of loss, an the loss was the part that materialized. Next time a CEO stands on a stage and promises a transformation, I will wait for the earnings to confirm the story before I put my money on the line. The market rewards patience, and I am ultimately learning to be patient enough to collect that reward.