Intel Drops 6%, JPMorgan Short: What the 18A Bet Now Costs
Intel fell 6%+ today as JPMorgan named it a top short. With Q2 earnings July 23, the foundry revenue number is all that matters now.
Key Points
- Intel dropped more than 6% today to approximately $110.32, and JPMorgan named it a top short after a 650% rally from its August 2025 low of $18.97 wiped out the valuation discount the foundry thesis once offered.
- External foundry revenue came in at just $174 million in Q1 2026 — the single number that exposes the gap between Intel's recovery narrative and its actual financial results.
- Q2 earnings on July 23 are the make-or-break event: an external foundry revenue step-up toward $300–$400 million is the only data point that credibly challenges JPMorgan's short thesis.
Intel fell more than 6% today to $110.32, and the move was not random sector noise — JPMorgan formally named the stock a top short idea, making the explicit argument that a 650% rally from the August 2025 low of $18.97 to the June 30, 2026 all-time high of $142.35 has already priced in a foundry and AI recovery that has not yet appeared in a single quarter of hard financial results.
How Intel Got Here
The 650% rally that ran from August 2025 to the end of June 2026 was one of the most dramatic reversals in large-cap semiconductor history, and it was built on three interlocking catalysts. First, the Trump administration took a 9.9% stake in Intel in August 2025, effectively removing the tail risk that the U.S. government would allow the country's only domestic advanced logic foundry to fail — a structural floor that justified a significant re-rating of the stock. Second, Intel's 18A process node began generating serious industry attention as a potential alternative to TSMC's N2 for customers who want U.S.-based production, with the geopolitical tailwind from the CHIPS Act and ongoing Taiwan tension concerns making domestic sourcing a real procurement consideration for defense and enterprise customers. Third, Intel announced a €5 billion — approximately $5.7 billion — investment to expand its Leixlip, Ireland campus for data center CPU and foundry production, signaling that the company was committing capital to the roadmap rather than retreating.
None of those three catalysts produced revenue. That is JPMorgan's core argument, and on the numbers, it is difficult to refute. Intel's external foundry business — the unit that must succeed for the IFS (Intel Foundry Services) thesis to be validated — generated $174 million in Q1 2026. Against TSMC's $39.62 billion quarter, that figure is not a foundation; it is a placeholder. The 18A node, where Intel's competitive hopes are most concentrated, reportedly may not achieve profitable yields until 2026 or 2027 depending on whom you ask, and the market has started to notice that profitable yields and production commitments from external customers are two different milestones, with the latter significantly harder to achieve.
The political dimension is now cutting both ways. The Trump administration's 9.9% stake structurally prevents Intel's worst-case scenario — a complete foundry shutdown or forced sale — but it also introduces a layer of political complexity that institutional investors price as risk, not reward. When Trump's involvement re-emerged as a news driver yesterday afternoon, the stock sold off sharply, suggesting the market is increasingly treating the government relationship as an overhang rather than a backstop. UBS raised its price target to $121 from $83, and TD Cowen lifted to $115 from $75, both acknowledging improved operational momentum — but both targets are below the June 30 high of $142.35, which tells you where the consensus view on fair value actually sits.
The July 23 Earnings Setup
Intel's Q2 2026 earnings report, due after the close on July 23, is the event that either validates or destroys the current price level. Management guided Q2 non-GAAP EPS of approximately $0.20, with revenue of $13.8–$14.8 billion. Analyst consensus sits at $0.10–$0.21 for EPS, a range wide enough to signal genuine uncertainty about execution. The revenue midpoint of $14.3 billion represents a business that is stabilizing rather than accelerating — which might be adequate for a semiconductor company trading at a reasonable multiple, but is not adequate for a stock that ran 650% in under a year.
The single metric that matters most on July 23 is external foundry revenue. Q1 came in at $174 million. For the bull thesis — the one baked into the stock's peak valuation — that number needs to show a credible step function upward, with $300–$400 million representing the threshold at which analysts can begin building a multi-year IFS revenue model with defensible assumptions. Anything below $250 million will likely be read as evidence that 18A is still in a pre-commercial phase, meaning the foundry recovery is at least two to three quarters further away than the peak price implied. JPMorgan's short thesis is essentially a bet that the July 23 report will not produce that step-up.
The Ireland expansion deserves specific attention as a secondary catalyst watch. The €5 billion Leixlip investment is real capital commitment — not a roadmap promise — and the campus expansion targets data center CPU production alongside foundry capacity. If management provides updated timelines on when Leixlip capacity comes online and what portion is committed to external foundry customers versus Intel's own product lines, that disclosure would add specificity to a thesis that has been frustratingly vague on production timelines. The distinction matters because capacity committed to Intel's own CPUs does not help the IFS revenue case; only capacity allocated to external customers — and ideally, named external customers — moves the needle on the foundry narrative.
What the TSMC Comparison Actually Shows
The comparison between Intel and TSMC this week is uncomfortable reading for Intel bulls. Bloomberg's coverage today highlighted that TSMC and ASML results are carrying unusual weight this week precisely because the AI chip trade is under valuation stress — meaning investors are actively reassessing which companies have the hard revenue to justify their multiples. TSMC answered that question with $39.62 billion and 36% year-over-year growth. Intel's foundry business answered with $174 million.
The regulatory backdrop adds another layer of complexity. The Supreme Court ruling that ended FTC independence — and that Axios reported will have major implications for AI regulation — changes how Intel might navigate any future partnerships or joint ventures involving its foundry business. A politically weakened FTC is theoretically more permissive toward deals, which could eventually benefit Intel if it pursues foundry joint ventures with large AI customers. But that is a long-horizon scenario, not a Q3 catalyst. The White House voluntary frontier model standards framework has an August 1 deadline, and while it is AI-model focused rather than chip-manufacturing focused, the broader regulatory environment shapes how hyperscalers think about domestic versus offshore supply chains — a consideration that could eventually route more business toward Intel IFS if geopolitical pressure intensifies.
The trade setup for July 23 is binary by nature. Bulls need external foundry revenue above $250 million and any named customer commitment for 18A production — even a single design win disclosed publicly would reset the sentiment. Bears need confirmation of what Q1 already suggested: that $174 million is the realistic near-term ceiling, not the floor. At $110.32 today, the stock is 22% below its June 30 all-time high and 7.8% above UBS's newly raised $121 target and just above TD Cowen's $115 target — a range that implies the sell-side believes the worst of the correction is done but that a return to $142 requires earnings proof, not just political support. Watch $105 as the first technical support level below today's close; a break there before July 23 would signal the market is pre-positioning for a disappointing foundry revenue print.
The Weekly Investor
Daily market analysis for active traders. Free.
