Intel CEO Lip-Bu Tan arrived in Taiwan on Monday for a set of meetings that will define the week — and may define the year. With Computex 2026 officially underway in Taipei, Tan held private sessions with Taiwan Semiconductor Manufacturing Co. management ahead of his first Computex keynote, scheduled for 1:30 a.m. ET Tuesday. At stake is not just Intel's product strategy but the investment thesis behind a stock that has risen more than 200% in 2026, putting the chipmaker's market capitalization above $614 billion and making the June 2 address one of the most closely watched corporate events in the semiconductor calendar.
Intel confirmed Tan would be in Taiwan but declined to specify his schedule. "As a general practice, our leadership regularly engages with customers, partners, and business leaders as part of ongoing business operations," a company spokesperson said, adding only that Tan would deliver the Computex keynote at 1:30 p.m. local Taipei time.
Intel TSMC Partnership: Rivals, Partners, and a Lawsuit in Between
The meetings carry unusual diplomatic weight. Intel and TSMC are simultaneously manufacturing rivals, supply-chain partners, and former joint-venture candidates — a configuration that makes every interaction between their leadership consequential. Reports surfaced in 2025 that the two companies explored an arrangement in which TSMC would take roughly a 20% stake in Intel Foundry. Those discussions predated the U.S. government's decision to convert approximately $9 billion in CHIPS Act grants into a 9.9% equity stake in Intel — a sovereign shareholder with its own industrial-policy interests that any future Intel-TSMC arrangement would now need to navigate.
Adding complexity to the relationship is the ongoing litigation involving Wei-Jen Lo, a former TSMC senior vice president who joined Intel as executive vice president in October 2025, three months after retiring from TSMC following a 21-year tenure. TSMC filed a lawsuit against Lo in November 2025 alleging he brought confidential trade secrets — including sensitive details on 2nm and A16 process technology — to the rival chipmaker. Taiwanese prosecutors subsequently raided Lo's residences. Intel has stood by the hire, with Tan calling the allegations baseless in an internal memo to employees. The lawsuit remains active. Whatever Tan discusses in his closed-door sessions with TSMC management, the legal dispute over Lo sits visibly in the background of every conversation.
Read more: Nvidia at Computex 2026: Jensen Huang Flies to TSMC as Vera Rubin Ramp Strains Taiwan Supply Chain
Intel 18A Foundry Gap: What $174 Million Means
Whatever diplomatic warmth is on display this week, the competitive arithmetic between Intel and TSMC is unambiguous. TSMC generated $35.9 billion in foundry revenue in the first quarter of 2026. Intel Foundry posted $5.4 billion over the same period — and of that total, only approximately $174 million, or roughly 3%, came from external customers. The remainder came from Intel's own chip designs. TSMC's market capitalization of approximately $1.86 trillion remains nearly three times Intel's current valuation.
TSMC's pricing power is accelerating. According to Commercial Times, citing supply chain sources, the foundry plans to raise the price of its 3nm node by up to 15% in the second half of 2026, with potential further increases of 5% to 10% in 2027. Nvidia, AMD, Google, and AWS are all deepening their dependence on TSMC's 3nm capacity for AI accelerators and custom silicon — the same class of customers Intel's foundry must attract. The preliminary chip-manufacturing agreement Intel reached with Apple in May 2026, reported by the Wall Street Journal, is Intel's most prominent external foundry win to date. Both companies declined to comment. Whether it translates into meaningful production volume at scale remains the central open question.
One complicating detail: reports suggest that more than 90% of Intel's forthcoming Nova Lake desktop processor tiles will be manufactured on TSMC's N2 process rather than Intel's own 18A fabs — a concession to the reality that Intel's in-house yields are not yet at production scale for that platform. Intel has said 18A yield improvements to industry-standard levels are expected in 2027.
Nvidia RTX Spark Dominates Opening Day
Intel faces immediate competitive pressure for the week's narrative. In his GTC Taipei keynote Sunday night, Nvidia CEO Jensen Huang unveiled the RTX Spark — an Arm-based superchip combining a 20-core Grace CPU architecture co-developed with MediaTek with an RTX Blackwell GPU, targeting AI-focused Windows laptops and compact desktop PCs. The platform delivers up to 1 petaflop of AI performance with 128GB of unified memory, Nvidia said. Partners already committed to RTX Spark systems include Dell, ASUS, Lenovo, HP, and Microsoft, with devices expected to ship in fall 2026.
The announcement marks Nvidia's first entry into the Windows client chip market, positioning the company as a direct challenger to Intel and Qualcomm in a segment Intel has historically owned. "Forty years of traditional PCs is now at an end," Huang said during the keynote. The line was received as both a joke about Nvidia's origins and a deliberate statement about competitive intent.
Intel is not without a response. The Arc G3 and Arc G3 Extreme, announced May 28 and built on Intel's 18A process node using Panther Lake architecture, are the company's first chips designed exclusively for gaming handhelds rather than adapted from laptop lines. Early hands-on evaluations at Computex have been favorable, with reviewers noting improved performance and power efficiency relative to AMD-based alternatives. Intel also revealed Monday that its Crescent Island inference GPU — built on the Xe3P architecture and capable of scaling to 480GB of LPDDR5X memory — is targeting customer sampling in the second half of 2026. A preview of Nova Lake, Intel's next-generation desktop processor family, is widely anticipated at Tuesday's keynote but has not been officially confirmed.
Dell AI Backlog Confirms the Opportunity Intel Is Chasing
Beyond the chip-designer competition, perhaps the most significant signal for Intel's long-term thesis arrived from a different quarter entirely. Dell Technologies reported first-quarter fiscal 2027 results on May 28 that exceeded nearly every analyst projection: $43.84 billion in revenue against a $35.46 billion consensus estimate, $16.1 billion in AI-optimized server revenue — a 757% year-over-year increase — and $24.4 billion in AI orders during the single quarter. The company exited the period with a record $51.3 billion AI server backlog.
Dell raised its full-year AI server revenue target to approximately $60 billion, implying 144% year-over-year growth. Full-year total revenue guidance moved to $165 billion to $169 billion. COO Jeffrey Clarke said on the earnings call that the AI business remained "exceptionally strong," explicitly characterizing second-half revenue moderation as supply-constrained rather than demand-constrained. JPMorgan analyst Samik Chatterjee raised his Dell price target to $500 from $280, maintaining an Overweight rating, and affirmed the supply-constraint characterization.
For Intel, this context is strategically favorable. Dell's numbers confirm that enterprise and hyperscaler demand for AI infrastructure is not softening — a rising tide of AI capital expenditure lifts CPU suppliers alongside GPU suppliers. Tan has argued publicly that CPUs are now the "orchestration layer and critical control plane for the entire AI stack." Whether Intel can capture that orchestration role at scale is what Tuesday's keynote is being asked to address.
What Tan Must Deliver Tuesday
The question hanging over the June 2 keynote is whether Tan can convert a remarkable stock run into a durable investment thesis. Intel's Data Center and AI segment posted $5.1 billion in first-quarter 2026 revenue, up 22% year over year, with total quarterly revenue of $13.6 billion beating the $12.3 billion consensus estimate by more than $1 billion.
Bulls point to the 18A process node entering volume production, the preliminary Apple manufacturing agreement, the Terafab deal with Elon Musk's consortium spanning SpaceX, Tesla, and xAI, and Nvidia's $5 billion investment in Intel from September 2025. Bears note that the foundry business remains deeply loss-making, 18A yields are not expected to reach industry-standard levels until 2027, and the stock now trades at a valuation that leaves little margin for execution errors.
Investors going into Tuesday's session want specificity: confirmed external foundry customers beyond Apple, progress metrics on 18A yields, and an update on Crescent Island's competitive positioning against Nvidia's inference-optimized products. Tan's meetings with TSMC this week — whatever their content — are a reminder that Intel's path forward runs directly through Taipei, whether as a competitor, a customer, or eventually a partner. The conversation is very much underway. The proof of concept is on the stage Tuesday.
Frequently Asked Questions
Why is Intel CEO meeting with TSMC at Computex 2026?
Tan's meetings with TSMC management reflect the complex relationship between the two companies, which are simultaneously manufacturing rivals, supply-chain partners, and former joint-venture candidates. A reported 2025 proposal would have seen TSMC take a roughly 20% stake in Intel Foundry. The meetings also occur against the backdrop of an active lawsuit: TSMC filed suit in November 2025 against former executive Wei-Jen Lo, who left TSMC for Intel, alleging he took trade secrets covering 2nm process technology.
What will Intel announce at Computex 2026?
Intel has confirmed CEO Lip-Bu Tan will deliver the keynote at 1:30 p.m. Taipei time Tuesday, focused on AI computing across PCs, the data center, and the cloud. Intel has already revealed the Crescent Island inference GPU with up to 480GB of LPDDR5X memory at Computex pre-keynote sessions. A preview of Nova Lake, Intel's next-generation desktop processor family, is widely anticipated but has not been officially confirmed. Progress updates on the Intel 18A process roadmap and foundry customer additions are expected to be central to the address.
Can Intel's stock turnaround survive the Computex keynote?
Intel's stock has gained roughly 206% year to date in 2026, driven by the Terafab deal, the preliminary Apple manufacturing agreement, Nvidia's $5 billion investment, and a Q1 earnings beat. However, Intel Foundry still generates only about $174 million quarterly from external customers, the 18A node is not expected to reach industry-standard yields until 2027, and the foundry business remains unprofitable. Whether the keynote delivers on the elevated expectations embedded in the stock price is the central investor question this week.
How does Intel's foundry compare to TSMC in 2026?
TSMC generated $35.9 billion in foundry revenue in the first quarter of 2026, with a market capitalization of approximately $1.86 trillion. Intel Foundry posted $5.4 billion over the same period, of which only approximately $174 million — roughly 3% — came from external customers. TSMC has announced a planned 15% price increase on its 3nm node for the second half of 2026, driven by surging demand from Nvidia, AMD, Google, and AWS — the same hyperscaler customers Intel's foundry is working to attract.
ⓒ 2026 TECHTIMES.com All rights reserved. Do not reproduce without permission.
