Overview

Intel’s 18A process node has reached volume production in the first half of 2026, and its enhanced variant, 18A‑P, entered risk production in June 2026. These milestones mark the most significant manufacturing advancement for Intel in years and are central to the company’s foundry turnaround strategy. Simultaneously, reports of a preliminary foundry agreement worth approximately $10 billion between Intel and Apple have intensified speculation that Apple will become an anchor customer for Intel’s leading‑edge nodes. This report examines the manufacturing readiness of 18A‑P, the status of external customer onboarding, the details and strategic rationale of a potential Apple deal, the financial and competitive implications for Intel and TSMC, and the broader industry and geopolitical context in mid‑2026.

Manufacturing Readiness

18A Volume Production and 18A‑P Risk Production

Intel’s 18A process is now in volume production at fabs in Oregon and Arizona [1][2]. At Computex 2026 (June 2026), Intel confirmed that 18A is being used to manufacture its first data‑center CPU, the Xeon 6+ (Clearwater Forest), which features up to 288 Darkmont E‑cores and was launched on June 1, 2026 [8]. The Arc G3 processor series for gaming handhelds, also built on 18A, was launched at the same event [9]. Intel’s Core Ultra 3 series, also on 18A, won the Computex 2026 Gold Award and is already in hundreds of commercial designs [11].

At the 2026 VLSI Symposium, Intel Foundry announced that the first performance enhancement of the 18A family, Intel 18A‑P, has entered risk production, meeting previously shared timelines [4][6]. 18A‑P delivers a 9% higher performance at iso‑power or 18% lower power at iso‑performance compared to baseline 18A, along with enhanced thermal characteristics and expanded design flexibility [6]. The node is fully design‑rule compatible with Intel 18A, enabling straightforward IP reuse [6].

Key Technical Innovations: RibbonFET and PowerVia

Intel 18A is the industry’s first process node to bring both gate‑all‑around (GAA) transistors (RibbonFET) and backside power delivery (PowerVia / BSPD) to market simultaneously [6]. At VLSI 2026, Intel quantified the benefits: an 11% routed area reduction versus frontside interconnect, a 10× dynamic voltage droop reduction, and the ability to achieve up to 6% frequency uplift or >15% dynamic power reduction compared to a comparable frontside interconnect technology [6]. Silicon results from CPU cores built on GAA and BSPD showed ~30% frequency improvement at low voltage (~0.5V) and reduced IR drop [6].

Specific advancements in 18A‑P include a new dual‑contact low‑resistance transistor option called “Power Boost,” 20–40% improved thermal resistance, 10–30% improved via resistance, mobility enhancement via PMOS strain engineering, and new low‑power and high‑performance transistor options [6]. Intel also presented long‑term research updates, including monolithic CFET inverters at 45nm gate pitch, 300mm monolithic integration of gallium nitride (GaN) power devices with silicon logic, and subtractive ruthenium interconnects with airgap integration achieving up to ~35% capacitance reduction [6].

Performance and Yield Comparison with TSMC N2 and Samsung SF2

TSMC N2 is the primary competitor. TSMC’s N2 yields were estimated at approximately 65% and climbed to around 70% by early 2026, targeting ~75% at maturity [2]. Intel’s 18A yield reportedly rose from around 50% to 55% by mid‑2025, with internal aspirations of reaching 65% to 70% by year‑end 2025 [2]. As of May 2026, yields remain a critical differentiator: Intel’s 18A yields are still 10–15 percentage points behind TSMC N2 [2]. Intel CEO Lip‑Bu Tan acknowledged that Intel had lost some of its historical manufacturing strength and recruited Sean Han from Samsung’s foundry business to improve yields [5]. Tan stated that yields are now improving at 7–8% per month, exceeding expectations [23].

Samsung SF2 (2nm GAA) is also a competitor. Samsung’s 2nm family includes SF2, SF2P, SF2P+, and SF2X (targeting AI and HPC) [16]. Samsung’s Taylor, Texas fab will begin production in 2027, manufacturing Tesla’s AI5 and AI6 chips, and will deploy advanced 2nm process technology [16]. However, industry observers note that TSMC’s N2P architecture is superior to Samsung’s second‑generation 2nm GAA, and Samsung’s 2nm GAA is reportedly more expensive to manufacture [19]. Google is in talks with Samsung to manufacture a memory‑facing I/O die for its next‑generation TPU “Icefish” using Samsung 2nm, with mass production targeted as early as 2028 [20].

In terms of technology position, Intel’s 18A has narrowed the gap with TSMC, and Intel’s future 14A node may adopt next‑generation High‑NA EUV lithography before TSMC, potentially offering a temporary edge [2]. Intel’s price‑to‑sales ratio has soared from roughly 1.6× a year ago to almost 10× currently, reflecting increasing investor confidence that Intel can serve as a credible second source for advanced semiconductor manufacturing in the United States [2].

Foundry Customer Onboarding

Apple Preliminary Agreement

Multiple credible sources confirm that Apple and Intel have reached a preliminary foundry agreement. The Wall Street Journal reported on May 8, 2026, that the two companies had reached a preliminary chip‑making agreement [23]. Benzinga reported on May 17, 2026, that “following over a year of negotiations, the two companies have finalized a deal” [22]. Thelec.net reported on May 20, 2026, that “Intel recently signed a preliminary $10 billion foundry agreement with Apple” [5]. Intel CEO Lip‑Bu Tan declined to comment on the Apple reports in a CNBC “Mad Money” interview on May 18, 2026, but emphasized the strategic importance of bringing advanced chip manufacturing back to the U.S. [23][25]. TechSpot noted on June 15, 2026, that Apple is negotiating with Intel to use its 18A foundry node to reduce dependence on TSMC [7].

Google TPU Order

Google has placed orders to manufacture more than 3 million tensor processing units (TPUs) by 2028 after extensive testing of Intel’s advanced packaging systems [14][27]. Intel will serve as a second source alongside TSMC [27]. This represents a major validation of Intel’s foundry capabilities, particularly in advanced packaging.

Nvidia Trials

Nvidia is running early trials of Intel’s 18A process and evaluating complex multi‑chip designs for AI data‑center components [14][27]. Like Google, Nvidia is seeking a backup chip supplier to TSMC, which is currently production‑constrained [27].

Other Customers

• Tesla / Terafab: In April 2026, Intel announced a partnership with Elon Musk to develop the Terafab semiconductor factory in Texas, involving Tesla, xAI, and SpaceX [5][14].
• AWS and Cisco: According to channel sources cited in Forbes, AWS and Cisco are current customers of Intel’s advanced packaging services [1].
• Microsoft: Microsoft is in discussions regarding Intel’s advanced packaging [1].
• Hitachi: Intel expanded partnerships with Hitachi focused on physical AI applications in robotics and industrial systems [14]. On June 5, 2026, Hitachi and Intel announced a strategic collaboration covering foundry tools, quantum computing, energy optimization, custom silicon, and factory automation [29].
• Other discussions: Apple, Google, Microsoft, Nvidia, and Tesla are in discussions with Intel Foundry for advanced packaging [1].

CEO Lip‑Bu Tan expects commitments from multiple foundry customers in the second half of 2026 [23]. CFO David Zinsner previously indicated that external foundry customer signals would become “more concrete” in H2 2026 and early 2027 [23].

Design Enablement and IP Portfolio

Intel has provided the 0.5 PDK (alpha‑stage design kit) for its 14A process to customers, with the 0.9 PDK (beta‑stage) set for external distribution in October 2026 [5]. The 18A‑P node is fully design‑rule compatible with Intel 18A, enabling straightforward IP reuse [6]. Intel’s advanced packaging technologies include EMIB (2.5D) and Foveros (3D die‑stacking), and the Rio Rancho, New Mexico facility is the most advanced U.S.‑based integrated packaging facility, now offering silicon photonics and expected to be the first site worldwide for volume glass substrate production [1].

Advanced Packaging Leadership

Intel Foundry’s advanced semiconductor packaging business is already outperforming its fabrication arm and driving early foundry success [1]. Tirias Research analyst Jim McGregor, after touring Rio Rancho, stated: “It didn’t feel like walking into an Intel facility. It felt like walking into a true semiconductor foundry that is willing to do anything for its customers” [1]. The Packaging and Test Business Group is already profitable, and CFO David Zinsner noted that early customer engagements suggest many advanced packaging opportunities could exceed $1 billion each [1]. Strategic partnerships exist with SK Hynix (HBM memory) and Amkor Technology (OSAT services) [1].

Apple Foundry Deal Analysis

Deal Terms and Timeline

The preliminary agreement is reported to be worth approximately $10 billion [5]. Negotiations have been ongoing for over a year (since approximately early 2025) [21][22]. The deal was finalized or reached a preliminary stage by May 2026 [22][23]. Specific terms have not been disclosed, but the agreement likely covers multiple generations of Apple Silicon.

Chips Involved: M7 on 18A‑P, Future iPhone SoCs on 14A

The first candidate is the M7 processor for MacBooks, targeting Intel’s enhanced 18A‑P process, with a potential market release in late 2027 [21]. Further collaboration is under discussion for iPhone SoCs using Intel’s next‑generation 1.4nm 14A process, contingent on the M7 partnership proving viable. This could eventually produce a mobile SoC for iPhones—possibly the A22 generation in 2028 or A23 in 2029 [21]. Apple could also choose less critical chips (modems, Wi‑Fi/Bluetooth) for Intel production, though specialized analog circuitry may complicate migration from TSMC [21].

Strategic Motivations for Apple

Supply chain diversification is Apple’s primary motivation. Over 90% of advanced processors are manufactured overseas, primarily in Taiwan [23][25]. The Iran war has created significant headwinds for semiconductor supply chains, with disruptions to helium (Qatar provided over 30% of the global market in 2025, export capacity reduced by Iranian strikes), energy, and freight costs [26]. CNBC reported on May 19, 2026, that “the Iran war is creating significant headwinds for semiconductor and tech hardware companies that power the AI boom” [26]. By engaging Intel as a second source, Apple gains negotiating leverage against TSMC in pricing negotiations [21]. Intel will likely compensate Apple through lower manufacturing costs, making the arrangement profitable for Apple overall [21].

Geopolitical risk management is also critical. The Trump administration’s push for domestic U.S. manufacturing provides strong incentives, as Intel is the only major foundry headquartered in the U.S. [21]. Intel CEO Lip‑Bu Tan stated: “90 plus percent of the most advanced processor is manufacturing outside the country. So, I think it’s important to bring some of them back” [23][25].

Performance and cost trade‑offs: Intel’s manufacturing processes are described as slightly behind TSMC, which could result in somewhat lower clock speeds and performance [21]. However, lower manufacturing costs from Intel could offset performance differences and make the arrangement profitable for Apple [21].

Strategic Motivations for Intel

Anchor customer validation: Securing Apple as a foundry customer would be the ultimate validation for Intel’s foundry ambitions, providing a marquee name that attracts other customers [21][23]. Intel Foundry earned only about $174 million from outside customers in Q1 2026, accounting for roughly 3% of total foundry revenue of $5.4 billion [2]. TSMC reported $35.9 billion in foundry revenue in Q1 2026, nearly all from external clients [2]. Intel Foundry recorded a loss exceeding $10 billion last quarter [2]. A $10 billion Apple deal would dramatically boost external revenue and credibility.

Revenue boost and path to profitability: Intel Foundry aims to break even in 2027 and be profitable by 2030 [1]. The Apple deal, combined with Google and Nvidia engagements, could accelerate this timeline. Intel’s “Rule of 45” targets combined revenue growth and operating margin above 45% [5].

National security narrative: The foundry strategy—manufacturing chips for external clients while rebuilding U.S. advanced chip production—is central to Intel’s turnaround [23]. Intel shares have surged over 300% since Lip‑Bu Tan was appointed CEO in March 2025 [23].

Strategic and Financial Implications

Impact on Intel Foundry Financials

An Apple deal would transform Intel Foundry’s revenue profile. In Q1 2026, Intel Foundry’s external revenue was negligible ($174 million) [2]. A $10 billion agreement over several years would represent a massive increase. Intel expects Q2 2026 revenue in the range of $13.8 to $14.8 billion, driven by strong AI demand [4]. The foundry business is targeting breakeven in 2027 and profitability by 2030 [1]. The advanced packaging business is already profitable and could contribute significantly [1].

Intel’s stock has surged over 300% since Tan’s appointment, and its price‑to‑sales ratio has jumped from roughly 1.6× to almost 10× [2]. Bank of America double‑upgraded Intel to Buy with a $135 target [30]. However, Wall Street analysts have a Hold consensus (11 Buys, 24 Holds, 3 Sells) with an average price target of $87.09, implying downside risk after a 408.75% rally over the past year [34].

Implications for TSMC’s Market Position

TSMC reported net income of $54.55 billion for 2025 on consolidated revenue of $120.95 billion, driven by strong AI demand [18]. TSMC’s operating margin was 58.1% in Q1 2026 [2]. An Apple‑Intel deal would erode TSMC’s near‑monopoly on leading‑edge logic for Apple, potentially reducing TSMC’s pricing power and forcing it to compete more aggressively on cost and capacity. However, TSMC’s scale, yield advantage, and customer trust remain formidable. AMD announced on May 21, 2026, that it has begun production ramp of its 6th Gen EPYC processor “Venice” on TSMC’s 2nm process, marking the first HPC product in the industry to enter production on TSMC 2nm [15]. TSMC’s Japan fab (JASM) achieved its first quarterly profit, signaling that TSMC’s overseas expansion is viable [24].

Intel’s Product vs. Foundry Business Balance

Intel must balance its internal product business (CPUs, GPUs) with its foundry ambitions. The foundry business requires significant investment in capacity and R&D, while Intel’s product business faces intense competition from AMD and Nvidia. CEO Lip‑Bu Tan has laid off ~34% of staff, paused factory expansions, flattened bureaucracy, and brought in new leadership from Qualcomm and Arm [31]. The U.S. government took a 9.9% equity stake in Intel in August 2025 by converting $5.7 billion in CHIPS Act grants and $3.2 billion in other awards [23][32]. Intel’s internal products, such as the Xeon 6+ and Arc G3, are already using 18A, demonstrating that Intel can serve both internal and external customers on the same node.

Risks and Challenges

• Yield gap: Intel’s 18A yields (~55%) still trail TSMC N2 (~70%) [2]. Closing this gap is critical for profitability and customer confidence.
• Capacity allocation: Intel must allocate sufficient capacity for both internal products and external customers without compromising either.
• IP protection: Apple and other external customers will require robust IP protection, especially when competing with Intel’s own product lines.
• Geopolitical tensions: The Iran war and potential escalation in the Taiwan Strait could disrupt supply chains. Intel’s U.S.‑based fabs offer a geopolitical hedge, but the company must navigate export controls and trade policies.
• Execution risk: Intel has a history of process delays. The 14A node is scheduled for risk production in 2028 and mass production in 2029, with a timeline “almost similar” to TSMC’s A14 [5]. Any delays could undermine customer trust.

Industry and Market Context

Broader Foundry Landscape in Mid‑2026

The semiconductor foundry market in mid‑2026 is dominated by TSMC, which holds a commanding lead in revenue, yield, and customer trust. TSMC’s Q1 2026 foundry revenue was $35.9 billion, compared to Intel Foundry’s $5.4 billion [2]. Samsung Foundry is also a competitor, but its 2nm GAA process is reportedly more expensive and less advanced than TSMC’s N2P [19]. Google is in talks with Samsung for a TPU component, but the main compute die will be on TSMC [20].

Intel’s 18A and 18A‑P represent the most credible challenge to TSMC’s dominance in years. The company’s advanced packaging business is already winning customers, and the combination of GAA transistors and backside power delivery gives Intel a technical edge that competitors are still working to match. Intel’s future 14A node, using High‑NA EUV, could provide a temporary advantage if executed on schedule [2].

CHIPS Act and Geopolitical Factors

The U.S. CHIPS Act has provided significant funding for domestic semiconductor manufacturing. In August 2025, the U.S. government took a 9.9% equity stake in Intel by converting $5.7 billion in CHIPS Act grants and $3.2 billion in other awards [23][32]. President Donald Trump said in a Fortune interview published June 15, 2026, that he should have asked for a larger stake in Intel when negotiating with CEO Lip‑Bu Tan [32]. Intel’s stock has since surged over 300% [23].

The Iran war is creating significant supply chain disruptions. CNBC reported on May 19, 2026, that “the Iran war is creating significant headwinds for semiconductor and tech hardware companies that power the AI boom” [26]. TSMC, Foxconn, and others have flagged supply chain disruptions, including helium shortages and higher energy and freight costs [26]. IDC analyst Francisco Jeronimo stated: “We can expect further negative impact this year…the price of gas, energy and freight are at an all‑time high and are likely to remain high for a few more quarters, even if the situation de‑escalates” [26]. These disruptions strengthen the case for onshoring advanced chip manufacturing, benefiting Intel.

Analyst Projections and Stock Market Reactions

Analyst sentiment is mixed but improving. Bank of America double‑upgraded Intel to Buy with a $135 target, citing agentic AI as a powerful demand accelerant that could expand the CPU total addressable market to over $170 billion by 2030 [30]. GF Securities identified Intel as a major AI winner at Computex 2026, noting its 18A manufacturing process is in volume production and its edge AI ecosystem has over 4,000 partners and more than 100,000 deployments [13].

However, Wall Street analysts have a Hold consensus on INTC (11 Buys, 24 Holds, 3 Sells) with an average price target of $87.09, implying 24.67% downside risk after a 408.75% rally over the past year [34]. Intel stock surged 11.2% to close at $110.27 on June 8, 2026, following reports of Google signing on as a foundry customer [27]. The stock had previously broken out of a cup base at a buy point of $54.60 on April 8, received a boost from its beat‑and‑raise Q1 earnings report on April 23, and hit an all‑time high of $132.75 on May 11 before pulling back [27]. April 2026 was Intel’s best month in the chipmaker’s 55 years on the Nasdaq, with the stock more than doubling [23].

Morningstar chief equity strategist Michael Field noted: “Any disruption so far has been completely overshadowed by the upswing in investor confidence in AI” [26]. William Blair analyst Sebastien Naji warned: “If the blockade continues through the summer, we are more likely to revisit the risks and impacts in future earnings periods” [26].