TradingKey - Against the backdrop of TSMC's long-standing dominance in the advanced process market, Intel (INTC) is attempting to return to the center of competition through Intel 18A.

Intel 18A is the first to introduce RibbonFET gate-all-around transistors and PowerVia backside power delivery technology, and has entered the mass production stage, with the first batch of products including Intel's self-developed Panther Lake processors. According to market reports, the 18A yield rate has improved to approximately 55% to 60%. Although there is still room for improvement compared with mature mass production levels, it is already sufficient to support initial product shipments and customer validation.

With its advanced process, U.S. domestic manufacturing capacity, and pricing lower than TSMC (TSM), 18A is currently viewed as one of the technologies most likely to break TSMC's monopoly on advanced processes. However, whether Intel can truly challenge TSMC depends not only on process parameters, but also on whether yield rates, design tools, customer onboarding speed, and production costs can simultaneously meet commercialization requirements.

Where Does Intel 18A’s Competitiveness Come From?

The biggest technical highlight of Intel 18A is the simultaneous adoption of RibbonFET and PowerVia. RibbonFET is Intel's gate-all-around transistor design, which improves current control while reducing transistor size; PowerVia moves the power delivery lines to the back of the wafer, thereby reducing interference between signal lines and power lines to boost chip performance and energy efficiency.

Intel hopes to leverage these two technologies to compete with TSMC's N2 node in terms of mass-production timing and certain performance metrics. Unlike TSMC, which places greater emphasis on reliable mass production, Intel has chosen to introduce two major architectures simultaneously, aiming to re-establish its image as a process leader through more aggressive technical upgrades.

Price is another key advantage for Intel (INTC) to win customers. According to market sources, Intel's 18A foundry quotes for some prospective clients may be roughly 25% lower than TSMC's. For chip design companies such as Apple (AAPL), Nvidia (NVDA), and Broadcom (AVGO), even if they do not transfer core products to Intel on a large scale in the short term, adopting 18A for testing can mitigate supply chain risks to some extent while enhancing their bargaining leverage in negotiations with TSMC.

In addition, the U.S. government's drive to reshore semiconductor manufacturing also brings opportunities to Intel Foundry. Microsoft (MSFT), Amazon (AMZN), and U.S. defense-related projects have already established 18A partnerships with Intel, making it one of the few suppliers alongside TSMC capable of simultaneously offering advanced nodes, advanced packaging, and U.S. domestic capacity.

18A Mass Production Does Not Equal Winning External Customers

Entering mass production on 18A is only the first step; what truly determines the future of Intel Foundry is whether external customers can successfully complete chip designs and generate stable orders.

Panther Lake is primarily an internal product for Intel, helping the company validate the yield, performance, and manufacturing processes of 18A, but it cannot fully prove that Intel possesses the capability to serve diverse customers. Foundry operations require adapting to a wide array of chip architectures, design rules, and third-party tools, and external customers often demand higher PDK maturity, IP ecosystems, and product delivery stability than internal units do.

Market rumors previously indicated that Broadcom maintained reservations about certain manufacturing processes and design tools after testing 18A. While this does not mean 18A has lost its competitiveness, it shows that the process remains some distance away from taking on large-scale high-end external orders. Even if yields reach 55% to 60%, Intel still needs to further improve production efficiency to lower unit chip costs and ensure stable delivery.

Meanwhile, financial pressures cannot be ignored. Advanced fabs require continuous capital investments of tens of billions of dollars, while Intel Foundry remains in a loss-making phase. If external customer order growth is slower than expected and 18A capacity utilization falls short, depreciation and R&D expenses will continue to drag down Intel's profits. Although lower pricing may help win customers, it could also compress gross margins, leaving Intel in a situation where 'orders increase but earnings improvements remain limited'.

Therefore, the core issue facing 18A is not whether it can produce chips, but whether it can sustain mass production for external customers with sufficiently high yields, reasonable costs, and reliable delivery capabilities.

Can Intel Challenge TSMC?

From a technical perspective, Intel 18A grants Intel renewed eligibility to compete in advanced process nodes, but challenging TSMC remains exceptionally difficult.

TSMC's advantage does not stem solely from the N2 node itself. Its true technological edge lies in years of accumulated high-yield manufacturing experience, a massive customer base, and an advanced packaging ecosystem built on CoWoS and 3DFabric. For AI chip customers, securing sufficient advanced packaging, HBM integration, and stable capacity is often just as crucial as transistor performance.

TSMC also boasts long-term customers including Apple, Nvidia, AMD (AMD), Qualcomm (QCOM), and Broadcom. Switching foundries requires chip design companies to perform redesign, verification, and software adaptation, which involves high costs and long cycles; therefore, they will not easily transfer core orders based solely on price advantages.

This means that 18A is more likely to serve as a secondary supply source alongside TSMC in the short term, rather than directly replacing TSMC. Even if Apple or Nvidia adopt Intel's process, they are likely to start with non-core chips, small-scale orders, or specific U.S. domestic projects, rather than immediately shifting flagship processors or AI GPUs.

In contrast, 18A may first impact Samsung. Samsung also seeks to challenge TSMC, but its advanced process nodes have long been constrained by yield issues and customer volume. If Intel can stably mass-produce 18A and leverage U.S. manufacturing, advanced packaging, and price advantages to win external orders, Samsung's position as the world's second-largest foundry will come under greater pressure.

Summary

Overall, the strategic significance of Intel 18A lies in helping Intel re-establish technical credibility and proving that its IDM 2.0 strategy and foundry transformation remain viable. If 18A can improve yields, refine its PDK, and secure more external customers, it will position Intel as a truly influential alternative supplier in the advanced process market.

However, the contest between Intel and TSMC will not end with the 18A generation. The true showdown may occur at the 14A node, when Intel must not only prove that its technology can achieve mass production, but also demonstrate that external customers are willing to place ongoing orders and that its foundry business can yield reasonable returns.

Therefore, success with 18A does not mean Intel is about to replace TSMC, but it could serve as a critical starting point for its return to the top tier of advanced nodes. In the short term, Intel's most realistic goal is to stabilize mass production, improve yields, and expand its external customer base; whether it can challenge TSMC over the long term depends on whether 14A can sustain technical progress and translate technology advantages into real scale, profitability, and customer trust.

Find out more