TSMC vs. Samsung vs. Intel: 2nm Process Showdown. TSMC Q1 Earnings to Decide the Winner

TradingKey - TSMC (TSM) TSMC will release its first-quarter earnings this Thursday (April 16), having already reported record quarterly revenue of NT$1.134 trillion (approximately $35.6 billion), a year-on-year increase of about 35% and the first time it has crossed the NT$1 trillion threshold.

Furthermore, LSEG projects that TSMC's net profit for this quarter will increase by 50% to NT$542.6 billion (approximately $17.1 billion), which would mark its fourth consecutive quarter of record earnings.

Analysts point out that the most critical technical metrics to monitor during the earnings call are the yield and ramp-up rate of the 2nm process, which will be the key battleground for TSMC to widen its lead over other chip giants.

What is the 2nm process node? Which companies compete with TSMC?

The 2nm process, or N2 process, is a specific semiconductor manufacturing node. While it should not be simply understood as the actual physical dimensions of a component being 2nm, this process represents the pinnacle of precision structures currently capable of mass production.

Compared to the 3nm process currently in use, 2nm will deliver an increase in transistor density; for the same workload, the latter is faster or more power-efficient.

Currently, the 2nm processes from TSMC and Samsung, and Intel's (INTC) 18A—essentially the 1.8nm process—are in the same performance tier despite their different names.

TSMC Q1 Earnings: Focus on 2nm Yield and Ramp-up Slope

By the end of 2025, TSMC's 2nm process had entered the mass production stage, with the Hsinchu Baoshan plant primarily catering to the initial demand from core customers such as Apple (AAPL) and others, while capacity at the Kaohsiung plant is still ramping up rapidly.

The performance-enhanced N2P version based on the first-generation 2nm process can increase speed by approximately 5% at the same power consumption and is expected to enter production in the second half of 2026. Even more advanced than N2P is A16, an angstrom-level chip equivalent to a 1.6nm process, which represents another major leap on TSMC's technology roadmap. Compared to N2P, A16 is 8–10% faster at the same voltage. This process is also slated for production in the second half of 2026.

Currently, among these three nodes, the 2nm process warrants the most attention, as TSMC and its competitors, Samsung and Intel, engage in fierce competition at this technology node. The 2nm process is currently the most advanced technology with the potential for the fastest commercialization.

This earnings call should focus on yield and ramp-up slope, as their mass production maturity will determine the trajectory of TSMC's average selling price (ASP) for 2026-2027. Yield refers to the percentage of qualified chips out of the total output; higher yields mean fewer scrapped chips and lower costs. The ramp-up slope refers to the speed at which capacity increases during the transition from pilot production to large-scale mass production; a higher ramp-up slope enables faster market capture.

These two indicators are reflected in the monthly capacity levels of the Hsinchu and Kaohsiung plants that have already entered mass production. If TSMC can maintain its capacity lead and reduce costs—given that 2nm unit prices are approximately 30%–50% higher than the current 3nm—the mass production maturity of the 2nm process will directly determine the direction of TSMC's average selling price for 2026-2027.

Analysis suggests that the 2nm process will become the node generation with the longest lifecycle in TSMC's history. From a technical perspective, 2nm is approaching physical limits and is difficult to surpass in the short term. Consequently, the 2nm family (including N2P and A16) will offer the highest performance among mass-produced chips, and TSMC's leadership in this area may allow its gross margin to stabilize above 60%.

TSMC vs. Samsung vs. Intel: Who Leads the 2nm Chip Race?

Yield: TSMC Maintains Absolute Lead

According to information from early 2025, TSMC's 2nm yield has stabilized between 60% and 70%.

According to the latest reports from April 13, Samsung's 2nm yield is only around 55%, lagging behind TSMC by approximately 10 percentage points and failing to reach the levels required to compete for critical fabless foundry orders. Industry insiders have even pointed out that when accounting for performance binning and losses in back-end packaging and testing, Samsung Electronics' 2nm yield for final products drops to just 40%, indicating that its 2nm process is far from mature.

Information from mid-2025 shows that Intel's 18A process had an initial mass production yield of only 55%-60%. Intel's CFO previously disclosed that yields will not reach commercially acceptable cost levels until the end of 2026, with industry-standard levels not expected until 2027.

Ramp-up Slope: TSMC Features the Steepest Curve

Driven by massive chip demand from Apple, TSMC must scale from the start of mass production to a monthly output of tens of thousands of wafers in a short period; consequently, its 2nm ramp-up slope is currently the steepest in the industry.

Intel is currently focused on 18A, but it is noteworthy that its 18A orders this year are primarily driven by internal demand—specifically for its own products like the Panther Lake processor. Therefore, its ramp-up slope depends largely on the market reception of its internal products, and it is expected to be more gradual compared to TSMC.

Among the three, Samsung faces the most severe issue: customer attrition. Latest reports indicate that while Samsung's 2nm yield has improved from 20% last year to the current 60%, Qualcomm (QCOM) continues to award orders for its next-generation Snapdragon 8-series flagship chips to TSMC due to Samsung's capacity constraints. Without the support of massive orders, Samsung's ramp-up slope is also expected to be relatively flat.

Cost: TSMC Boasts Lowest Costs and Highest Gross Margins

TSMC is widely recognized as the manufacturer most capable of driving down 2nm costs, thanks to its exceptional yields and massive shipping volumes. High yields mean less waste, while high volumes allow for better amortization of facility and equipment depreciation costs. However, for customers, TSMC's products are the most expensive, resulting in the highest gross margins for the company.

Samsung primarily relies on the vertical integration of its entire supply chain. It typically offers customers more attractive foundry pricing than TSMC, largely because Samsung controls everything from design and foundry services to memory and packaging, allowing for internal profit hedging to lower overall chip costs.

Currently, Intel is less concerned with foundry margins, as its primary customers are internal. This bypasses the need for foundry margin negotiations, allowing the company to focus solely on the final gross margins of the overall product.

Technical Milestones: Intel Leads

Although TSMC has implemented backside power delivery in its A16 process, Intel was the pioneer of this technology, having fully integrated it at the 18A node.

TSMC does not currently utilize backside power delivery in its base 2nm process, which means that in specific high-performance computing (HPC) scenarios, Intel's architecture may hold an advantage in energy efficiency.

Customer Ecosystem: Absolute Dominance by TSMC

Currently, TSMC is the most sought-after player in the 2nm market; not only has Apple reserved the majority of its initial capacity, but NVIDIA (NVDA) is also deeply tied to TSMC for its next-generation AI architecture.

Although Samsung struggles to compete with TSMC, it did secure a $16.4 billion order from Tesla (TSLA) in July of last year to produce Tesla's AI6 chips using the 2nm process.