On April 22, 2026, Taiwan Semiconductor Manufacturing Co. (TSM) presented its latest semiconductor manufacturing innovations at its yearly technology symposium held in Santa Clara. The company’s shares closed 5% higher that day.
Taiwan Semiconductor Manufacturing Company Limited, TSM
The centerpiece of the presentation featured two new manufacturing nodes: A13 and N2U. The A13 process represents an evolution of TSMC’s current A14 technology and specifically targets artificial intelligence chip manufacturing, with mass production scheduled for 2029. Meanwhile, N2U serves as a cost-effective alternative designed for consumer electronics including smartphones and laptops, as well as AI applications, with rollout planned for 2028.
While both technologies enable the production of more compact and faster semiconductors, the performance improvements at the individual chip level are characterized as incremental rather than revolutionary.
Perhaps more significant is what TSMC has decided to forgo. The semiconductor giant announced its intention to bypass ASML’s latest “high NA” extreme-ultraviolet lithography systems for the foreseeable future. These cutting-edge machines carry a price tag of approximately $400 million per unit—essentially double the cost of the older EUV equipment currently in TSMC’s facilities.
Kevin Zhang, serving as TSMC’s deputy co-chief operations officer, explained to Reuters that the company’s research and development efforts have identified methods to extract additional performance from existing EUV technology. “This is definitely a strength,” Zhang stated.
ASML’s stock price dropped approximately 1% in response to the announcement. The reduced urgency for equipment upgrades translates to near-term challenges for the Netherlands-based semiconductor equipment manufacturer.
The packaging announcements from Wednesday’s event also garnered significant interest. TSMC revealed that by 2028, its facilities will possess the capability to integrate 10 large computing chips with 20 high-bandwidth memory stacks within a single package. For context, contemporary AI processors such as Nvidia’s Vera Rubin—manufactured by TSMC and scheduled for release this year—incorporate just two large chips and eight memory stacks.
This multi-chip stacking approach represents the industry’s primary strategy for sustaining performance growth as the benefits from transistor miniaturization diminish. Dan Hutcheson from TechInsights characterized this evolution as Moore’s Law “morphing from a monolithic, single die in a package to multi-die in a package.”
However, chip stacking technology presents significant challenges. Thermal management issues and mechanical stress resulting from different materials’ expansion rates can lead to package warping or fracturing. Ian Cutress from More Than Moore pointed out that Nvidia’s Rubin processor had encountered precisely these types of complications. He also observed that TSMC did not explicitly address its solutions for these thermal and mechanical challenges.
TSMC currently trades at a P/E ratio of 32.18x and holds a GF Score of 96 out of 100. Corporate insider transactions over the previous 12 months reveal 33 purchases with no sales.
Notwithstanding these positive indicators, GuruFocus’s GF Value model presently categorizes the stock as “significantly overvalued.”
The company’s customer base for these new manufacturing processes encompasses industry leaders including Apple, Nvidia, AMD, and Google. TSMC maintains dominance in the global foundry sector with approximately 70% market share.
ASML verified the 1% decrease in its share price on the day TSMC made its announcements.
The post TSMC (TSM) Stock Surges 5% After Revealing New Chip Strategy Without ASML’s Expensive Equipment appeared first on Blockonomi.
