While Taiwan Semiconductor Manufacturing (TSM.US) shares have surged approximately 116% over the past year—an impressive figure in any sector—it has become a relative underperformer in the current AI chip rally when compared to Intel's over 500% gain and Micron's over 800% increase. What has caused the world's leading foundry to trail behind once-forgotten Intel and the cyclical memory stock Micron? What strategic logic and risks lie behind TSMC's "steady bull" trajectory?
Strong Fundamentals, Yet Not the Most Explosive Play
From a fundamental perspective, TSMC's performance is flawless. For the first quarter of 2026, TSMC reported consolidated revenue of approximately NT$1.13 trillion (about $359 billion), a year-over-year increase of 35.1%. Net profit after tax was around NT$572.5 billion, up 58.3%. Its gross margin reached a historic high of 66.2%, with an operating profit margin of 58.1% and a net profit margin of 50.5%. Revenue from High-Performance Computing (HPC) surpassed 60% for the first time, replacing smartphones as the largest revenue source. Chairman C.C. Wei has raised the full-year revenue growth target to "over 30%" and expects 2026 capital expenditure to approach the upper end of the $52-$56 billion range.
However, the market has been notably restrained in its valuation of TSMC. As of June 19, TSMC's price-to-earnings ratio is about 39.8x. Although this is above the industry average, compared to the explosive gains of Micron (up over 260% year-to-date) and Intel (also up over 260% year-to-date), TSMC is clearly not the most volatile play in this AI-driven market. The fundamental reason is that TSMC is the "shovel seller" rather than the "gold digger." Micron benefits from DRAM contract prices surging over 300% in the past five quarters, while Intel benefits from a market re-rating based on expectations for its 18A process and foundry business transformation. Although TSMC's foundry business is in high demand, its pricing strategy is more stable. While reports suggest TSMC is preparing another price hike of about 15% in the second half of the year, this is considered moderate compared to the "soaring" prices of memory chips.
Capacity Constraints and Order Shifts: Strategic Restraint or Inability to Keep Up?
The core dilemma TSMC faces is infinite demand against finite capacity. Reports indicate that due to AI infrastructure demand far exceeding TSMC's advanced process capacity, Samsung Electronics is receiving an increasing number of chip foundry orders from global clients like Google, AMD, and Tesla. TSMC has allocated the vast majority of its advanced node capacity to major clients such as Nvidia, Apple, AMD, Broadcom, Marvell Technology, and MediaTek, leaving little room for new clients or small-batch orders. A significant "supply chain shift" is underway. Google is reportedly in discussions with Samsung for the potential production of its next-generation Axion processor and is considering moving some of its TPU foundry orders from TSMC to Samsung as early as 2028. AMD is also reportedly discussing future CPU production with Samsung, and Nvidia's Groq LPU is already being manufactured by Samsung.
At a shareholder meeting in early June, Chairman Wei admitted that even with ongoing global capacity expansion, the company cannot meet AI-driven chip demand in the coming years. When asked about the peak of capital expenditure, he stated, "I don't know, but the next few years are very good. I don't see any indicators [of a stop] now." TSMC is not idle on expansion. Between 2025 and 2026, TSMC is simultaneously constructing and renovating 18 twelve-inch wafer fabs globally, with new capacity focused on 2nm and 3nm. However, building chip manufacturing infrastructure is a "time-consuming and capital-intensive process," making it difficult to fundamentally alleviate the capacity bottleneck in the short term.
TSMC's attitude towards the outflow of orders is intriguing. Analysis suggests management does not seem eager to "plug" every capacity gap through aggressive expansion or sharp price hikes, preferring instead to maintain long-term, stable customer relationships. This "strategic restraint" may mean ceding market share in the short term but could foster healthier customer trust and pricing power in the long run. TSMC can still employ strategic pricing to improve its profit margins.
Geopolitics: The Sword of Damocles Hanging Over TSMC
TSMC's concentration of capacity in Taiwan remains a structural risk that investors cannot ignore. On the geopolitical front, TSMC recently received a positive signal: the U.S. Department of Commerce has granted its Nanjing fab a 2026 annual export license, allowing it to receive U.S.-controlled equipment without separate applications. This "annual license" ensures the operation of TSMC's production lines in mainland China is not interrupted. However, long-term risks persist. The U.S. government has indicated plans for broader, higher-rate tariffs. TSMC's capacity in Arizona is still ramping up and cannot fully replace its Taiwan hub in the short term. Balancing the political pressure for "U.S. manufacturing" with the cost efficiency of "Taiwan capacity" will be a core strategic challenge for TSMC in the coming years.
Intel's Comeback: The Real Threat to TSMC?
Intel's stock has skyrocketed over 600% from a low of around $19 in mid-2025 to over $133 currently. The core logic behind this reversal is the market's renewed confidence in Intel's advanced process manufacturing capabilities. On June 16, 2026, Intel disclosed at the VLSI Symposium that its 18A-P process has officially entered the risk production phase. The 18A process utilizes gate-all-around transistors and backside power delivery, offering a 9% performance improvement at the same power. More critically, 18A has completed key stages including design tape-out, customer tape-outs, and internal product introduction.
Simultaneously, market rumors suggest Intel may reach a strategic partnership with Apple to design and manufacture chips in the U.S., and Intel is deepening its collaboration with UMC to include 3nm advanced processes. Analysts note that Qualcomm and AMD are the most likely clients to consider Samsung, while Apple and Broadcom are evaluating Intel's technology. However, Intel's foundry business remains in a "cash-burning" phase. In Q1 2026, Intel Foundry revenue totaled $5.4 billion, up 16% year-over-year, but external customer revenue was only $174 million, with the rest being internal chip sales. The foundry segment reported an operating loss of $2.4 billion for the quarter. On a GAAP basis, Intel still recorded a net loss of $3.7 billion in Q1.
TSMC's technological moat remains deep. Its 3nm process accounted for 25% of wafer sales in Q1, 5nm for 36%, and 7nm for 13%. Combined, advanced processes (7nm and below) represented 74% of wafer sales revenue. The chief analyst at SemiAnalysis points out that AI workloads are evolving towards complex agents, and CPUs are facing "extremely severe capacity shortages." In this context, TSMC's position as the world's most reliable advanced process supplier is difficult to challenge in the near term.
Valuation and Outlook: A 'Buy' Opportunity at 23x P/E?
From a valuation perspective, the market has not priced in excessive future growth for TSMC. As of June 19, TSMC's P/E ratio is about 39.8x. However, considering its expected 2026 revenue growth of over 30% and gross margins stable above 66%, its forward P/E is closer to the industry average. Previous analysis suggested TSMC's valuation remains in a reasonable range around 23x, very close to the semiconductor industry's average P/E of 21.5x. Chairman Wei expressed "considerable confidence" in visibility through 2030. TSMC expects capital expenditure over the next three years to be "significantly higher" than the past three years, with FY2028 capex potentially approaching $75 billion. Such a massive capital expenditure plan demonstrates management's confidence in long-term AI demand but also implies significant execution risk. If an unexpected economic downturn occurs or the AI investment frenzy subsides, high depreciation and amortization could significantly pressure profit margins.
Nevertheless, TSMC's stock price has decisively broken through the key resistance level of $390 without retreating, showing strong upward momentum. Several prominent Wall Street strategists acknowledged in a jointly released report that their earlier downgrade of TSMC in April "may have been premature." With the proliferation of custom chips by tech giants like Amazon and the intense competition between Google's AI cloud and Nvidia-backed Neoclouds, visibility for foundry capacity saturation is extremely high over the next 3 to 4 years.
Comments