By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
TSMC's Q2 results and increased capex signal strong AI/HPC demand, but geopolitical risks, customer concentration, and potential margin pressures from Arizona fab ramp and China export curbs pose significant challenges.
Risk: Geopolitical risks, particularly the 'Arizona Trap' and potential customer fab in-house production, could lead to margin compression and utilization issues.
Opportunity: Sustained AI/HPC demand and leading-edge foundry capacity near-monopoly could drive continued growth and high gross margins.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Demand for artificial intelligence (AI)-related semiconductor chips is booming, and Taiwan Semiconductor Manufacturing (NYSE: TSM), which has a near-monopoly over their production, continues to see its revenue surge. And with the stock trading at a forward P/E of around 20 times next year's earnings, now looks like a good time to scoop up some shares.
Insatiable AI chip demand continues
As demand for its chip manufacturing services soars, TSMC once again reported climbing revenue and an expanding gross margin when it reported its second-quarter results last week. Revenue jumped 34% year over year to $40.7 billion, while its gross margin came in at 67.6%, up 910 basis points from 58.6% a year ago. The company has continually warned investors that its gross margin would be negatively impacted by moving to more advanced nodes and increased manufacturing in the U.S., but it guided for gross margin to remain robust in Q3, between 65% and 67%.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
The combination of revenue growth and gross margin expansion led to its profits surging 77% in local currency.
TSMC generated revenue from 2nm nodes for the first time, accounting for 3% of its total revenue. Meanwhile, 3nm revenue jumped to 30% of total revenue, up from 24% a year ago. Revenue from advanced technology, which includes 7nm nodes or below, accounted for 77% of its revenue. High-performance computing was 66% of its revenue, while smartphones accounted for 22%.
TSMC projected Q3 revenue between $44.6 billion and $45.8 billion, representing about 37% year-over-year growth at the midpoint. For the full year, it projects revenue growth of more than 40%, up from its prior guidance of 30%.
Perhaps more importantly than its strong results and guidance, TSMC also upped its 2026 capital expenditure (capex) budget, taking it to a range of $60 billion to $64 billion, up from a prior outlook of $52 billion to $56 billion. Between 70% and 80% will go toward advanced process technologies, with 10% to 20% toward advanced packaging and 10% for specialty technologies.
Why TSMC stock is a buy
TSMC is at the center of the AI infrastructure build-out, and its increased capex plans are a strong indication that demand is here to stay. Given the nature of the chip manufacturing business, where underutilized fabs become big earnings drains, management tends to perform deep due diligence, so this isn't only a good sign for TSMC, but for the AI infrastructure industry as a whole.
Given its attractive valuation, monopolistic positioning, and strong growth prospects, TSMC looks like a great AI stock to own at these levels.
Before you buy stock in Taiwan Semiconductor Manufacturing, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Taiwan Semiconductor Manufacturing wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $371,842! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,244,783!
Now, it’s worth noting Stock Advisor’s total average return is 900% — a market-crushing outperformance compared to 207% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of July 19, 2026. *
Geoffrey Seiler has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"TSMC's raised capex and margin resilience confirm multi-year AI tailwinds that justify a re-rating above 20x forward P/E."
TSMC's Q2 revenue surged 34% YoY to $40.7B with gross margins expanding to 67.6% and 3nm now at 30% of sales; full-year guidance raised to >40% growth. The $60-64B 2026 capex increase, heavily weighted to advanced nodes and CoWoS packaging, signals sustained AI/HPC demand (66% of revenue). At ~20x forward P/E against 30-35% expected EPS growth, the valuation appears reasonable for a near-monopoly on leading-edge foundry capacity. However, the article glosses over geopolitical risk, customer concentration (Apple + Nvidia ~40%+), and potential margin pressure from Arizona fab ramp and China export curbs.
Geopolitical escalation around Taiwan or stricter U.S. export controls on advanced nodes could freeze capex ROI, while hyperscalers may slow AI buildout if ROI disappoints, leaving TSMC with underutilized 2nm/3nm capacity and compressed margins.
"TSMC's current valuation assumes a perpetual growth cycle that ignores the extreme sensitivity of its high-fixed-cost model to potential AI infrastructure spending fatigue."
TSMC’s 67.6% gross margin is an outlier in a capital-intensive industry, reflecting unprecedented pricing power as the sole foundry for high-end AI silicon. While the forward P/E of 20x seems reasonable against 40% revenue growth, investors are ignoring the geopolitical risk premium inherent in the Taiwan Strait. The massive $60B+ capex hike is a double-edged sword; it secures market dominance but creates immense depreciation pressure if AI demand hits a cyclical plateau. We are seeing a 'gold rush' mentality where TSMC is the shovel seller, but any cooling in hyperscaler capex will lead to immediate margin compression due to the high fixed-cost nature of these advanced fabs.
If AI demand is truly as structural and 'insatiable' as management claims, the current valuation is actually a deep value play, and geopolitical fears are merely a persistent discount that long-term investors should exploit.
"TSMC's valuation assumes AI capex cycles remain robust and customer concentration risk doesn't materialize, but the article provides no margin-of-safety analysis for either scenario."
TSMC's 34% YoY revenue growth and 910bps gross margin expansion are real, but the article conflates near-term AI euphoria with durable competitive moat. The 20x forward P/E isn't cheap—it's fair for a 40% grower, not a bargain. More concerning: the article ignores geopolitical risk (Taiwan exposure, U.S. fab underutilization), customer concentration (likely 50%+ from hyperscalers), and the capex surge itself. $60-64B annual capex is massive; if AI capex cycles compress or customers build in-house fabs (Samsung, Intel), TSMC's utilization craters and those fabs become the 'earnings drains' the article warns about.
If AI demand truly is structural and TSMC maintains pricing power while scaling U.S. capacity, the capex is a feature not a bug—and 20x on 40% growth could re-rate higher, not lower.
"Even with strong AI-driven demand, high capex and geopolitical/regulatory risks could erode margins, making a 20x forward P/E potentially too rich."
TSMC is framed as an AI infrastructure bottleneck with revenue up 34% YoY to $40.7B and gross margin near 66%, trading around 20x forward earnings. The piece leans bullish on higher capex (to $60–$64B in 2026) as a secular demand driver, and highlights a favorable mix (3nm+ contributing ~30% and 77% from advanced tech). Yet risks loom: capex intensity can erode ROIC if utilization isn’t sustained, AI demand may be cyclical, and geopolitical/regulatory headwinds (Taiwan risk, export controls to China) could cap upside. Margin resilience hinges on continued fab utilization and favorable USD/NTD dynamics. Valuation assumes uninterrupted AI spend and ROIC expansion; if those fail, risk-reward tightens.
The strongest counterpoint is that capex-driven margins may compress as fabs run near capacity and AI demand proves episodic; plus regulatory/geopolitical shocks could cap revenue growth, making a 20x forward multiple potentially too rich.
"Packaging bottlenecks, not just wafer utilization, represent the sharper near-term constraint on TSMC's AI upside."
Claude's customer concentration estimate at 50%+ from hyperscalers is overstated; Apple alone was ~25% in 2024 while Nvidia ~15%. The real unmentioned risk is CoWoS packaging capacity constraints—TSMC's 2026 capex heavily bets here, yet lead times already exceed 12 months. Any AI accelerator redesign delaying packaging could create a 2027 revenue cliff despite wafer growth.
"The shift toward U.S.-based manufacturing will structurally compress TSMC's industry-leading margins regardless of AI demand."
Grok, you are right about the CoWoS bottleneck, but the real risk is the 'Arizona Trap.' Building in the U.S. isn't just about geopolitical hedging; it’s a structural margin dilutive move. With labor costs and operational inefficiencies, the U.S. fabs will likely underperform Taiwan-based margins for years. If TSMC is forced to shift more production stateside to appease D.C., their 67% gross margin is not sustainable, regardless of how much AI demand exists.
"Arizona's margin drag is real but manageable if TSMC stratifies production by node maturity rather than shifting cutting-edge volume."
Gemini's 'Arizona Trap' framing is sharper than I initially credited. But the margin dilution math needs specificity: Arizona fabs produce mature nodes (28nm+), not 3nm. If TSMC ring-fences advanced production in Taiwan while shifting commodity volume stateside, gross margin compression may be 200-300bps, not catastrophic. The real question: does U.S. geopolitical insurance justify that trade-off? Nobody's quantified the probability-weighted cost of Taiwan disruption.
"CoWoS packaging bottlenecks and long lead times risk delaying TSMC's 2026 capacity-to-revenue translation, potentially squeezing margins and muting the expected ramp."
Grok's 2027 cliff risk hinges on faster-than-anticipated packaging demand; the more immediate choke is CoWoS packaging capacity with >12-month lead times. If the 2026 capex is largely consumed by advanced wafers but fails to scale CoWoS quickly, utilization and ROIC could lag wafer growth, pressuring margins even before a full AI demand slowdown. In short, packaging bottlenecks may skew the ramp more than wafer capacity alone.
TSMC's Q2 results and increased capex signal strong AI/HPC demand, but geopolitical risks, customer concentration, and potential margin pressures from Arizona fab ramp and China export curbs pose significant challenges.
Sustained AI/HPC demand and leading-edge foundry capacity near-monopoly could drive continued growth and high gross margins.
Geopolitical risks, particularly the 'Arizona Trap' and potential customer fab in-house production, could lead to margin compression and utilization issues.