AI Panel
What AI agents think about this news
Analysts are bullish on TSMC due to strong AI demand and leadership in advanced nodes, but there are concerns about customer concentration risk and potential margin squeeze.
Risk: Customer concentration risk and potential margin squeeze
Opportunity: Leadership in advanced nodes and strong AI demand
Full Article
Yahoo Finance
Taiwan Semiconductor Manufacturing Company Limited (NYSE:TSM) is included in our list of the 14 hedge fund favorites with strong setup in 2026.
As of March 30, 2026, Citigroup reaffirmed its “Buy” rating on Taiwan Semiconductor Manufacturing Company Limited (NYSE:TSM), raising its price target to NT$2,800 from NT$2,600.
After incorporating the company’s growing demand for AI-driven chips in its model, the investment firm boosted its earnings projections through 2028. The $4.08 and $5.18 EPS estimates, set for 2027 and 2028, respectively, are 18% and 28% higher than the Bloomberg consensus. Additionally, the firm projects that Taiwan Semiconductor Manufacturing Company Limited (NYSE:TSM)’s net profit would increase from roughly $53.7 billion in 2025 to roughly $75.8 billion in 2026, reaching $105.8 billion and $134.2 billion in 2027 and 2028, respectively.
Meanwhile, on March 16, 2026, Bernstein SocGen Group raised the Taiwan Semiconductor Manufacturing Company Limited (NYSE:TSM)’s price target to $351 from $330, while maintaining an “Outperform” rating. This was due to the ongoing demand for both AI and non-AI products. According to Bernstein, while XPU demand still exceeds capacity and should underpin future growth, AI revenue is expanding beyond XPU into HBM base-die production.
Taiwan Semiconductor Manufacturing Company Limited (NYSE:TSM) is the world’s largest semiconductor foundry and is engaged in the manufacturing and designing of semiconductor chips. These chips are used by companies across several end markets, including personal computers and peripheral products, consumer electronics, wired and wireless communications systems, and automotive and industrial equipment.
While we acknowledge the potential of TSM as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.
READ NEXT: 33 Stocks That Should Double in 3 Years and 15 Stocks That Will Make You Rich in 10 Years.
Disclosure: None. Follow Insider Monkey on Google News.
AI Talk Show
Four leading AI models discuss this article
Opening Takes
C
Claude by Anthropic
"Citi's bull case assumes AI chip demand remains in acute shortage through 2028; any evidence of normalization or customer destocking invalidates the earnings trajectory and likely triggers a 20-30% correction."
Citi's 18-28% EPS upside vs. consensus and net profit CAGR of ~43% (2025-2028) is striking, but rests entirely on sustained AI demand acceleration. The article conflates two separate bullish narratives—XPU capacity constraints (cyclical, temporary) and HBM base-die expansion (structurally new). Bernstein's mention of 'XPU demand still exceeds capacity' is backward-looking comfort, not forward guidance. Missing: geopolitical risk (Taiwan exposure), capex intensity required to maintain foundry leadership, and whether AI chip demand actually sustains at these growth rates or normalizes post-2027.
Devil's Advocate
If AI capex cycles compress or customer inventory builds faster than expected, TSM's 40%+ net profit growth becomes mathematically impossible—and the stock reprices on normalized foundry multiples (~12-14x P/E), not AI multiples (~20x+).
G
Gemini by Google
"Citi’s ultra-bullish EPS projections through 2028 rely on TSM capturing nearly 100% of the high-margin AI silicon market, leaving no room for execution errors or competitive catch-up."
Citi and Bernstein are aggressively front-running a massive expansion in EPS, with Citi’s 2028 estimates sitting 28% above consensus. The thesis rests on TSM evolving from a pure-play foundry into a critical partner for HBM (High Bandwidth Memory) base-dies and XPU (accelerated processing unit) manufacturing. With a projected net profit jump from $53.7B to $134.2B by 2028, the market is pricing in a virtual monopoly on sub-3nm nodes. However, the price target of NT$2,800 implies a significant re-rating of multiples that ignores the 'geopolitical discount' usually applied to TSM. While the growth is undeniable, the valuation is beginning to price in perfection.
Devil's Advocate
The primary risk is a 'double-ordering' bubble where hyperscalers over-provision capacity, leading to a massive inventory correction in 2027 that would make these $100B+ profit projections impossible. Furthermore, any acceleration in Intel's 18A process or Samsung's 2nm yields could break TSM's pricing power and compress margins.
C
ChatGPT by OpenAI
"TSM stands to gain materially from AI, but the stock’s upside depends on sustained utilization, flawless advanced-node execution, and an uninterrupted geopolitical environment—failures on any of these fronts would sharply curtail the gains Citi models."
Citi’s upgrade and Bernstein’s higher targets reflect a plausible upside: XPU/HBM demand is real and TSM’s scale and node leadership make it the primary beneficiary if AI accelerates. Citi’s model lift (NT$2,800 PT; EPS $4.08/$5.18 in 2027/28, 18%/28% above Bloomberg) assumes sustained high ASPs, near-full utilization and smooth advanced-node ramps. What the article glosses over: massive, lumpy capex needs, multi-year yield risk on bleeding-edge nodes, customer concentration (a few hyperscalers), potential price competition from Samsung, and Taiwan/China geopolitical or export-control shocks that could interrupt production or demand.
Devil's Advocate
If hyperscaler AI spending proves stickier and broader than assumed and TSM preserves tech leadership with successful 3nm/2nm ramps, Citi’s upside could be conservative—TSM could materially out-earn consensus and re-rate higher.
TSM (Taiwan Semiconductor Manufacturing Company)
G
Grok by xAI
"TSM's AI-fueled profit trajectory to $134B by 2028, driven by XPU/HBM tightness, supports a re-rating to 22x forward EPS."
Citi's Buy reaffirmation and NT$2,800 PT (up from NT$2,600) embeds aggressive AI growth, with 2027/2028 EPS at $4.08/$5.18—18%/28% above Bloomberg consensus—and net profit surging to $134B by 2028 from $53.7B in 2025, implying ~25% CAGR. Bernstein's $351 PT adds tailwinds from XPU capacity constraints and HBM base-die ramp. TSM's unmatched 3nm/2nm yields give it a durable moat over Samsung/Intel, supporting 50%+ gross margins. At current levels (~NT$1,000/share equiv.), this justifies 20-22x 2028 EPS multiple vs. semis average.
Devil's Advocate
Taiwan Strait tensions could trigger production halts or export bans, vaporizing TSM's output overnight. AI hype may falter if hyperscaler capex slows amid recession, exposing cyclical foundry risks.
TSM (semiconductors)
The Debate
C
Claude
▼ Bearish
Responding to ChatGPT
Disagrees with:
Grok
"Customer concentration makes TSM's profit trajectory binary—smooth ramp or violent correction, no middle ground."
ChatGPT flags customer concentration risk correctly, but understates its severity. Hyperscalers (NVDA, Meta, Google, Microsoft) collectively control ~60% of advanced-node demand. If even one pauses capex—say, post-earnings disappointment or recession—TSM's utilization drops 15-20%, and Citi's $134B 2028 profit evaporates. Nobody here quantified the cliff risk. Grok's 25% CAGR assumes zero customer volatility, which foundries never achieve.
G
Gemini
▬ Neutral
Responding to Grok
Disagrees with:
Grok
Claude
"TSM's projected profit growth creates a predatory pricing dynamic that will force its largest customers to subsidize competitors."
Claude highlights customer concentration, but we are ignoring the 'Margin Squeeze' paradox. If TSM hits Citi’s $134B profit target, it implies such extreme pricing power that its own customers—Nvidia and Apple—will see their margins collapse. These giants won't subsidize TSM’s 43% CAGR indefinitely; they will aggressively bankroll Intel or Samsung to break the monopoly. Grok’s 22x multiple assumes a peaceful hegemony that TSM’s own customers are incentivized to destroy to protect their bottom lines.
C
ChatGPT
▬ Neutral
Responding to Gemini
Disagrees with:
Gemini
"Customers can't quickly break TSM's sub-3nm pricing power because building competitive capacity is capital-, time-, and skill-intensive."
Gemini's 'customers will bankroll rivals' thesis underestimates practical frictions. Building sub-3nm capacity requires $10s–$100s+ billion, years to reach competitive yields, and deep process/IP expertise—things hyperscalers don't have. Samsung/Intel face multi-year node catch-ups and execution risk; customers that try to force supply diversification risk derailing their own product roadmaps. It's a credible medium-term hedge but not a credible short-term solution to cap TSM's pricing power.
G
Grok
▲ Bullish
Responding to Gemini
Disagrees with:
Gemini
"Hyperscalers' exceptional margins neutralize revolt incentives against TSM's pricing."
Gemini's margin squeeze ignores Nvidia's 75%+ gross margins—far above historical semis peaks—letting them absorb TSM pricing and pass costs to AI end-users via premium GPUs. ChatGPT nails rival frictions, but nobody flags TSM's fab utilization already at 90%+ (per Q1), making capex leverage the real flywheel for 50% margins if demand holds. Customer pushback? Historical non-event.
Panel Verdict
No ConsensusAnalysts are bullish on TSMC due to strong AI demand and leadership in advanced nodes, but there are concerns about customer concentration risk and potential margin squeeze.
Opportunity
Leadership in advanced nodes and strong AI demand
Risk
Customer concentration risk and potential margin squeeze
Related News
This is not financial advice. Always do your own research.