16 April 2026

TSMC Raises 2026 Revenue Growth to Above 30% on Surging AI Demand, Unveils Aggressive 3nm Capacity Expansion

Q1 2026 Earnings Call, April 16, 2026

Taiwan Semiconductor Manufacturing Company delivered first quarter results that exceeded guidance and raised its full-year revenue growth outlook to above 30% from previous expectations, driven by what management characterized as "extremely robust" demand for AI chips that continues to outpace the company's ability to add capacity. The world's largest contract chipmaker also unveiled plans to build three new 3-nanometer fabrication facilities globally and guided capital expenditures to the high end of its $52 billion to $56 billion range for 2026.

The company's first quarter revenue reached $35.9 billion, up 6.4% sequentially, while gross margin expanded 390 basis points to 66.2%, surpassing the high end of guidance by 120 basis points. More significantly, TSMC now expects second quarter revenue of $39 billion to $40.2 billion at the midpoint, representing 32% year-over-year growth, with gross margins holding near 66.5%.

Unprecedented AI Demand Drives Historic Capacity Expansion

In a departure from historical practice, TSMC is adding substantial new capacity to its 3-nanometer node even after it has reached what would normally be considered full build-out. CEO C.C. Wei announced the company is executing a global expansion plan that includes a new 3nm fab in Taiwan's Tainan coming online in the first half of 2027, a second Arizona facility using 3nm technology starting volume production in the second half of 2027, and a second Japan fab utilizing 3nm scheduled for 2028. The company is also converting existing 5nm tools to support 3nm capacity.

Wei emphasized the demand drivers remain concentrated in high-performance computing and AI applications. "The shift from generative AI and chat mode to agentic AI and command and action mode is leading to another step-up in the amount of tokens being consumed," Wei explained. "This is driving the need for more and more computation, which supports the robust demand for AI silicon." The company noted it has received "very strong signals and positive outlook" from cloud service providers, who are the ultimate customers driving semiconductor demand.

When pressed on how long supply constraints might persist, Wei was candid about the timeline: "It takes 2 to 3 years to build a new fab... we expect this to continue to be very tight." He indicated that even with aggressive expansion plans, supply will likely remain constrained through 2027.

Capital Intensity to Remain Elevated But Controlled

CFO Wendell Huang revealed that TSMC now expects 2026 capital expenditures toward the high end of the $52 billion to $56 billion range, representing potential spending of $56 billion this year alone. That figure exceeds 50% of the $101 billion the company spent over the entire 2023-2025 period. Looking ahead, Huang stated that CapEx "in the next 3 years will be significantly higher than the past few years."

However, management sought to reassure investors that capital intensity would remain under control. Huang noted that "if we do our job right, then we will continue to see revenue growth outpace CapEx growth" in coming years, suggesting the company does "not expect in the next several years a sudden surge in capital intensity." The implication is that while absolute CapEx dollars will rise substantially, revenue growth should keep pace or exceed it, maintaining capital intensity in a reasonable range rather than seeing it spike dramatically higher.

3nm Economics Reach Inflection Point

A significant milestone highlighted during the call was the expected crossover of 3-nanometer gross margins to corporate average levels in the second half of 2026. Huang confirmed this development, noting that after full depreciation as with previous nodes, "gross margins are generally very high." Given that TSMC's corporate gross margin now stands at 66.2% and is guided to 66.5% for the second quarter, this crossover represents a substantial profitability achievement for a node that only entered volume production in late 2022.

The 3nm technology now represents 25% of wafer revenue, with strong contributions from both smartphone and HPC applications. Management emphasized that capacity expansion is being driven by HPC and AI demand specifically, with Wei stating plainly when asked what applications justify the unprecedented capacity additions: "The answer is simple, is still the HPC AI applications."

Margin Structure Holds Despite 2nm Ramp and Overseas Dilution

TSMC reiterated that its 2-nanometer technology, which entered high volume manufacturing in the fourth quarter of 2025, will dilute gross margins by 2% to 3% for the full year 2026. Additionally, the company continues to forecast 2% to 3% dilution from overseas fab ramp-up in early stages, widening to 3% to 4% in latter stages. Management also flagged potential cost pressures from the Middle East situation, noting that "prices for certain chemicals and gases are likely to increase," though it was too early to quantify the impact.

Despite these headwinds, second quarter gross margin guidance of 65.5% to 67.5% suggests the company is successfully offsetting dilution through higher utilization rates, cost improvement efforts, and the improving economics of 3nm. TSMC maintained its long-term margin targets of 56% and higher gross margins through the cycle and ROE in the high 20% range through the cycle, targets that were already raised from previous guidance.

Competitive Response: No Shortcuts in Foundry Business

When questioned about competitive threats from Intel's foundry ambitions and initiatives like the terafab announcement, Wei offered a measured but confident assessment. "Both Intel and Tesla, they are TSMC's customers," he noted, adding that "we view Intel as a formidable competitor and do not underestimate them." However, he emphasized that "there are no shortcuts, the fundamental rules of the foundry game never change. They need technology leadership, manufacturing excellence and customer trust and most of all, the service."

Wei reinforced the time horizons involved: "It takes 2 to 3 years to build a new fab. No shortcuts. And it takes another 1 to 2 years to ramp it up. Again, that's a fundamental of foundry industry." Regarding concerns about losing customers due to capacity constraints, Wei emphasized that TSMC views customers as partners and works to ensure their success rather than pursuing aggressive pricing to maximize short-term margins.

Advanced Packaging Capacity Equally Constrained

Beyond front-end wafer capacity, TSMC revealed that advanced packaging capacity is also extremely tight. Wei stated, "our advanced packaging capacity is very tight also. So we have to work with our OSAT partners we hope that we can increase the capacity to support our customers." The company is currently supplying what it describes as "the largest reticle size packaging" in the industry and is developing even larger solutions, though management declined to provide specific details on panel-level packaging timelines.

When asked about competitive threats from alternative packaging approaches, Wei acknowledged that competitors offer attractive technologies but maintained confidence in TSMC's roadmap. The company is working on addressing technical challenges such as warpage and mechanical stress in larger die sizes, with Wei stating, "A good challenge, and we like it, the harder the better because of TSMC in technical engineering, and we have confidence that we can work with our customers to solve all the issues."

A16 Development on Track for 2028

Looking further ahead, Wei provided an update on the company's A16 node, its second-generation transistor structure that represents another four-node stride from N2. The technology will deliver 10% to 15% speed improvement at the same power or 25% to 30% power improvement at the same speed, along with close to 20% chip density gains compared to N2. "Our A16 technology development is on track and progressing well," Wei stated, with volume production scheduled for 2028 and "high level of customer interest and engagement from both smartphone and HPC applications."

Material Supply and Energy Security

Addressing concerns about the Middle East situation's impact on supply chains, Huang detailed TSMC's risk mitigation strategies. For specialty chemicals and gases including helium and hydrogen, the company sources from multiple suppliers across different regions and maintains safety stock inventory. "We do not expect any near-term impact on our operations for material supply," Huang stated.

On energy supply for Taiwan operations, the company noted it works closely with Taiwan Power and the government to ensure stability. The Taiwan government has secured sufficient LNG supply through at least May and is working on diversification and backup plans. "We do not expect any near-term disruption or impact to our operations," Huang confirmed.

Mature Node Strategy: Focus on Specialized Technologies

Wei clarified TSMC's approach to mature nodes, emphasizing the company is building "high-yield capacity for specialized technologies rather than just normal capacity." Examples include capacity in Japan for CMOS image sensors and ESSMC in Germany for automotive and industrial applications. Notably, the company plans to wind down its 6-inch Fab 2 and 8-inch Fab 5, using available space to optimize support for leading-edge applications. "Even without fab 2 and fab 5, we still have enough capacity to fully support our existing customers," Wei stated, signaling a clear strategic shift toward higher value-added segments.

Platform Mix Reflects AI Dominance

First quarter revenue mix showed HPC increasing 20% quarter-over-quarter to account for 61% of revenue, up from previous quarters and reflecting the AI surge. Smartphone decreased 11% to 26% of revenue, while IoT increased 12% to 6%, automotive decreased 7% to 4%, and DCE increased 28% to just 1%. The HPC platform's dominance underscores the concentration of growth in AI-related applications and helps explain management's conviction about the sustainability of current demand levels.

PRO Investment deep dive

Taiwan Semiconductor Manufacturing Company Deep Dive

The semiconductor industry has entered the era of extreme computational density, and Taiwan Semiconductor Manufacturing Company remains the undisputed pivot point for the global digital economy. As of April 2026, TSMC’s technical and operational hegemony has reached a level of maturity that arguably transcends traditional cyclical analysis. By successfully transitioning to 2-nanometer production using gate-all-around transistor architecture, the company has effectively widened the moat separating its capabilities from those of its closest peers. While Samsung Foundry and Intel Foundry continue to iterate their roadmaps, they operate at a distinct disadvantage in both yield stability and customer trust, the latter of which is the most critical currency in the modern foundry business.

The Architecture of Dominance

TSMC’s competitive advantage is no longer just about lithography and node shrinking; it is about the integration of complex ecosystems. The move to 2-nanometer mass production is significant, but the real structural strength lies in the company’s absolute control over advanced packaging. Technologies such as Chip-on-Wafer-on-Substrate are now the primary bottlenecks in the AI hardware supply chain. By aligning its wafer fabrication capacity with massive, dedicated packaging volumes, TSMC has essentially become a "one-stop shop" that is impossible for high-performance computing customers to circumvent. This vertical integration of assembly, testing, and production provides a level of certainty that competitors cannot replicate, effectively locking in the most important customers in the AI and high-end smartphone sectors for multiple product generations.

Furthermore, TSMC’s pure-play business model acts as a powerful economic flywheel. Unlike Samsung or Intel, which are burdened by their own chip design ambitions and the inherent conflicts of interest that create, TSMC remains a neutral partner. This neutrality, combined with their history of high-volume, reliable execution, means that leading-edge design houses view TSMC not as a supplier but as an extension of their own R&D departments. This relationship creates a self-reinforcing cycle of data sharing, process refinement, and yield optimization that allows TSMC to iterate faster than any potential entrant. Even with significant capital expenditures projected to stay elevated in the $50 billion-plus range annually, the company’s ability to command premium pricing on its most advanced nodes ensures that its profitability metrics remain well above those of its peers, even as they attempt to catch up.

The Competitive Landscape and Geopolitical Realities

The foundry race has essentially settled into a "one dominant player and several struggling competitors" structure. Intel’s foundry service ambition, while necessary for the company’s survival, remains a high-risk, unproven pivot. The technical hurdle of establishing process parity with TSMC is compounded by the commercial hurdle of winning external customers who are often direct competitors to Intel’s own design business. Similarly, Samsung Foundry has demonstrated resilience but continues to struggle with yield consistency, which leaves them relegated to second-source status for critical high-performance designs. For now, there are no credible threats from new entrants that possess the scale, intellectual property, or institutional knowledge to challenge the TSMC stronghold at the leading edge.

The most pertinent threat is not technological; it is geographic. The concentration of TSMC’s most sophisticated facilities in Taiwan creates an asymmetric risk profile that is impossible to hedge fully. While the company is accelerating its footprint in the United States, Japan, and Germany, the overwhelming majority of its "crown jewel" production—specifically the 3nm and 2nm nodes—remains tethered to its home island. This geographic dependency is the primary systemic risk to the bull case. While the company has demonstrated remarkable management track records in navigating geopolitical tensions, the reality is that its entire operational infrastructure rests on a foundation that remains vulnerable to external political shocks. Investors must weigh the company’s undeniable technical lead and financial momentum against this permanent, underlying geopolitical volatility.

Opportunities for Future Growth

The company’s roadmap beyond 2nm provides a clear path for sustained revenue expansion. The introduction of 1.6nm, which will leverage backside power delivery to optimize energy efficiency, is already being integrated into the AI chip design cycles of the next two years. The push toward 1.4nm (A14) for 2028 reflects an industrial-scale commitment to maintaining this trajectory. These future nodes are designed specifically for the increasing power requirements of generative AI, where thermal management and energy efficiency are no longer optional features but are critical for the economic viability of training and inference clusters. If TSMC continues to deliver on these node transitions, it will successfully transition from being a simple contract manufacturer into a fundamental bottleneck in the global AI hardware pipeline, cementing its role in every significant compute cycle for the next decade.

The Scorecard

TSMC’s fundamental position is exceptionally strong, characterized by clear technological leadership, unmatched customer loyalty, and a business model that creates a virtuous cycle of margin expansion as new nodes reach maturity. The company has demonstrated a consistent ability to manage capital allocation efficiently while scaling complex manufacturing processes at record speeds. With AI demand driving a multi-year growth trajectory, the company is effectively printing cash while its competitors fight for the scraps of the market. The high gross margins and the ability to maintain pricing power, even as capacity expands, underscore the deep moat protecting its business.

The primary risk remains the structural reality of geopolitical exposure, which persists as a permanent discount on the company’s potential. Execution risks are notably low, but the industry-wide transition to increasingly complex architectures like backside power delivery and 3D stacking does create a non-zero risk of a major operational stumble. While these risks are present, they are overshadowed by the reality that the global tech economy has no viable alternative to TSMC. For the long-term, the company remains the foundational bedrock upon which the next decade of digital advancement will be built, provided the geopolitical status quo holds.