Taiwan Semiconductor – How Can You Not Be Bullish Now? (Q1 Review)
Q1 Was Stunning. The Outlook Is Even Better!
Semiconductor giant TSMC, one of my top portfolio holdings, released its Q1 results exactly one week ago and delivered a stunning report.
Revenue growth remained blistering, comfortably beating both consensus and management’s own guidance. Margins hit all-time highs across the board, reflecting extraordinary pricing power and operational leverage. Management raised its full-year outlook and signaled that AI-driven demand shows absolutely no signs of slowing. And the company’s reinvestment metrics are sublime.
In short, TSMC continues to fire on all cylinders, and the outlook is as strong as ever. The company has proven its quality and confirms my belief that this is a sensational business to own in the decade ahead.
Today, I want to take a close look at TSMC’s Q1 results, breaking down the numbers and assessing developments in order to update my view of the company, thesis, and financial forecast.
Does my thesis hold up, and is now a good time to buy? Let’s find out!
This is a paid-exclusive analysis. Want full access? Consider upgrading InvestInsights Premium for $9.50 a month or $95 a year! This gets you:
6–8 Research Reports per month, 2x more than free subs do, including 3-4 Deep Dives and 2 mid-week analyses (that comes down to just $1 per analysis)
Full visibility into my personal portfolio, including allocations, transactions, and conviction levels
An overview of target prices and ratings, updated weekly
InvestInsights Premium is designed to help you find best-in-class compounders, identify deep-value opportunities, avoid costly mistakes, and stay focused on fundamentals when markets get noisy, with full transparency into how I make decisions over time.
Financial & Performance Review
Let’s jump straight into the numbers!
TSMC reported a Q1 revenue of $35.9 billion, beating the consensus by a comfortable $410 million, surpassing management’s guidance, and reflecting YoY growth of 41%, which is quite sensational for a business as dominant as TSMC and with a revenue run-rate exceeding $140 billion.
As visible below, TSMC has pretty much been delivering blinding growth ever since early 2024, which can be entirely attributed to the booming demand for leading-edge process technology, or simply the most advanced chip technology, entirely driven by the AI boom and the subsequent data center build-out.
This demand drove 40%+ revenue growth for TSMC in Q1, matching the pace we saw in most of 2025 and showing no signs of demand weakness. TSCM simply scales output, driving immediate growth as demand for its advanced chips continues to outpace supply.
This pace is likely to hold throughout 2026 as the supply shortage persists.
Impressively, with this growth, TSMC continues to outpace the market by some distance, despite already dominating it. As shown in the Counterpoint data below, TSMC absolutely dominates global semiconductor manufacturing, commanding a 72% market share and growing.
While the company captured just 63% of the market as of Q1 2024, this has steadily grown to 72% as of Q4 2025, which is nothing short of impressive. That is a 9-percentage-point gain in two years in an industry nearing $200 billion.
The key to TSMC’s success, and the enabling factor behind its 40%+ revenue growth today, is its technological dominance in advanced nodes.
You see, while TSMC controls “only” 72% of the entire foundry market, its dominance rises to over 95% (pretty much 100%) in advanced nodes, or those below 3nm. Here, the company effectively operates a monopoly, and this is where demand is booming.
The reason is straightforward: AI.
Training and running large AI models requires processors of extraordinary computational density and energy efficiency. You need to pack billions of transistors into a tiny space while keeping power consumption manageable, because at data center scale, electricity costs are existential.
That’s only achievable at sub-3nm nodes, which is exactly why NVIDIA’s Blackwell and upcoming Rubin, Google’s TPUs, Apple’s A-series and M-series chips, and AMD’s MI-series are all designed at 3nm or below. There is simply no alternative if you want competitive performance.
You can think of a transistor as a light switch. It turns electrical signals on and off, and that switching is what performs computation. A modern AI chip needs to perform trillions of these operations per second. Now imagine you’re building a city. You can either spread it across a vast countryside or build a dense Manhattan with skyscrapers. The denser your city, the shorter the distances between buildings, meaning people (signals) travel faster and use less energy getting from A to B.
So, smaller nodes mean lower energy use per signal (shorter distance to travel). Therefore, you simply don’t compete in performance if the node is above 3nm.
And TSMC is practically the sole producer at scale, which is why its 3nm capacity is entirely sold out and capacity-constrained. This explains why TSMC is far outpacing the industry – it dominates the vertical with the strongest demand by a mile.
Most importantly, this dominance in advanced nodes isn’t expected to fade either, which is hugely important for its outlook and customer commitments.
TSMC is already advancing beyond 3nm and had its 2nm node entering high-volume production in late 2025 with good yield, ramping according to plan. For reference, this new node delivers a 10-15% power benefit at the same power, a 20-30% power improvement at the same speed, and more than a 15% increase in chip density over the latest and most advanced 3nm node.
That is a high level of innovation and, crucially, the competition isn’t close.
TSMC’s 2nm yield is now stable at 60-70%, which is good. For reference, yield is the percentage of chips on a wafer that come out working correctly, which is a crucial measure in high-volume manufacturing.
Samsung also has a 2nm node, but it’s struggling with its yield, sitting at only around 55%, or 10 percentage points below TSMC, which is a huge gap, simply failing to reach levels required to compete for critical orders. Furthermore, when accounting for performance binning and losses in back-end packaging and testing, Samsung’s effective yield for final products drops to just 40%, indicating its 2nm process is far from competitive.
And that is the closest peer TSMC has. Intel has its 18A process, but yields will not reach commercially acceptable cost levels until the end of 2026, with industry-standard levels not expected until 2027. Again, TSMC started successful high-volume manufacturing in late 2025.
By current estimates, Samsung is still 2-3 years behind TSMC in terms of technology, and Intel by 3-5 years, which, at the pace of innovation in this industry, is light-years.
A competitor can announce a 2nm node any day. Getting to a 60–70% yield on it consistently at volume takes 2–3 years of painful iteration. Samsung made slight gains in 2024 and 2025, but its overall technology remains far behind TSMC. On top of that, TSMC’s $56 billion in 2026 CapEx is impossible to match, meaning the gap is more likely to expand than close.
The moat is real, it’s wide, and it’s widening.
At 3nm, TSMC continues to work hard to increase output to meet demand, executing a multi-year build-out plan. It has a 3nm fab in Taiwan coming online in the first half of 2027, a 3nm facility in Arizona with construction complete and volume ramping up in the second half of 2027, and a 3nm fab in Japan coming online in 2028.
And management already has capacity commitments for all of these, giving it a significant growth runway.
The company is also already working on its A14 node, adding another 10% to 15% speed improvement at the same power or 25% to 30% power improvement at the same speed, and close to 20% chip density gain over 2nm. Customer interest is huge, with Nvidia already securing capacity coming online in 2028.
Let’s then move to the P&L, which was similarly sensational.
