TSMC's Green Signal
A Programming Note: I will take a personal day tomorrow, so there won’t be any new post tomorrow.
TSMC’s 4Q’25 call was quite reassuring for AI bulls. Many investors look at TSMC as an important barometer for assessing the cycle as TSMC is deeply incentivized to ensure they are not left holding the bag by spending tens of billions of dollars in capex right at the peak and then suffer through the underutilized fabs for the entire cycle. TSMC has spent ~$101 Billion cumulative capex in the last three years, and they were quite confident that the number will be “significantly higher” in the next three years. Considering they guided for $52 Billion to $56 Billion capex in 2026, it is not hard to see that indeed this number may be lot higher in the next three years.
Given the elevated risk for TSMC, TSMC management takes more of a 360-approach to receive feedback not from just its customers, but customers’ customers to gauge the health of their end markets. Clearly, whatever TSMC has seen was encouraging enough for them to increase capex and revenue outlook. From the call:
Our customers continue to provide us with a positive outlook. In addition, our customers' customers who are mainly the cloud service providers are also providing strong signals and reaching out directly to request the capacity to support their business. Thus, our conviction in the multiyear AI megatrend remains strong, and we believe the demand for semiconductor will continue to be very fundamental…To address the structural increase in the long-term market demand profile, TSMC works closely with our customer and our customers' customer to plan our capacity.
Based on our planning framework, we raised our forecast for the revenue growth from AI accelerator to approach a mid- to high 50s percent CAGR for the 5 years period from 2024 to 2029. Underpinned by our technology differentiation and broad customer base, we now expect our overall long-term revenue growth to approach 25% CAGR in U.S. dollar terms for the 5-years period starting from 2024.
The first question in the call was about what TSMC is hearing from its customers, and customers’ customers. While investors seem to be starting to express a tad bit skepticism on TSMC’s customers’ customers i.e. hyperscalers, TSMC management seems much more convinced about the ROI on such spending. Some excerpt from the Q&A:
…you essentially try to ask us, say, whether the AI demand is real or not. I'm also very nervous about it. You bet because we have to invest about USD 52 billion to USD 56 billion for the CapEx, right? If we didn't do it carefully, and that would be big disaster to TSMC for sure. So of course, I spend a lot of time in the last 3, 4 months talking to my customer and end customers' customer. I want to make sure that my customers demand are real. So I talked to those cloud service providers, all of them.
The answer is that I'm quite satisfied with the answer. Actually, they show me the evidence that the AI really help their business. So they grow their business successfully and healthy in their financial return.
So no doubt, I also asked specifically that what's application, right? I mean that's -- for one of the hyperscalers, they told me that, that helped their social media software. And so the customer continue to increase. So I believe that. And with our own experience in the AI application, we also help to our own fab to improve the productivity.
…so all in all, I believe in my point of view, the AI is real, not only real, it's starting to grow into our daily life. And we believe that is kind of -- we call it AI megatrend, we certainly would believe that. So you -- another question is can the semiconductor industry to be good for 3, 4, 5 years in a row, I'd tell you the truth, I don't know.
But I look at the AI, it looks like it's going to be like an endless, I mean, that for many years to come. No matter what, TSMC stick on the fundamental technology leadership, manufacturing excellence, and we work with customers to get their trust. And I think that fundamental thing position TSMC to be very good future growth, let me say that, 25% CAGR as we projected, and we used to be conservative.
One analyst asked about TSMC’s assumptions about the token growth behind their revenue CAGR outlook. TSMC’s CEO C.C. Wei gave an endearing response:
I also try to understand what is the tokens of growth. But my customers, their product improvement continue to increase…it’s a well-known from Hopper to Blackwell to Rubin, that almost double, triple their performance. So the one they can support the tokens of growth or the one they can continue to support the compute power is enormous.
And so I lose the track to be frank with you.
You and me both, Mr. Wei!
One point that I would highlight here is the fact that TSMC is upgrading their capex outlook for 2026 likely indicates they believe AI still has legs to continue the momentum and this is more of an affirmation of their outlook on AI in the medium term. From the call:
If you build a new fab, it takes 2 to 3 years to build a new fab. So even we start to spend the 52 billion to 56 billion, the contribution to this year almost none and to 2027, a little bit. So we actually are looking for 2028, 2029 supply.
Our headache right now, if I can call it a headache, is a demand and supply gap. We need to work hard to narrow the gap.
TSMC also reminded that silicon still remains the key bottleneck, not power:
Today, from my point of view, still the bottleneck is TSMC's wafer supply. Not the power consumption, not yet. So we also look at carefully. To answer your question, say that TSMC's wafer can support how much of the gigawatt, still not enough. They still have abundant of power supply in the U.S.
One particular dimension about power “bottleneck” narrative that I didn’t appreciate before is this “prisoner dilemma” dynamic highlighted by Jeremie Eliahou Ontiveros in a recent Stratechery interview:
“The way we put it is it’s sort of a prisoner’s dilemma. If everyone could coordinate, we could all find power in a more easy way, but because everyone is spamming the utilities all around the country, actually all around the world, it creates this vicious circle where because you have requests everywhere, you need to put requests elsewhere because you don’t know if you’re going to get one in this location, and so it just only gets worse. And so that’s why you’ve seen this crazy behavior where every state has seen tens of gigawatts of data center requests for every month, which is insane because if you aggregate the number or total number of data center requests close to a terawatt in the US alone, which for context, the peak load in the US is 750 gigawatts, so we’re saying we are requesting more than double that!
…they’re fake, and it’s just not possible, the US is not going to be able to handle all of that. But the reason they’re fake is that prisoner’s dilemma”
Don’t get me wrong; power can still prove to be a bottleneck, but the situation may be not as dire as many might have assumed.
TSMC’s commentary seemed also quite encouraging for semi cap equipment companies. The rising complexity in chip manufacturing in advanced nodes is a boon for semi cap equipment companies such as ASML, Applied Materials, Lam Research, and KLA Corp. etc. From the TSMC call:
“the cost of tools are becoming more expensive and process complexity is increasing. As a result, the CapEx dollar required to build 1,000 wafer per month capacity of N2 is substantially higher than 1,000 wafer per month capacity for N3. The CapEx per k cost for A14 will be even higher.”
The other data point that I thought was very encouraging is the confirmation from TSMC that Arizona’s fabs’ yield or defect density is now almost equal to its fabs in Taiwan.
All in all, TSMC not only provided a buoyant outlook for their own business in the short to medium term, it also largely assuaged investors that the “AI trade” is still very much on!
In addition to “Daily Dose” (yes, DAILY) like this, MBI Deep Dives publishes one Deep Dive on a publicly listed company every month. You can find all the 65 Deep Dives here
Current Portfolio:
Please note that these are NOT my recommendation to buy/sell these securities, but just disclosure from my end so that you can assess potential biases that I may have because of my own personal portfolio holdings. Always consider my write-up my personal investing journal and never forget my objectives, risk tolerance, and constraints may have no resemblance to yours.
My current portfolio is disclosed below: