January 17, 2021 - 6:36pm EST by
2021 2022
Price: 125.23 EPS 0.78 0.95
Shares Out. (in M): 5,078 P/E 0 0
Market Cap (in $M): 645,000 P/FCF 0 0
Net Debt (in $M): -16,000 EBIT 0 0
TEV (in $M): 630,000 TEV/EBIT 0 0

Sign up for free guest access to view investment idea with a 45 days delay.




We believe the shares of Taiwan Semiconductor Manufacturing Corporation (TSMC, ticker TSM) are an attractive investment. We see 32% upside by the EOY 2021. 


TSMC is a foundry company, they own and operate the fabs that take in raw silicon wafers and produce chips. TSMC’s customers are companies like AAPL, AMD, NVDA, XLNX, TXN that design chips and outsource their manufacturing to TSMC. The global foundry market has multiple players but TSMC’s positioning in the market is unique. 


TSMC has over the years evolved into the by far the  best leading-edge foundry in the world. Their track record in developing and ramping cutting edge production processes (nodes) is unparalleled. They will be the clear market leaders at the upcoming 5nm and 3nm nodes and their lead will expand over time. We believe the value of this leadership position will compound over time and lead to substantial appreciation in TSMC shares. 


We are in the early innings of the deployment of Machine Learning into the economy. We already see it’s effects on the internet with automated chatbots and uncannily persuasive recommendation engines. Over time the same underlying data processing/decision making technologies will arrive in many more areas of the economy in the form of smart factories, automated medical diagnoses, advanced facial recognition, autonomous vehicles and many other applications. 


The common thread between these applications is the importance of powerful and power-efficient semiconductors across every level of the data chain (from IoT devices to hyperscaler servers and everything in between). TSMC is the only company that has a demonstrated capability to consistently deliver cutting edge chips on a predictable and reliable schedule. As a result, all the established chipmakers, as well as every new entrant into the industry is dependent on TSMC to produce new generation products. 


Falling behind in nodes can be substantially damaging to a company’s prospects. For many years INTC was the node-leader in the industry but their internal issues and managerial missteps have resulted in them ceding the leadership to TSMC. As a result, AMD which was a perpetual little brother in the datacenter CPU market is now taking substantial share and is considered likely to become the industry leader in the near future. This market position transformation goes to highlight how critical leading-node manufacturing is in this industry.


TSMC’s value chain position as the clear leader in cutting edge chips means their business model is more attractive than that of their customers and upstream suppliers. Their customers are in an extremely competitive, fast changing market. Using Datacenter GPUs as an example, NVDA is the current leader but AMD and INTC are chasing them. Moreover, startups like SambaNova, Graphcore and Cerebras are also going after share. NVDA’s hyperscaler customers like AMZN and GOOG are also coming up with their own chips. NVDA is well positioned but is going to have to fight. TSMC, on the other hand, is going to be the one producing the chips for whoever the winner is, the competition is good for them as it increases both demand and customer diversity.


TSMC’s suppliers are (with the exception of ASML) competitors in 2-3 player markets for the various types of machines required for fab production. Semiconductor capital equipment is also a well structured market but TSMC has enough alternatives to ensure that their suppliers don’t overly eat into TSMC’s pie. 


TSMC’s competitors at the leading edge are really only Samsung and INTC. However, neither of them are really competitors in that they aren’t purpose designed foundry companies. Both of their fabrication divisions are internally focused (Samsung is trying to change this, INTC’s strategy is up in the air). Samsung is the only other company that even offers near-leading edge fabs to third parties, but they only have two large scale third party customers (QCOM and NVDA) and have a long way to go to demonstrate that they can be a viable sole-source foundry for a company with market-leader ambitions. 


The lead TSMC has established in 5nm and 3nm is only going to increase over time. New nodes become substantially harder (the scientific problem of jamming more and more transistors into smaller spaces is even harder than it sounds). Catching up vs an established leader requires a significant screwup by the leader. TSMC is laser focused on maintaining their leadership position and being the best foundry in the world, INTC and Samsung have much more complex business models and competing priorities. We are confident that TSMC’s leadership position will be maintained and expanded over time. 




Having established TSMC as being the leader in leading edge manufacturing, what does that mean for their numbers?


High end manufacturing is very expensive, TSMC’s capex burden is substantial, but it comes with well above market topline and cash flow growth. We think TSMC will grow their topline by north of 15% per year for the next 5 years+ with the blockbuster 3nm node combining with the increasing semiconductor penetration of the economy. If you want to break this down, semi demand will be growing 8-10% over the next few years, this includes more GDP-levered analog semis. High end digital semis will see demand growth well above 20% per year, sales to this kind of customer make up 40% plus of TSMC’s revenue resulting in an overall mid to high teens topline growth rate.


Due to TSMC’s competitive position they will get meaningful pricing on new nodes which will drive Gross Margin expansion and they will get some operating leverage on R&D/SG&A. 


Our Numbers are conservative in that we don’t assume a an accretive use of Free Cash Flow. TSMC will convert net income to FCF at 70%+ through cycle so there will be ample opportunity to return capital to shareholders and further enhance this return picture. We see TSMC generating 18% of its current market cap in FCF over this forecast period. 


I believe TSMC is maybe the highest quality long dated asset in the market. Ther competitive advantage gets more pronounced over time due to the extremely difficult engineering challenge that they address. As such even in the out years TSMC’s outlook will still be mid to high teens EPS growth, high 50% Gross margins, and an ever increasing portion of tech industry value creation accreting to the leader in high-end semiconductor manufacturing. As this gets proven out I believe the stock will still trade at 30x P/E in 2025. This results in the below upside on an EO 2021 basis. This is all on the ADR. 



I don’t see TSMC trading below $100/share, the market is starting to understand both how strong demand for semis is going to be and how unparalleled an asset TSMC is. Anything close to 20x 2022 EPS should be bought aggressively. 


I see TSMC as a buy and hold asset, it will compound over time, this story only gets better as technology marches on. 




Competition - Samsung is chasing them, INTC could try to get their mojo back. INTC is not a foundry so even if they get back to the lead it only matters for AMD and maybe NVDA of TSMC’s customers, there will still be massive demand for wafers from TSMC. Samsung is a more “real” threat with their strategy targeting the GAAFET transition as an opening for them to take a leadership position. This will have to be carefully watched but TSMC’s manufacturing prowess is unmatched and the burden of proof here is on Samsung and INTC. 


Also demand is so high for high end semis that even if one or both make some strides TSMC will be able to deliver on something close to the earnings picture I show above. 


China - TSMC is a Taiwanese company, and the bulk of their assets sit on the island. If China were to attack/repossess Taiwan the shares would likely become worthless. This is a major risk, and one that his hard to handicap. In order to hedge it i would recommend owning a small position in the semicap names. If TSMC gets literally destroyed the wave of spending on fabs around the world as every country realizes that semis are strategic assets would result in an unbelievable bonanza for semicap names. 


Operational - Like any manufacturing company, operations are a risk. TSMC needs to maintain its exact engineering and production standards in a field that gets progressively more difficult.


*I have no position in TSMC in any capacity personal or professional

*I have no non-public information about TSMC and there is no non-public information in this note


*Nothing in this note should be construed as investment advice or a recommendation to do anything at all with your money or money that you are in control of


I do not hold a position with the issuer such as employment, directorship, or consultancy.
I and/or others I advise do not hold a material investment in the issuer's securities.


Increasing evidence of the extremely high level of demand at 5nm/3nm nodes (recent capex increase was a good example of this). There will likely be more capex increases and generally more announcements of chips at those nodes from the semiconductor design companies. 

Strong earnings results as TSMC's price and operating leverage continue to drive a strong operating performance.  

    show   sort by    
      Back to top