TSMC – Surfing USA
January 2023 (ADR) $96
Post our published piece on TSMC last year we have been researching the company in greater detail. As a result, we share with clients numerous reference pieces we think you should be reading to better understand the company. The last page of this note includes links to many of these. What follows is some reflection post this further study.
Mental models – The secret weapon of those of us with more moderate IQ’s
We have been to many a Berkshire AGM and listened to Charlie Munger talk about mental models. We have also followed his books and writings on the subject closely over the years. Indeed, when your author is asked to say what differentiates Holland Advisors analytical approach from others, he answers as follows. “Most other analysts look at investments by sector or geography, we look at them by mental model or business model.” We seek out those models we have seen successfully used before in maybe a new marketplace. Munger once spoke of the fun you can have with mental models using them to outthink those with higher IQ’s. This to us is clearly an appealing prospect. Indeed, we think it likely that almost everyone that has ever penned a research piece on TSMC will have a higher IQ than your author. As such we are about to try and put Munger’s idea to the test!
We have now spent a good amount of time on TSMC. The more pages we turn, the more we like it. Numerous mental models come to mind as we study it. The first is one we heard Charlie describe when he was once discussing Google. He talked of “surfers and wallowers” i.e. if a business is surfing in front of a wave (having found some form or advantage; scale, technological or otherwise), no competitor franticly paddling behind it is going to be able to catch up. In large cap tech there are plenty of examples of such situations: Google’s search engine, Microsoft’s Windows, Apple IOS etc. Most such business are priced by markets reflecting these advantages however. Investors, we think should pay attention when they are not priced as such. Today TSMC looks a good example of a business that is surfing further and further away from competitors, but it is not priced for the likely outcome that will result.
We conclude that TSMC has:
- A growing scale and process advantage
- Improved competitive tailwinds
- Regulatory assistance
A growing scale Lollapalooza
In our first piece we noted that we saw TSMC as a rare example of one of Hamilton Helmer’s 7 Powers, i.e. a process power company. The more we have studied this part of TSMC’s service offering the greater our conviction has become. It is clear listening to TSMC management past and present, that its advantages are not single source (i.e. a brand or lowest units cost). Instead, they are multi-faceted. Some being tangible (availability of newest technology, the ability to afford it and R+D dollars).
Many are however intangible. Examples of this would include long standing customer trust and the ability to constantly attract the brightest minds in the sector to make them want to work for TSMC. Also a designed system of continuous improvement so such staff feel free to create and challenged to stay at the leading edge of their industry technology’s advancement. Each of these factors in isolation might not make TSMC such a formidable competitor. It is their combined effects that really create a powerful Sustainable Competitive Advantage. Lollapalooza.
A mixture of these tangible and intangible factors can be seen in the company R+D expense. At c.7-8% of revenues this clearly rises in absolute dollars as the group’s revenue grows. However this only explains a part of the companies R+D advantage. As every advanced chip designed in the world (ex-Intel) pretty much uses TSMC for manufacturing, its client base has almost all of the best and brightest ideas/minds in semiconductors. That TSMC engages in much R+D with clients allows it to deliver great win-win scenarios to customers, whilst staying at the forefront of chip manufacturing innovation. Client chip ideas can be worked on and helped funded by the supplier they trust the most. As we observe below the same collaborative thinking and development is not today available to any Chinese would be competitors. Seemingly not to Intel either.
Improved competitor tailwinds
Another mental model we have learnt from our Buffett studies is the pricing power that can accrue to a business if it is a survivor in a market that consolidates from say five players into three, or three players into two. Indeed, this was the thinking that lay behind much of Buffett’s investments in newspapers during the 1970’s. If there were two highly competitive newspapers in a city and one went under/was consolidated, the remaining one was a money printing machine with far greater long term pricing power over advertisers. This same model we have used in our analysis of Sports Direct in the UK (with JJB going bust, only JD Sports remained), hence why Sports Direct/Frasers can now take pricing up… #“where else are they going to go”. This thesis has also played out in the US airline market, which post much consolidation today sees far less discounting. Instead today the sector enjoys higher fares/profits. It will be the same factor that drives RyanAir’s future profitability also.
First AMD and now INTEL…. Running hard to stand still
The more we have read about the recent history of company like TSMC and its sub-sector of semi-fabrication, the more we see this sector consolidation mental model in action. Massive consolidation has already occurred resulting in a very small number of remaining players. The background of AMD and its spin out of what become Global Foundries is instructive. Despite a long-held desire to hang on to in-house fabrication in the end the economics of the division forced its spin out. Then the stand alone cost of such an independent company keeping up with the likes of TSMC hampered it further. Just a few years of sub-scale investment vs an industry leader in such a fast advancing industry can leave a company behind in a way it has no hope of catching up from. This was the fate of Global Foundries.
Arguably the same fate, or similar transition is now occurring at Intel, albeit in slow motion. We observe that Intel lost the scale war against TSMC many years ago. It then also arguably went on to lose a skill war in the last 2-4 years as a good number of its highly respected engineers left the company. That this occurred as the company was being run more for profit stabilisation/maximisation rather than innovation for the benefit of future customers has now been revealed. The result was the recent (2021) re-hiring of a more technological focused CEO to revive the company’s production abilities. Only time will tell as to how successful he will be. Arguably the cards are stacked against him. Even just to make such a conclusion against a company like Intel, once the gorilla of its industry, speaks to the power of market position TSMC now displays.
Importantly such big changes in the status, scale and skillset of Intel only occurred quite recently (last 2-4y). Looking at the resulting challenges Intel now faces in the context of the very long investment periods TSMC have undertaken investing in plant, IP, R+D and staff would suggest any fab production recovery at Intel will be a long time coming. In addition, arguably all Intel’s current actions are designed to achieve is a maintaining of its current status quo i.e. to be able to remain a chip manufacturer as well as a designer. As such a, perhaps unlikely, success achieved by Intel in rebuilding a leading edge chip production ability would not impact TSMC negatively, as its market share of chip production would stay the same as today. In contrast were Intel to fail in such a challenge and finally outsource manufacturing to TSMC, as was mooted in 2020, this would be a huge benefit to TSMC. Such an endgame for Intel would not just give a short-term fillip of further scale to TSMC. It would arguably gift it a scale and technological lead that could last for decades. Its manufacturing advantage would surely then be assessed as almost unassailable.
China competition threat: Dead on Arrival
Our reading of Chip War[1] helped our understanding in the area of potential Chinese competitor threats a great deal. What we think is important to recognise here with reference to the TSMC competitive environment is not that a visible powerful competitor has been undermined. Instead, it is that any ability to create a future competitor has been stopped before it even had chance to succeed. That a priority of Chinese policy was/is to be a leader in leading edge technologies’ is widely understood. A such the Chinese state (by fair means or foul) has engaged in a determined multi-year effort to try to takeover, copy and/or build the necessary skills it needs to design and produce leading edge semi-conductors. The recent US Government policy has not only stopped US chip design/technologies being exported to China, it has also spread that constraint to Dutch and Japanese important parts suppliers. This has effectively meant the idea of any Chinese backed production of advanced semiconductors is dead before its arrival.
To achieve such production excellence under a China state sponsored/supported entity would never have been easy even with access to US technology hardware and software. This is due to the long term learned skill base and industry co-operation required to do so. Without US/Dutch technologies a Chinese semiconductor manufacturing company, be it state backed or not has zero chances of success. With access to huge capital, long time frames for potential success and availability of US technologies, building a Chinese competitor to TSMC would not have been easy, and even if successful might still have taken c. 10 years. Now any such threat looks dead.
3 becomes 2… Becomes 1?
The combination of a much-weakened Intel and a dead-on-arrival Chinese aspirations to establish a new well-funded competitor we think are powerful factors. Both of these events are arguably quite recent in their occurrence – (Intel exodus of talent occurred in 2020 leading to the change of CEO, appointed Jan 2021 and US CHIPS Act was only enacted in 2022). However the implications from these changes to any investor assessment of TSMC we think are long lasting. Whilst TSMC was clearly already the leader by far in this field, Intel’s efforts show how hard it is even for a past market leader to catch up, despite being very very determined to so do. (#Surfing vs Wallowing).
That a global state venture of prominent ambition, funded by the deepest pocket in the world might now also have to accept failure in its challenge is important when considering what TSMC’s competitor positioning might look like ten or even twenty years out. In another industry where the competitive advantage is say a brand or current technology such moments might not matter as company’s SCA may change in time. However, when your advantage is due to the use of scale and process power the ability for future competitors to catch you becomes reduced, maybe significantly so. Such missteps by competitors/good fortune in policy changes are today we creating potential gulfs in the size of TSMC relative moat size in the future.
TSMC meet Amazon
In our Amazon work we were impressed by its fixation not on its competitors but instead on the customer.
We’re divinely discontented with customer experiences, whether they’re our own or not. We believe these customer experiences can always be better, and we strive to make customers’ lives better and easier every day – Andy Jassy, our emphasis
“We seek to be Earth’s most customer-centric company. We are guided by customer obsession rather than competitor focus – Jeff Bezos, our emphasis
This same customer centric approach exists inside TSMC. Whilst its founder (Morris Chang) has now once again stepped aside, the culture that he ingrained seemingly remains strong inside the company. A healthy paranoia of complacency clearly exists:
“Improve every year or you will be eliminated”. TSMC CEO November 2022
This is also evidenced not just in its internal culture but in its relationship with customers. Customers clearly love TSMC, particularly its responsiveness. Its consistent customer feedback surveys and the transparent way TSMC runs open customer relations give an insight into why customers feel this way.
“We are a service business not just a production business” TSMC CEO FY 2021 analyst call
Maybe this culture can also be seen in the CEO’s recommendation of a business video to watch to better get a feel for the company’s culture and strategy. In a presentation last November he cited: ‘Why the mighty fall’ by Jim Collins. This is clearly a company that has internalised Munger’s view on corporate death.
“Work out where you are going to die… and never go there”. Charlie Munger
We reassert a few points we made in notes on companies like Schwab and Amazon. This being that a huge commitment to give the customer great service and great prices has multi-faceted outputs. It means that a customer is pleased with you and gives you more business. That customer also tells their friends of your combination of service and value and they give you new sales without a need for marketing or gimmicks. There is a third element/effect of these customer centric actions however… the effect it is having on your competitors.
Not only if customers are switching to you are you gaining more scale advantages, but such a move is taking important scale away from your rivals. In turn making any dis-economy they already possess even more painful. In short, if you are struggling to compete with a business-like Amazon, Schwab or TSMC today, tomorrow will be worse!
Observing TSMC competitors in the last few months we see big recent-year changes in a weaker Intel and much weaker China state-funded ability to compete. What we also see however is how the Scale Economy/Process Power model of TSMC presses its advantages home. With each ratchet up in scale they deliver yet more value, more innovation and service to customers. Buffett often talks of ignoring many other metrics and just asking yourself “is this company’s moat narrowing or widening?” In steady state TSMC’s moat looks wide today. The interesting bit is it seems set widen more and do so for many years to come.
Regulatory tailwinds
Arguably much has changed in the regulatory environment for TSMC in recent months. Whilst we have addressed some of the knock-on competitive impacts of The CHIPS Act above it was clearly started by regulatory/retaliatory actions.
The war of words and then trade that broke out post the US accusing China of stealing its technology/IP is interesting. Alarming as it may read to some, these events are not without precedent. Again we thank Chip war for the insights it gave us here. The same copying/stealing took place by the Soviet Union many years ago. It was notable then that such an approach did not foster the internal innovation the USSR aspired to create. The opposite occurred with a greater reliance on US technology as a copying mentality outweighed one of internal innovation. Arguably evidence of China’s efforts is this regard are yielding a similar outcome today.
We note the innovation that can be achieved with trusted employees in partnership with long-standing suppliers/customers is very hard to foster/replicate in a state sponsored environment where the same parties do not have an equivalent level of trust. This is notable in China today, i.e. seemingly they just cannot replicate voluntary/commercial networks of deserved trust that exist between US/EU counterparties. Whilst this is not a piece extolling the virtues of capitalism and free markets, this appears to be a real-life example when even a well-funded and organised planned economy just cannot compete with open architecture entrepreneurism.
We observed in our first piece that the US CHIPS Act’s global enforcement, whilst not of TSMC’s political choosing stands to help it greatly. (Indeed TSMC tries very hard to be politically neutral). Its impact however is something that will be a long-term beneficiary to TSMC by the weakening of Chinese competitors as we observed above and the de-facto establishing of TSMC as the global gold standard in semi-conductor production. On this last point we note that very few companies breaking ground on a new US factory get both the Apple CEO and the US President turning up for the opening!
We would also now obverse the contrast between TSMC and other global technology companies from today’s starting point of regulatory risk. Many large tech businesses like Amazon, Meta, Apple and Alibaba are seeing far higher regulatory scrutiny of their activities with overseers looking for areas where a monopoly has been abused or just allowed to get too big. Against this backdrop TSMC is being courted with tax incentives by multiple jurisdictions. The 2022 US CHIPS Act penalised its would be Chinese competitors, while at the same time offering generous tax breaks if the company opened operations based in the US. This was soon followed by a new Taiwanese tax incentive tax package of their own. It is clearly designed to ensure the bulk of TSMC research and development stays within Taiwan (the effect of which is TSMC’s current temporary 10% income tax rate, will now become semi-permanent). Lastly a tax incentive of some form by the EU to attract a fabrication plant to mainland Europe is seemingly on the horizon also.
What goes wrong?
As we have reflected more on TSMC we make a few observations on why our positive thesis might not pan out.
Political risk
Much is written on the China/US relationship and the important geographic position that Taiwan plays within that. We feel our view in this area can add relatively little. We only reflect that current trade relationships and indeed the relationship of China with investors more widely is strained currently. Whilst we see some signs of a thaw in US/China relations (see FT pieces below) these of course can be wrong. With US/China trade tensions elevated and the Russian/Ukraine conflict bringing Taiwan into the spotlight again at least maybe risks are properly considered and we suggest fully discounted as a result. To consider the financial model of TSMC and its ability to still be able to invest future capital at high rates in isolation, is to see a wonderful business model of high ROIC and plentiful ability to deploy more capital at similar to historic rates. What might such a company be valued at were based in another geography/had less political risk. We will suggest =/> 20x earnings. If we accept 2023 as a weaker year in the semiconductor cycle and instead look to 2024, we find TSMC expected to report $6.9 of EPS per each ADR. The resulting PE, even after the recent share price pop-up is 13.7x. This clearly shows a sizable risk discount is priced into the shares today already (we note looking at the companies past PE’s it has often been at much higher multiples). Today’s lower multiple is despite record high levels of return on capital and high gross margins.
Future fading returns
If you want to understand Mr Market’s current concerns, we recommend listening to the recent full-year analyst call. Whilst there are discussions on different product rollouts and the outlook for the 2023 semiconductor cycle, another item clearly leads analysts thinking. This is the potential declining returns on capital that will result from investments being made outside Taiwan (i.e. in the US/Europe). Clearly the company’s past discipline in sticking with its home market for a low and consistent cost of production is something investors have gotten used to. Today they are being told by TSMC that the build cost of plant in Arizona is 5x that of one in Taiwan. Although this does not seem a like-for-like comparison, clearly there is a higher cost of build outside Taiwan and the company admits this fact. It also admits how important the company is to its customers and how these customers now want some of the semiconductors they buy to be able to be sourced locally to them. Investors are worried this will lead to an uneven marketplace and in time lower returns on capital – they might be right of course. When questioned, the company admits that it will seek to balance these higher upfront costs with:
- some internal efficiencies
- by using the tax breaks new geographies offered to attract the investment and
- with the prices paid for these locally sourced semiconductors likely needing to rise.
Whilst none of this makes for easy modelling, it is easy to see both sides of the argument i.e. by investing in high cost countries you will lose your competitive strength. Alternatively clients are asking for this sourcing and saying they will pay for it, so as a client/service orientated company there is a point you must align yourself with their needs. We are unclear on how this ends up and will admit to some worry. Five years from now, a government or company that promised premium prices for locally sourced chips no longer wants play ball. Equally we are comforted by:
- that the company has pointedly reasserted its gross margin and ROE targets despite the admission of higher overseas costs;
- also that as the group observes it might one day get to c.20% of semis being produced outside Taiwan. This speaks to how big and growing its Taiwan operations will still be. This should ensure that were a credible alternative supplier to one day emerge TSMC still has c.80% of its capacity still on equivalent to lowest unit cost.
Price is what you pay, value is what you get
In our first piece we outlined the scale of compounding that TSMC has been able to achieve in the past. All of it, we note at high returns on capital. Also, that it/ the sector are suggesting another multi-year wave of growth is now ahead. The result of this is an ability to again deploy huge amounts of capital at still very exciting ROICs. Indeed as we roll the company’s likely financials forward it is the absolute quantum of profit the business might make we are attracted to, rather just the percentage ROIC it might report. Systems like Holt et. al. don’t like ROIC declines (if that is what might occur). The reason being is as a measure it often points to a new competitor or some weakening in the company’s business model. Were TSMC to grow at the rates it suggests in the coming years (15-20%) and were the ROIC it reported to have faded a little due to a few overseas investments that its clients requested, is that a problem for investors? That depends. If there are signs of wider weakness in the companies SCA, then maybe. However, everything else we are analysing about this company points to a strengthening not weakening position. Were such absolute profits to have grown strongly but return levels to dip a little we still assess the shares will be far higher than today. That is a Put Up More To Make More model.
Investor returns will also be a function of the price they have paid. A glance at the group’s past multiples show that rarely have they been as cheaply rated in these last 6-12 months and yet the growth and future ROIC prospects are still excellent. Price is what you pay, the value of an amazing, maybe even unassailable market position with strong compounding prospects is what you get.
Please do take a look at our original piece and the many links to useful documents and videos below.
New reading
The above reflections should be read in conjunction with our original piece – Holland Views: What does Warren see (December 2022). These thoughts came from the following sources. All we think are time well spent for those researching the company.
Read Chip wars by Chris Miller.
We have suggested this book a few times now. I had not finished it when we published our first note. It is both interesting as to the history of this sector/company. It is also very informative and up to date having only been published in 2022
Listen to Morris Chang
The founder of TSMC is interviewed in this link. It seems this interview has informed more than a few podcasts we have listen to on TSMC. So it is important to hear the message from the horse’s mouth.
Listen to current TSMC CEO Conference speech last autumn
Current TSM CEO Mr CC Wei spoke at a conference in Taiwan last autumn. Whilst it is translated the translation is far from perfect (I am looking for a link to a better one). That said it still provides a great insight into the companies thinking on sustainable competitive advantage and the evolution of the semiconductor market.
List to the company’s most recent year end conference call with investors
This is also an important listen, particularly with reference to whether the planned deployment of capital outside Taiwan will/will not change the company’s financial outlook.
Read some of the recent FT articles on China
They suggest, a thawing of relations between US + China
https://www.ft.com/content/e592033b-9e34-4e3d-ae53-17fa34c16009
https://www.ft.com/content/f6c47c54-b928-46ee-aa63-fb903e74b2e6
These downloads from HBR (and elsewhere) were also helpful
https://www.hbsp.harvard.edu/product/W25607-PDF-ENG?Ntt=
Kind regards
Andrew Hollingworth
Disclaimer
This document does not consist of investment research as it has not been prepared in accordance with UK legal requirements designed to promote the independence of investment research. Therefore even if it contains a research recommendation it should be treated as a marketing communication and as such will be fair, clear and not misleading in line with Financial Conduct Authority rules. Holland Advisors is authorised and regulated by the Financial Conduct Authority. This presentation is intended for institutional investors and high net worth experienced investors who understand the risks involved with the investment being promoted within this document. This communication should not be distributed to anyone other than the intended recipients and should not be relied upon by retail clients (as defined by Financial Conduct Authority). This communication is being supplied to you solely for your information and may not be reproduced, re-distributed or passed to any other person or published in whole or in part for any purpose. This communication is provided for information purposes only and should not be regarded as an offer or solicitation to buy or sell any security or other financial instrument. Any opinions cited in this communication are subject to change without notice. This communication is not a personal recommendation to you. Holland Advisors takes all reasonable care to ensure that the information is accurate and complete; however no warranty, representation, or undertaking is given that it is free from inaccuracies or omissions. This communication is based on and contains current public information, data, opinions, estimates and projections obtained from sources we believe to be reliable. Past performance is not necessarily a guide to future performance. The content of this communication may have been disclosed to the issuer(s) prior to dissemination in order to verify its factual accuracy. Investments in general involve some degree of risk therefore Prospective Investors should be aware that the value of any investment may rise and fall and you may get back less than you invested. Value and income may be adversely affected by exchange rates, interest rates and other factors. The investment discussed in this communication may not be eligible for sale in some states or countries and may not be suitable for all investors. If you are unsure about the suitability of this investment given your financial objectives, resources and risk appetite, please contact your financial advisor before taking any further action. This document is for informational purposes only and should not be regarded as an offer or solicitation to buy the securities or other instruments mentioned in it. Holland Advisors and/or its officers, directors and employees may have or take positions in securities or derivatives mentioned in this document (or in any related investment) and may from time to time dispose of any such securities (or instrument). Holland Advisors manage conflicts of interest in regard to this communication internally via their compliance procedures.
