TMT Trend: Chip Wars and Semiconductor Deglobalization

Perhaps the single best way to really stand out during tech investment banking interviews is to show that you understand that the remit of tech investment bankers extends far beyond dealing just with flashy social media companies. This is especially true today, given that dealmaking involving unprofitable tech companies has fallen significantly relative to the past few years when SPACs were all the rage.

The reality is that part of the reason why deal flow in tech investment banking is so much stronger than in most other coverage groups is due to how many relatively distinct industries fall under the tech coverage umbrella. As we discussed in the TMT banking primer, tech investment bankers will deal with companies with wildly different characteristics: from internet start-ups that have eye-wateringly high gross profit margins to semiconductor manufacturers that are as capital intensive as any company in any industry.

Given just how capital intensive the semiconductor industry is - along with how tightly regulated it is – there’s naturally quite a bit of concentration in the space. This has the natural consequence of making deal volume quite low, but those deals that do happen are often large, complex, and very time-intensive to put together.

So that’s why you’ll see some MDs really specialize in the semiconductor space and why you’ll see them frequently get poached by banks as they try to increase their own share of the relatively limited deal flow pie.

Part of what makes the semiconductor space so fascinating is not only the mind-bendingly level of complexity that goes into making the actual semiconductors – or even just creating the inputs to make them – but also how tangled up semiconductors have become with geo-politics. Thereby making nearly every deal one that needs to be viewed through both a finance and geo-political lens.

Indeed, it’s relatively rare that regulatory developments in an industry begin to capture front-page headlines in the popular press. But, as you’ve likely seen, this is exactly what’s happened over the past year as the so-called chip wars have begun.

The “chip wars” is a turn of phrase coined to refer to the pseudo arms race taking place between China, the US, and (to a lesser degree) the EU to dominate semiconductor manufacturing. The reason being is simple: not only are chips increasingly being used in everything – from cars to phones to refrigerators – but advancements in chips are what allows for breaking into new frontiers in the world of artificial intelligence and weapons design (to name just a few). 

In this post we’ll quickly be covering some of the latest developments in the chip wars (all of which would be impressive to bring up during regular interviews or group placement interviews) while also highlighting a few fantastic graphs put together in a recent Deutsche Bank research report (incidentally DB just hired away a lead banker from RBC who focused on semiconductors). 

On This Page

I’ll try to keep this post relatively short and high-level. But I’ll drop links to more detailed articles along the way if you want to dive in further (since the chip wars aren’t really new, but rather have been brewing under the surface for years now).

    The US Takes a Stand on Semiconductors

    On October 7 of 2022, the US took action that was inconceivable just a few years ago: banning domestic companies from exporting any equipment, software, or technology used to produce advanced chips and supercomputers to certain semiconductor manufactures in China. 

    More specifically, the new rules mean that high-end chips (or the inputs that could create them) that are produced in the US can only be exported to China with a special license (that will be very difficult, or perhaps impossible, to obtain).

    The move sent shockwaves across the industry, as building up the domestic semiconductor industry is a central part of China’s five-year plan and building up their domestic industry requires leveraging US software, technology, and general intellectual property – at least for now. Further, semiconductor chips are the world’s most traded product (given that they’re now an indispensable component of almost every electronic device) and denying China access to more advanced chips will set back their aspirations of moving to the bleeding edge of areas like artificial intelligence by years. Thus, putting them at a distinct disadvantage relative to the US.  

    Additionally, an under-discussed element of the new rules is that they will also prevent individuals who are US citizens or US green card holders from actively supporting the development or production of advanced chips in China without a US government license. Here's a good write-up from Bloomberg on this aspect of the new rules.

    So, this move isn’t just setting a small speedbump along China’s path. Rather, it’s something much more comprehensive that is meant to insulate China from taking advantage of the equipment, intellectual property, and even people who can help them advance their fledgling semiconductor industry. This was immediately recognized by China who made it clear they viewed these new rules as an actively antagonistic act -- one that will likely involve some level of retaliation in the months or years to come.

    While the new rules got an (understandably) chilly reception from China, for the US the introduction of the rules had a clear strategic rationale: instead of trying to maintain competitive advantage over China by outpacing their technological advancements (i.e., trying to always stay one step ahead of them) it was decided to try to break China’s advanced chip supply chain due to a fear of how exactly more advanced chips would be utilized and what China developing any form of competitive advantage in the making of chips would portend for the US.

    From Globalization to De-Globalization

    Within the US, there is an increasing concern that – in the decades to come – China will begin to dominate the global production of semiconductor chips, and that could be used as leverage over the US given the relatively small amount of manufacturing that takes place domestically (even though much of the design of chips still occurs within American borders).

    There is an obvious historical parallel here that makes this worry not entirely unfounded. As globalization became the word du jour in the 1970s, many American companies decided to export the manufacturing of their goods to cheaper countries while maintaining their design, marketing, strategy, etc. activities domestically.

    Semiconductor Historical Timeline

    This had the benefit – both to consumers and to companies – of dramatically lowering costs. In the world of semiconductors, the point in time many would highlight in which globalization really took off would be 1976 when the Taiwanese government had RCA – a major US semiconductor company – transfer semiconductor technology to Taiwan.

    In the ensuing decades, Taiwan increased fabrication capacity dramatically and today TSMC represents ~55% of total semiconductor production and ~90% of advanced chip production. Meanwhile, as Deutsche Bank noted recently, since just 1990 the US share of global semiconductor fabrication fell from 37% to just 12% today.

    Currently it’s being recognized that while the globalization of semiconductor manufacturing brought about unprecedented efficiency gains, the supply chain for the US is now resting on an increasingly shaky foundation. It’s one thing for much of the design of semiconductors to still occur within American borders, but if there’s no way to manufacture them then that hardly does one much good in the advent of any meaningful geo-political conflicts arising. 

    Note: When it comes to research and design, the US is still the clear global leader with more than 80% market share in design automation software and more than 40% market share in semiconductor manufacturing equipment (keep in mind, advanced semiconductor equipment can cost into the hundreds of millions).

    This is partly why two critical acts were passed by Congress over the past two years: the 2022 CHIPS Act and the 2021 FABS Act. Both are meant to elevate the manufacturing capacity of the US (in the decade to come) to some level nearing self-sufficiency (i.e., trying to have the design and equipment created domestically put to use domestically, as opposed to having it exported to other countries to be utilized).

    When folks talk about how we could have structurally higher inflation moving forward due to deglobalization, this is one area that is often pointed to. Because in order to insulate oneself from global supply chains, by producing as much as possible domestically, that will require what is being produced domestically to cost considerably more (as there’s a reason why production was done in other countries and then imported back in to begin with!).

    Here's how Deutsche Bank puts it… 

    “The Russia-Ukraine conflict has revealed what could be the formation of three international blocs: (i) US and allies; (ii) China and allies; and (iii) rest of the world. If these geopolitical silos perpetuate, and if the US continues its strategy of targeted decoupling, the globalised semiconductor supply chain may also become fragmented along such lines. Additionally, at least in the long term, China could become more important for many non-US chip companies…”

    “These companies might struggle if the currently integrated supply chain is split into separate semiconductor ecosystems. The first would likely be centered around US semiconductors, the second based on a growing Chinese domestic industry. Taiwan may be caught between these two ecosystems, which could become yet another point of contention due to Taiwan’s industry dominance. TSMC, Taiwan's largest semiconductor company, has 54% market share in semiconductor manufacturing…” 

    “However, over 97% of TSMC's long-term assets, which include its fabrication facilities, are located in Taiwan. However, it is neither in China nor the US interest to damage Taiwanese chipmaking capacity or cut off access to Taiwanese chips. Additionally, neither would be able to easily replace Taiwan- tech parity would likely take years; achieving the same cost efficiency would be improbable, and it would likely take many years to yield a positive return…”

    The Impossible Complexity of Global Semiconductor Supply Chains

    It’s one thing to talk about deglobalization and the decoupling of supply chains, but it’s quite another thing to actually go about creating a comprehensive domestic supply chain. Because, as a result of globalization, many different areas of the world have become highly specialized in increasingly niche areas of semiconductor manufacturing.

    Right now, as DB notes, there are in excess of fifty points across the semiconductor supply chain where one region holds more than 65% of the global market share. In other words, it’s impossible to conceive of any one country being able to become entirely self-sufficient across the entire supply chain, or even a group of large allies being able to be so. And this doesn’t even get into how wildly inefficient, from a cost and manufacturing perspective, doing everything within one country or a small group of countries would be.

    Here's a great graphic illustrating the geo-political spheres of influence at play…

    Global Supply Chain Semiconductor Map

    To put a finer point on this, the Semiconductor Industry Association estimates that if each region tried to become truly self-sufficient – assuming that this is possible to begin with – their domestic semiconductor prices would need to rise 35-65%. Which, of course, would feed through to higher inflation in the economy as so many goods now have chips making up a non-trivial percent of their production costs (e.g., cars, appliances, etc.).

    What shouldn’t get lost in any conversation about semiconductors is the sheer size of the industry. Here’s a great chart illustrating global semiconductor revenue from Deutsche Bank. What it illustrates is that within just a few years, global revenue will be closing in on one trillion dollars a year. And that underestimates just how critical semiconductors are to the general economy, since without them, or with them costing much more than they do now, much of our modern way of living would radically change. 

    Global Revenue Semiconductors


    Today, there is one theme in the world of semiconductors that overhangs all others: the chip wars and just how far semiconductor supply chain decoupling will go. From a deal making perspective, this is both positive and negative. It’s positive in that it may lead to some forced divestment occurring or domestic conglomerates trying to do some downstream or upstream integrations, both of which would be positive from a dealmaking perspective.

    However, the negative is that due to the geo-political implications of the chip wars, almost all semiconductor deals will be heavily scrutinized by regulators and the universe of those who will actually be able to do acquisitions will shrink from what it once was (indeed, we’ve already seen the break-up of some prominent semiconductor deals over the past few years).

    Anyway, whether you’re interested in doing TMT banking or just tech banking, it’s worth paying attention to what’s happening in the semiconductor space. Not only because it would be impressive to bring up an interest in the semiconductor industry in an interview (since not many will!) but also because it’s an incredibly interesting area to get exposure to if you have the opportunity (since it involves companies with incredibly complex business models who are operating in one of the most complex regulatory and geo-political environments imaginable).

    But, obviously, your interviewers won’t be proactively asking you about anything we’ve discussed here. So make sure to go through the classic TMT investment banking interview questions, tech investment banking interview questions, etc. if you’re currently getting ready to do some interviews.

    Leave a comment

    Please note, comments must be approved before they are published