top of page
  • Facebook
  • Instagram
  • Twitter
  • TikTok

The toil of semiconductors

This article is authored by ATE Research.


The bulk of the publicity behind the magnificent seven can be attributed to the game changing technologies they continuously roll-out. However, from the data centers that run Amazon Web Services (AWS) to the algorithm behind Google’s search engine, none of it would be possible without semiconductors. This has created an eco-system behind big-tech that often goes unnoticed. In turn, substantial shareholder value has been created across a variety of businesses that each serve a crucial role in the designing, building, and distribution of semiconductors. This is what makes semiconductors a promising place for anyone to store their money for the next 5-10+ years.

Data, Data, and Data…

The modern economy is driven by one thing, data. Data is the ultimate commodity since it’s the only resource that gains value as it grows in abundance. This explains why institutional investors are fiercely gripping onto the magnificent seven. Big tech collects the most data, so it makes sense that their growth prospects are perceived to be unlimited. Although this is a flawed thought process from an investment perspective, it gets across my point. The collection and allocation of data, mainly emanating from big tech, will require more semiconductors with increased processing capacities for a long time to come. Moore’s Law is the driving force behind this phenomena, which is the observation that the number of transistors, or connections, on a semiconductor will double every 1.5-3 years. Although this unnatural law will be coming to an end sooner or later due to the laws of physics, it doesn’t mean that semiconductor innovation is dead, in fact we’re far from it. The prime example of this is NVIDIA, tying in my discussion of big tech. NVIDIA’s revenue grew at a compound annual growth rate (CAGR) of 18.1% between Fiscal 2019-23 (2018-22) while net income had a CAGR of only 1% through that time. Meanwhile, worldwide semiconductor industry revenue had a CAGR of 4.9% between 2012-23 and 4.7% between 2018-22 (Statista). These numbers get across multiple ideas, which are (1) the lack of earnings behind the U.S.’ biggest companies relative to their valuation (2) NVIDIA’s representation of the semiconductor industry’s revenue expansion and product innovation (3) the robustness of the semiconductor industry’s growth from market peak to peak (2018-22) and over long periods (2012-23).


NVIDIA is a great business worth over $1.5 trillion. If they were to rival Apple and Microsoft at $3 trillion, a “maximum” return of 100% is possible. This limit of upside is what attracts me elsewhere, ASML and TSMC. TSMC is the most advanced and voluminous manufacturer of semiconductors in the world while ASML builds lithography machines that make it all possible. Both of these companies have a significantly smaller market cap (valuation) than NVIDIA but are operationally equal or superior. In effect, NVIDIA is worth triple TSMC and 5 times ASML while outputting less than TSMC and 2.5 times ASML when adjusting for NVIDIA’s projected Fiscal 2024 net income of around $20 billion. To summarize, NVIDIA is an awesome group making game changing technology, but investor confidence has turned NVIDIA shares into speculative vehicles, not investment opportunities. Likewise, NVIDIA is a fabless designer of CPU’s and GPU’s, so their business model isn’t representative of the semiconductor manufacturing process. Thankfully, this has turned the medias lenses away from companies that truly carry the industry like ASML and TSMC. With lack of attention comes a shortchange of investor appreciation, which possibly applies to ASML and TSMC as well as a few other businesses I’ll now introduce you to.

Beyond The Leaders…

Smaller companies generate higher rates of return, just like e=MC². These organizations get less attention than ASML, NVIDIA, and TSMC, are more fairly valued, and will play a bigger role than in the past in terms of perpetuating semiconductor evolution. A few of these companies are Amkor Technology, ASE Group, and Photronics. Amkor and ASE are both involved in the packaging of semiconductors, which involves connecting a semiconductor with a printed circuit board (PCB). An example of packaging companies spawning innovation is Amkor’s SiP (System in Package) offering. SiP allows for more features with reduced package size, a harbinger of value for integrated device manufacturers (IDM’s) like TSMC, Intel, Global Foundries, UMC, and Samsung. Meanwhile, ASE and TSMC have been strategic allies since 1997. Since then ASE has become the world's largest outsourced semiconductor assembly and test (OSAT) company while TSMC is the globes dominating semiconductor foundry. As for Photronics, they’re the only independent manufacturer of photomasks, which are a vital component to the photolithography process. Amkor, ASE, and Photronics are all worth less than $20 billion and trade below a TTM P/E of 20 all while the semiconductor feedback loop is regaining its momentum. Other examples of value to be found in the present semiconductor ecosystem are STMicroelectronics and Infineon, who both have exposure to automotive IC’s, which is widely considered to be the highest growing segment of the semiconductor market.


The biggest company in any industry is the one by which all others are judged. In this instance, NVIDIA’s market cap (valuation) foretells of all the great things that’ll emerge out of the semiconductor industry over the next 5-10 years. The businesses that will actually account for this transformation haven’t had this same viewpoint applied to them though. My case lies here.


bottom of page