Summary of Key Points
This article challenges the common belief that "the most critical links in a supply chain are the most profitable" and that "upstream companies are more valuable than downstream ones" by highlighting the unusual situation where ASML (the world's only manufacturer of EUV lithography machines), with a gross margin of 52.8%, earns less than its customer TSMC (66.2%) and even less than its downstream customer NVIDIA (around 75%). The article identifies three common misconceptions in industrial planning:
1. When everyone rushes to the same "correct path," profits can be eroded.
2. Not all industries can establish monopolistic barriers through technical expertise alone.
3. Even with a monopoly technology, it doesn't necessarily lead to substantial profits, as factors such as customer composition, cost positioning, and business models play a role. The article concludes that the key to industrial strategy is not about "choosing the right path" but about finding a niche where there are few competitors and where one can remain competitive for a long time—this position is earned through persistence, not selected.
Breakdown and Interpretation
1. Why Does ASML, Which Monopolizes the Upstream, Earn Less Than Its Downstream Customers?
ASML is the only company capable of producing EUV lithography machines, which are essential for advanced chip manufacturing. Yet its gross margin is 14 percentage points lower than that of TSMC and more than 20 percentage points lower than that of NVIDIA. This contradicts the intuition that the further upstream a company is in the supply chain, the more it should earn. The article explores the underlying logical issues in industrial planning.
2. Misconception 1: Everyone Choosing the "Correct Path" Leads to Profit Erosion
The article points out that city-level industrial plans often emphasize sectors like artificial intelligence and biomedicine as promising areas. However, when too many companies compete in these fields, profits can be diminished. For example, in the photovoltaic industry, excessive production has driven down component prices below cost, resulting in losses for the entire sector. The same is true for lithium batteries and energy storage technologies, with the current focus on large-scale models and computing power centers.
This phenomenon is similar to the "winner's curse"—the bidder with the highest bid often overestimates the value of the asset. In industrial policy, cities may all aim for the same high-end segments, collectively reducing profits. Economists call this "rental value dissipation"—when everyone queues up at the same popular restaurant, the long waiting time negates the perceived benefits, leaving no one truly benefiting.
3. Misconception 2: Not All Industries Can Establish Monopolistic Barriers Through Technical Expertise
To avoid being overwhelmed by competition, a company must find a segment that is difficult for others to enter. However, this depends on whether the technical expertise can be sustained over time:
- Semiconductors as a Case Study: ASML's EUV technology took 20 years to develop and involves hundreds of thousands of components, funded by customers like Intel, TSMC, and Samsung. Their expertise has made them increasingly competitive; later entrants cannot catch up without the same investment.
- Biomedicine as a Counterexample: Once a leader in gene sequencing, Illumina now faces competition from companies like PacBio and BGI Genomics, causing its stock price to drop by 60% in five years. Biomedicine requires constant innovation, with each new drug development being a new gamble. This is exemplified by the "Eroom Law" (the reverse of Moore's Law): the number of new drugs produced per billion dollars in research investment halves every nine years.
Therefore, not all industries can become monopolistic like semiconductors, and blindly pursuing a "ASML-like" position may be futile.
4. Misconception 3: Even with Monopoly Technology, High Profits Are Not Guaranteed
Even though ASML monopolizes EUV technology, its low gross margin is due to several factors:
- Customer Lock-in: ASML has only a few key customers (multibillion-dollar giants) that contribute to its research and development efforts, limiting its ability to charge higher prices.
- Cost Structure: ASML's machines are a cost factor for TSMC, which seeks to reduce costs by purchasing them. In contrast, NVIDIA's GPUs generate revenue for cloud providers, allowing them to charge premium prices.
- Business Model Differences: ASML operates in manufacturing (high fixed costs), while NVIDIA sells design services with nearly zero marginal costs due to outsourcing. Technological monopoly does not equate to high profits; business models play a crucial role.
5. The Right Approach to Industrial Strategy
The article suggests that the focus should be on finding a niche with few competitors where one can remain competitive for a long time. This is not about choosing the most profitable path but about persisting in a market dominated by a few players. For cities and industrial parks, the goal should be to create environments that help companies generate more revenue (by attracting customers, generating orders, and providing use cases), rather than simply trying to be cheaper (since the lower-cost limit is often zero profit).
In summary, instead of competing for the same "correct paths," it's better to find untapped markets and wait for others to be deterred from entering. That’s where true value lies.
(The entire article avoids using technical jargon and uses everyday examples like popular restaurants and lotteries to make complex financial concepts accessible to a wide audience.)