Distilling Frenzy #3: Value Creation vs Value Capture, Hardware vs Software

It is now conventional wisdom that competition among firms creates value for society.[1] There was a point in the 20th century when this was in doubt—might central planning avoid the wasteful, zero-sum aspects of inter-firm competition?—but we are in 2020 now. From a whole-of-society vantage point, economists have the privilege of overlooking how this value is allocated.[2]

In business, however, the difference between value creation and value capture is huge. For example, I have long known that Lee Kuan Yew considered air-conditioning the greatest invention of the 20th century. But Byrne Hobart of The Diff has given this trivia an interesting twist:

The challenge, sometimes, is not creating a valuable product, but creating a valuable product and capturing the value. Government contractors run into this issue all the time: building a better website for filing unemployment claims would make millions of people better-off, but building a website that’s barely functional but technically meets the spec pays just as well and is a lot easier.

Or, to take another example: the economies of Singapore, Dubai, and Atlanta can only function because of air conditioning. The industry has created trillions of dollars in value, but that’s mostly captured by the buyers, not the sellers. Sometimes not even the buyers: when department stores started adding air conditioning, the first department store to do so benefited from incremental customers and sales. But once every store had them, market share returned to normal; it was just a cost of doing business. A benefit to the customers, but not to the owners.

In a sense, this is simply how civilization works. We get to free-ride on the ideas and innovations of the people who came before us. Humanity’s greatest hits are transmitted in the form of intangible information, via textbooks and Sunday school and memes, as well as in the form of tangible artifacts like air-conditioners and plastic bags and iPhones. Cesar Hidalgo poetically calls the latter “crystals of imagination” in Why Information Grows.

Information Transfer Economics: A review of Cesar Hidalgo's Why ...

And so it goes.


This distinction between value creation and value capture is highly illuminating in tech.

Consumer-facing software companies, like economists, sometimes have the luxury of ignoring the distinction. In the early years of the consumer Internet, the standard playbook is roughly:

  1. Create a zero-marginal-cost product that is highly valuable to consumers.

  2. Figure out a way to capture a portion of that value.

This is why, in Facebook’s early days, Mark Zuckerberg’s disinterest in monetizing the social network and keen interest in growing the social network were in fact the optimal strategy. As Antonio García Martínez puts it in Chaos Monkeys:

Before 2013, if you wanted to know how Facebook made money, the answer was very simple: a billion times any number is still a big fucking number. Facebook monetization was laughable compared with Google’s on a purely CPM basis. But usage was ungodly. Up there with heroin, carbohydrates, or a weekly paycheck: that’s how addictive and rewarding Facebook was.

As such, it is entirely credible that when Facebook launched Libra, monetization was not a top priority. Facebook has no lack of billion-user apps through which it could capture the value it helped to create. (Assuming Libra takes off, of course, which is a big assumption.)

However, transitioning from value creation to value capture is much easier if company owns the customer relationship. For example, after opening up the platform for free-wheeling third-party experimentations, Twitter significantly tightened its APIs for third-party clients in 2012 because it figured out that its preferred method of value capture is serving ads. To do so, it had to make sure users access Twitter via its own client.

This direct relationship with the consumer also underlies the network effect of many consumer software companies: the value of many software products increases with the number of users (up to a point). Unlike enterprise customers which are wary of lock-in, consumers are usually too scattered to coordinate and switch to a different provider. (This is why exceptions like this are newsworthy.)

For companies that do not own the customer relationship, however, their ability to capture the value they helped to create is significantly attenuated. I came across this Tweet by Jonathan Blow that has stuck with me since:

Unlike software companies, hardware companies are forced to be much more conscious of the distinction between value creation and value capture.

Hardware companies’ have a weaker relationship with end users: their products are typically purchased by users to run software written by other companies. The network effects accrue to the software company instead. While there are economies of scale in hardware manufacturing, many hardware components are commodities sold by East Asian suppliers. Hardware companies thus find themselves in a Red Queen race where they must constantly run to stay in the same spot. Compaq’s descent into commoditization is a cautionary tale.

Semiconductor companies have an even more attenuated relationship with end users, who typically purchase processors as part of a fully assembled product. If you know the difference between Snapdragon and Exynos, you’re probably not the average user. This disadvantage is why you can find Intel’s stickers on PCs—Intel wishes to strengthen its relationship with end users—and why you can’t find these stickers on a Mac. (More here on the history of Intel and TSMC.)

Apple is the exception that proves the rule. As the only BigTechCo with a primarily hardware-based business model, Apple’s core advantage comes from its integration of software and hardware. Not only do Apple products run on a proprietary operating system, Apple has always tried to write the software for the most commonly used apps on its hardware. From this perspective, it is unsurprising that the first iPhone was launched without an App Store and Apple continues to pour money into their Maps app the same way Microsoft continues to pour money into Bing.

Not coincidentally, Apple is obsessed with owning the customer relationship, which has allowed the company to capture the value created by its hardware innovations. By “hardware innovations”, I refer to both the revolutionary—launching the modern smartphone market—and the incremental. An example of the latter would be Apple’s in-house chip design, especially in light of the recent WWDC. Starting with the iPhone, then the iPad and wearables, and now the Mac, Apple’s proprietary chips have boosted the longevity of its hardware’s product differentiation and helped to preserve its own margins.

In other words, hardware companies must lean into software to capture value created by their hardware in an enduring way. (Across the Pacific Ocean, DJI arguably has done the same.)


The distinction between value capture and value creation also explains the locus of the US-China trade war.

When the trade war first began, there was a certain dissonance among China’s tech/VC circles. The advent of mobile and the resulting land grab have created many mega-cap software companies in China. Yet, instead of taking aim at those companies, why has the US focused its sanctions on Huawei instead?

Part of the explanation is surely that Huawei has a much more significant international presence overseas compared to many of these software companies, whose core business remains China-centric.

The better explanation is that Huawei has in recent years surpassed Nokia and Ericsson to become the world’s biggest maker of cellular-tower hardware, internet routers and related telecom equipment. Huawei’s technological lead in telecom equipment is precisely the type of infrastructure work that creates immense value, even if—like other hardware companies—Huawei’s ability to capture this value is limited. Economics is positive-sum but politics is zero-sum. The US is wary of a world in which a value-creating technological bottle-neck is controlled by a company hailing from China.

In contrast, the highly profitable business models of Tencent and Alibaba are “merely” the result of China’s large consumer economy and the inherent ability of software to disproportionately captures a larger proportion of the value it creates. They do not represent any fundamental technological breakthrough unrivaled by Western companies. (No, mobile payment doesn’t count.) As Dan Wang puts it:

I find it bizarre that the world has decided that consumer internet is the highest form of technology. It’s not obvious to me that apps like WeChat, Facebook, or Snap are doing the most important work pushing forward our technologically-accelerating civilization. To me, it’s entirely plausible that Facebook and Tencent might be net-negative for technological developments. The apps they develop offer fun, productivity-dragging distractions; and the companies pull smart kids from R&D-intensive fields like materials science or semiconductor manufacturing, into ad optimization and game development.


In closing, here are some observations based on this essay on Morris Chang (张忠谋), founder of TSMC. (English translation by Joy Dantong Ma and Jeffrey Ding.)

The essay itself is more biographical than analytical, but the following points were noteworthy:

  • The geographical center-of-gravity of chip manufacturing has followed a broad trajectory that is not unlike that of manufacturing generally: US → Japan → Korea → Taiwan → Mainland China (?). It is this last step that is currently ongoing and that the US-China trade war seeks to forestall.

  • New entrants to the semiconductor industry typically ride on the wave of an industry shift. Intel itself was founded as a memory company and found success in its pivot to microprocessors, after Japanese companies made DRAMs a commodity. The TSMC-pioneered contract manufacturing business model both hastened and was made possible by the unbundling of chip design and chip manufacturing. China itself is counting on the slowing of Moore’s law and the rise of ASIC to take a slice of the semiconductor pie.

  • Much of the expertise in semiconductor manufacturing is embodied in the executives and engineers working at a handful of companies. At one point, Samsung caught up with TSMC by poaching a Taiwanese TSMC executive. The essay itself is worth reading for the resulting drama and intrigue alone. (Said executive is now at SMIC, the largest semiconductor foundry in mainland China.)

In this vein, I have recently come across Digits to Dollars, whose posts on ARM’s JV in China and China’s investments in semiconductors are very insightful.


[1] To clarify, the following discussion of “value” is not intended to contain moral judgment. Rather, it is closer to “utility” or “wealth” or “consumer (or producer) welfare” that is generated in the exchange of goods and services.

[2] In the same way, Coase theorem states, and I quote Wikipedia, that “if trade in an externality is possible and there are sufficiently low transaction costs, bargaining will lead to a Pareto efficient outcome regardless of the initial allocation of property.” For regular people, however, the initial allocation of property is important because it determines who has to pay whom!

[3] I, for one, was struck by how long it took for competitors to match the first-generation AirPods’ feature of taking a phone call that played on both earbuds, a feature enabled by the W1 chip. Most other wireless earbuds at that time could only take phone calls on a single earbud and some even struggled to synchronize music playback on both earbuds.

[4] ByteDance is now the ideal test case: what kinds of obstacles will the US impose now that it has achieved significant overseas market share? As I write this, TikTok is currently banned in India—no one knows for certain if this is temporary or permanent.