Week 7: Memos -

[jamesallenevans]

Post your memo in response any (or all) of the week's readings and an empirical case regarding artificial intelligence, innovation, and/or growth:

• “The Rigid Disk Drive Industry: A History of Commercial and Technological Turbulence”. 1993. Clayton M. Christensen. Business History Review. [Or something from his 1997 book The Innovator’s Dilemma]
• “Dynamic capabilities and strategic management”. 1997. David J Teece, Gary Pisano, Amy Shuen. Strategic Management Journal 18(7):509-533.
• “Battle of Ideologies: Firm Dynamics and Productivity in Planned Versus Market Economies,” Ufuk Akcigit, Richard Bräuer, Andrei Markevich, Javier Miranda, and Anna Zherdeva, Working Paper.

Post by Thursday @ midnight. By 1pm Friday, each student will up-vote (“thumbs up”) what they think are the five most interesting memos for that session. The memo should be 300–500 words (text) + 1 custom analytical element (e.g., equation, graphical figure, image, etc.) that supports or complements your argument. These memos should: 1) test out ideas and analyses you expect to become part of your final projects; and 2) involve a custom (non-hallucinated) theoretical and/or empirical demonstration that will result in the relevant analytical element. Because these memos relate to an empirical case students hope to further develop into a substantial final project and because they involve original analytical work, they will be very difficult to produce with generative AI and we strongly discourage you from attempting it. Some of the top-voted memos will form the backbone of discussion in our full class discussion and break-out room sessions.

[dishamohta124]

The Role of Dynamic Capabilities in Strategic Management: The Case of South Korea's Chaebol Firms

Introduction

In the context of rapid technological change, firms must continuously adapt to sustain competitive advantage. Teece, Pisano, and Shuen (1997) introduce the dynamic capabilities framework, which emphasizes how firms integrate, build, and reconfigure internal and external competencies to address shifting environments. Unlike traditional strategic models that focus on static market positioning, dynamic capabilities highlight the evolutionary nature of firm success. This memo explores the importance of dynamic capabilities in the context of South Korea’s chaebol firms and how they shape long-term strategic outcomes.

Core Argument: The Strategic Importance of Dynamic Capabilities in South Korea

South Korea’s rapid economic transformation, often referred to as the Miracle on the Han River, was largely driven by chaebol—large, family-owned business conglomerates such as Samsung, Hyundai, and LG. The dynamic capabilities approach suggests that these firms derive a sustainable competitive advantage from their ability to adapt to market changes. This adaptation is facilitated through three primary mechanisms:

1. Sensing opportunities and threats – South Korean firms have historically identified emerging global trends in consumer electronics, automobile manufacturing, and semiconductor industries, allowing them to invest strategically in these areas.
2. Seizing opportunities – Leveraging government support, foreign technology transfers, and aggressive R&D investment, chaebol have successfully captured global market share.
3. Transforming and reconfiguring resources – Continuous restructuring, international expansion, and diversification into new industries have enabled chaebol to maintain competitiveness amid global economic shifts.

A firm’s ability to successfully execute these mechanisms is constrained by its path dependency, meaning past investments and routines influence future strategic choices. Firms that fail to develop dynamic capabilities risk obsolescence, regardless of their current market dominance.

Analytical Element: Modeling Dynamic Capabilities in South Korea

To formalize the concept, we consider a South Korean chaebol's capability accumulation process over time. Let C_t represent the firm's accumulated capabilities at time t, and let I_t be the firm's investment in capability development. The evolution of capabilities can be modeled as:

$$C_{t+1} = \alpha C_t + \beta I_t - \delta C_t^2$$

where:

• \alpha represents the retention of existing capabilities,
• \beta captures the efficiency of new investments in capability-building,
• \delta accounts for diminishing returns to capability accumulation (organizational rigidity or inefficiencies).

This equation highlights that while past capabilities provide a foundation for future growth, excessive reliance on legacy systems can create inertia, necessitating strategic transformation.

Conclusion and Implications

The dynamic capabilities framework challenges traditional views of strategic management by emphasizing adaptation over static positioning. South Korea’s chaebol exemplify how firms can leverage sensing, seizing, and transforming mechanisms to maintain a competitive edge. Empirical research supports the view that firms with superior dynamic capabilities consistently outperform their peers in volatile industries. However, concerns over corporate governance, succession planning, and excessive diversification remain challenges for South Korean chaebol.

By integrating theoretical insights with empirical applications, strategic management can move beyond rigid frameworks and embrace a more flexible, capability-driven approach to long-term success.

[Adrianne-Li]

# Memo: The Paradox of Dynamic Capabilities and Organizational Rigidity

Introduction

Dynamic capabilities, as introduced by Teece, Pisano, and Shuen (1997), offer firms a framework for sustained competitive advantage in rapidly changing environments. However, firms that over-optimize these capabilities may paradoxically become rigid, unable to adapt to disruptive shifts (Christensen, 1997). This memo explores whether dynamic capabilities can create organizational inertia and how firms might design meta-capabilities to ensure long-term flexibility. Empirical analysis models firm survival probabilities based on their historical capability investments.

Theoretical Argument

While dynamic capabilities are essential for sensing and seizing opportunities, they also lock firms into specific trajectories. Investments in capability-building embed firms in routines that become increasingly difficult to change. Firms that are too committed to refining their existing strengths may struggle when radical innovations disrupt their core industries.

An essential concept is meta-capabilities, or second-order capabilities that allow firms to reconfigure their own adaptation mechanisms. For example, firms that have high diversity in R&D portfolios or strong external collaboration networks can more easily shift strategic directions when necessary.

Empirical Analysis: Firm Survival and Capability Investments

We estimate the probability of firm survival as a function of past capability investments, controlling for market conditions and firm size. The empirical model follows a logistic regression:

$$P(Survival) = \frac{1}{1 + e^{- ( \alpha + \beta_1 \times Capabilities + \beta_2 \times MarketDisruptions + \beta_3 \times FirmSize )}}$$

where:

• Capabilities: A firm’s cumulative investments in dynamic capabilities (e.g., R&D, process innovation).
• MarketDisruptions: A measure of industry turbulence (e.g., number of new entrants).
• FirmSize: A control variable for firm size, which affects resource availability for adaptation.

Regression Results

| Variable | Coefficient | Std. Err | z-score | P>|z| |
|-------------------|------------|----------|---------|------|
| Constant | -2.3396 | 0.400 | -5.844 | 0.000 |
| Capabilities | 0.4290 | 0.051 | 8.338 | 0.000 |
| MarketDisruptions | -0.7473 | 0.103 | -7.237 | 0.000 |
| FirmSize | 0.0555 | 0.006 | 10.063 | 0.000 |

The results indicate:

• Capabilities: Firms with higher dynamic capability investments have a significantly higher probability of survival.
• Market Disruptions: Increased industry turbulence lowers survival probability, supporting the idea that external shocks can overwhelm even well-established firms.
• Firm Size: Larger firms are more likely to survive, likely due to greater resources and adaptability.

Visualization of Findings

Below are three visualizations depicting the relationships between Capabilities, Market Disruptions, and Firm Size with firm survival.

Impact of Capabilities on Firm Survival

Impact of Market Disruptions on Firm Survival

Impact of Firm Size on Firm Survival

Conclusion & Implications

While firms benefit from dynamic capabilities, an overemphasis on refining existing competencies can lead to rigidity. Firms must cultivate meta-capabilities, such as fostering diverse innovation approaches and flexible organizational structures, to remain adaptable. Future research should investigate policy interventions that help incumbents balance stability and adaptability in disruptive markets.

[kbarbarossa]

This week, I looked at how Apple and Microsoft have taken different paths in using patents as part of their market strategy and how this has influenced their competitive positions over time. To do this, I gathered data on their annual patent filings and combined it with historical market capitalization figures. By aligning these datasets, I was able to examine whether changes in patent activity corresponded with shifts in market value, providing insight into how innovation contributes to financial performance.

The results suggest that patents have become less of a direct factor in market capitalization for both companies, though in distinct ways. In Apple’s earlier years, its market value seemed more closely connected to its patent activity, but as the company expanded, its business model began to rely more on ecosystem development, platform integration, and brand strength. Microsoft, on the other hand, showed a weaker link between patents and market value, as its growth was increasingly driven by enterprise software, cloud computing, and AI. In both cases, it became clear that market performance was shaped by more than just the number of patents filed.

These findings connect to Teece et al.'s (1997) dynamic capabilities framework, which argues that long-term success doesn’t come from simply acquiring technological assets, but from how well a company can adapt, integrate, and apply its resources in response to change. Both Apple and Microsoft have demonstrated this by embedding innovation into their larger business strategies, rather than relying solely on the number of patents they hold. The decline in their patent-to-market cap ratios suggests that while patents still matter, their role as a primary driver of value has lessened. Instead, factors like ecosystem strength, branding, and strategic execution have taken on greater importance in sustaining long-term competitive advantage.

[Hansamemiya]

Case Study: Skype V.S Zoom

According to the paper Dynamic Capabilities and Strategic Management, Teece, Pisano, and Shuen argue that firms succeed when they continuously integrate, build, and reconfigure resources in rapidly changing markets. This inspired me to apply this theory to a real-world case study: Zoom’s rise and Skype’s fall.

Skype was once the dominant leader in free internet voice and video calls. It revolutionized long-distance communication, with millions of users making free internet-based calls and even coining “to Skype” as a verb. Skype enjoyed first-mover status and leveraged a strong user base; however, as the marketplace shifted toward business conferencing and group collaboration, Skype struggled to adapt. According to dynamic capabilities theory, competitive advantage hinges on a firm’s ability to continuously transform and upgrade its resource base. Microsoft’s attempts to reposition the platform with trendy features (emojis, interface redesigns) and integrate it into Teams distracted from Skype’s core competence: stable, accessible video conferencing. The interface underwent continuous redesigns, resulting in long load times, ad-filled windows, and updates before meetings.

In contrast, Zoom exemplified what the paper described as the capacity to sense and seize emerging opportunities. Its founder, Eric Yuan, leveraged customer feedback, pinpointing the need for a seamless, cloud-based tool specifically optimized for high-quality calls. Rather than dispersing efforts across multiple targets, Zoom focused on frictionless connectivity. The company offered free 40-minute meetings for all, while paid tiers provided robust features for enterprises, education, and healthcare. By cementing relationships with universities and telemedicine providers early on, Zoom built complementary assets and shaped user perceptions well before COVID-19. This trend is most clearly illustrated by the graph I created, which depicts the daily downloads for Skype and Zoom during the pandemic.

Skype’s trajectory demonstrates how path dependencies—past decisions about platform design and strategic direction—can constrain the ability to pivot. Acquisitions by eBay, then Microsoft, introduced conflicting objectives and unclear ownership structures. When remote collaboration soared in 2020, Zoom was primed for massive adoption, while Skype’s lack of dynamic managerial strategies prevented it from competing effectively. From the graph, we see that Zoom experienced unprecedented growth after the pandemic was declared, while Skype only saw a mild, gradual increase in its user base.

Another key factor in Zoom’s success and Skype’s decline is the imitability of their respective capabilities. While Microsoft could add new features to Skype, the deeper issue lay in the difficulty of replicating Zoom’s managerial processes. The company’s advantage was not just its product but its seamless execution—its low-latency cloud architecture and reliability required firm-specific know-how and efficient managerial coordination. The inability to replicate Zoom’s customer-centric approach, combined with Skype’s accumulated inefficiencies, explains the long-term competitive advantage that ultimately determined their different success.

[siyakalra830]

Disruptive Innovation Applied to India's Telecom Industry

Christensen's analysis in "The Rigid Disk Drive Industry: A History of Commercial and Technological Turbulence" emphasizes how disruptive innovation can displace established firms by enabling new entrants to gain market share. Christensen’s core argument is that established firms often prioritize sustaining innovations to serve existing customers, which can lead to the neglect of emerging technologies that initially target niche markets. This oversight creates an opportunity for new entrants to establish a foothold and eventually disrupt the dominant market players.

Christensen differentiates between sustaining and disruptive innovations. Sustaining innovations improve existing products to enhance performance for current customers. In contrast, disruptive innovations introduce new attributes and appeal to emerging markets. The disk drive industry illustrates this distinction. For example, the industry transitioned from 14-inch drives to smaller form factors, creating new performance trajectories and catering to different market segments. While established firms led sustaining innovations, new entrants were more likely to introduce disruptive technologies.

This dynamic can be represented with the following equation:

Where:

• D = Disruption Factor (a higher value indicates greater disruptive potential)
• k = Industry-Specific Constant (accounts for industry dynamics)
• T_new = Performance Improvement Rate of the New Technology
• T_est = Performance Improvement Rate of the Established Technology
• M_est = Performance Mismatch for the Established Market
• M_new = Performance Mismatch for the New Market

The disruption factor D depends on the relative performance improvement rates (T_new/T_est) and the relative performance mismatch (M_est/M_new) between established and new markets.

A real-world example outside the disk drive industry is India’s telecommunications sector, which was transformed by Reliance Jio’s disruptive entry in 2016. Before Jio’s arrival, established players like Bharti Airtel and Vodafone-Idea focused on incremental improvements to voice and SMS services while offering limited mobile data at high prices. Jio disrupted the market by providing affordable 4G data and free voice calls, targeting underserved rural and low-income segments.

Using the disruption equation:

T_new: Jio’s performance improvement stemmed from its extensive 4G infrastructure rollout and adoption of advanced data technologies.

T_est: Incumbent firms lagged in adopting 4G and prioritized optimizing older 2G and 3G networks.

M_est: Established firms offered services mismatched to market demand due to high data costs and limited rural coverage.

M_new: Jio aligned its offerings with customer needs for affordable data, free calls, and better rural connectivity, reducing performance mismatches.

Jio’s high (T_new/T_est) and favorable (M_est/M_new) ratio enabled it to disrupt the market. It captured market share rapidly and compelled competitors to lower prices and expand 4G services. Over time, Jio moved from serving niche segments to becoming a dominant market player. This case demonstrates how disruptive innovations allow new entrants to gain traction and challenge incumbents by better aligning technological advancements with emerging market needs.

[malvarezdemalde]

Smartphone Innovation

The history of the smartphone industry closely mirrors Christensen’s analysis of the rigid disk drive industry. Nokia and BlackBerry, once dominant players, were disrupted by new entrants that initially served a niche market but later redefined industry standards. This memo examines how these incumbent firms failed to respond to technological and market changes, ultimately leading to their decline.

In the early 2000s, Nokia controlled between 30% to 40% of the global mobile phone market. BlackBerry, with its enterprise-focused messaging capabilities, commanded over 50% of the U.S. smartphone market in 2009. Both firms catered to their core customer base (business users for BlackBerry and general mobile consumers for Nokia) by refining existing technology rather than exploring new user needs.

However, Apple’s 2007 launch of the iPhone and Google’s Android ecosystem introduced a paradigm shift. Initially dismissed as niche products, touchscreen smartphones with app-based ecosystems rapidly improved and captured mainstream markets. By 2013, BlackBerry’s market share had plummeted to around 3%, while Nokia, despite a last-minute pivot to Microsoft’s Windows Phone, saw its mobile division sold off in 2014.

Christensen’s theory suggests that incumbents often ignore disruptive technologies because they focus on serving existing high-end customers rather than emerging, lower-margin markets. In this case, BlackBerry’s reliance on physical keyboards and its slow adoption of touchscreen technology alienated consumers who prioritized app ecosystems and media consumption. Nokia’s Symbian OS struggled with poor developer support and usability compared to iOS and Android, making it unattractive to app developers and, in turn, consumers.

Between 2010 and 2015, the smartphone industry transformed. Apple’s iPhone shipments increased from 47.4 million in 2010 to 231.5 million in 2015. Android became the dominant OS, growing from around 1% market share in 2009 to almost 60% by 2014. By 2015, Samsung, Apple, and emerging Chinese brands (Huawei & Xiaomi) controlled 52% of the global smartphone market, while BlackBerry and Nokia had become negligible players. An analysis of total company sales from 2005 to 2015 further illustrates this shift (see graph attached below). Nokia’s sales peaked in 2008 and then steadily declined, while BlackBerry’s sales followed a similar trajectory with a peak in 2010 before plummeting. Meanwhile, Apple’s sales showed exponential growth, overtaking both competitors and continuing an upward trend.

This case underscores the necessity for firms to recognize and react to disruptive innovation early. Nokia and BlackBerry’s failures stemmed not from a lack of technological capability but from an inability to shift business models and prioritize emergent user needs. Companies today that seek to retain their market share in the future must remain vigilant against complacency in the face of changing technologies.

As demonstrated in the smartphone industry, Christensen’s framework is also relevant outside of the rigid disk drive industry. Nokia and BlackBerry were displaced not by direct competition but by new entrants that initially served niche markets and rapidly iterated to capture mainstream adoption. Understanding disruptive innovation dynamics is crucial for firms looking to maintain long-term competitive advantage in rapidly evolving industries.

[Graph data from FactSet]

[willowzhu]

The Rigid Disk Industry shares a figure on the relationship between cost per megabyte versus cumulative capacity produced of disk drives. The numbers show that every time cumulative output in the industry doubled, costs per megabyte declined to 53 percent of their former level. Christensen argues that the enormous decline in the cost per megabyte of memory was an important industry growth driver.

The relationship between cost and production/growth in the disk drive industry makes me curious as to what extent these patterns could show up in modern day AI ecosystems. In my question for Tuesday, I discussed the tension between incumbents and entrants in AI innovation, as big firms like Meta and Google are struggling against the disruptive ideas of smaller startups like Deepseek and OpenAI, and how this mirrors those same tensions between incumbent and entrants in disk drive innovation. Could it be that the relationships between cost and production/growth are parallel as well?

For today’s memo, I would like to theorize about the influence of policies, specifically new tariffs, on innovation in the semiconductor industry. Will a rise in production cost promote production or stifle it? Will it encourage more innovation or suppress it with reduced collaboration amongst countries?
To help answer this initial hypothesis, I propose a slightly more nuanced version of Schumpeter’s Creative Destruction Model that is two-fold: short term impacts and long term impacts of tariffs on innovation.

Short Run

Tariffs will increase production costs and disrupt pre-existing AI supply chains
P= old price of semiconductors
C= cost of production without tariffs
T= tariff rate
D= old demand for semiconductor chips
S= supply of semiconductor

New Price P’= C (1+T)
*this factors in the tariff distortion

The result of this is a decrease in demand, dependent on the price elasticity of demand
New Demand D’= D - ε (P’-P)
where ε = price elasticity of demand

Long run

We return to this profit maximizing function for entrepreneurs:

*but factor in a new price with the tariff distortion.

Similar to the analysis we did on innovation with politicians vs without politicians, I hypothesize that the tariffs would disrupt cost and production in the short run. However, in the long run, they could encourage more innovation and more profit, as when At increases over time, the benefits of operating the fixed cost technology will outperform the upfront investment of more expensive chips.

[mskim127]

Case Study: Tesla and its rise to the top

Tesla’s focus on vertical integration, software-centric design, and direct-to-consumer sales—coupled with its entry into underserved premium markets—mirrors patterns seen in Christensen’s study of disk drive innovators.

Large incumbents perfected the manufacture of internal combustion engine (ICE) vehicles over many incremental improvements. Tesla took a radically different approach by reconsidering the manufacturing processes of electric vehicles from the ground-up. By considering electric vehicles not as an ICE offshoot but as an entirely new architecture, Tesla was able to implement innovations that took advantage of characteristics unique to EVs.

I focus on three key innovations. First, was integrated hardware and software, a byproduct of fully electric drivetrains, which allowed Tesla to post over the air updates that enhanced performance post purchase and allowed them to create new revenue streams like subscription based FSD. Second, it designed its vehicles around a structural battery pack as opposed to retrofitting ICE designs. His allowed them to maximize energy density (and range) and interior space, two areas where EVs traditionally struggled to compete with ICE vehicles due to design constraints. Third, Tesla simplified production through the implementation of large single-piece castings (Gigacastings as they call them) which allowed them to reduce parts counts and reduce manufacturing complexity. Much like how specialized disk drive manufacturers where able to outperform industry giants through the implementation of novel technologies, these production innovations gave Tesla key advantages in terms of cost of production, reflected in their gross margins:

“The Rigid Disk Drive Industry: A History of Commercial and Technological Turbulence” finds that established disk drive leaders like IBM were initially successful due to their vertical integration, albeit eventually succumbing to more nimble competitors. Similarly, vertical integration is a key aspect of Tesla’s technology and cost efficiency. The company manufactures its own battery cells in Gigafactories and sells directly to consumers as opposed to competitors that operate the traditional dealership model. While disk drive firms disintegrated as the industry matured, Tesla’s vertical integration remains a strength, although exactly where this leads them is yet unclear.

However, Tesla is not exempt from disruption by competitors be it the likes of BYD or from incumbents. Scaling production while maintaining quality remains a challenge, as seen in early issues with Model 3 manufacturing. Furthermore, it seems the industry as a whole is ramping up in R and D spending with incumbents expected to invest $515 billion into EV production by 2030. While Tesla’s cost advantage and R and D spending remains intact, as Christensen’s paper notes, it is the direction of this R and D that will be crucial in determining the success of Tesla in the long run. Tesla’s focus on AI (Optimus robot, Dojo supercomputer) and energy storage (Powerwall) suggests it is heeding this advice, ensuring it is preempting disruption.

[jacksonvanvooren]

Patent Trends in Key Disk Drive Manufacturers

Introduction

Christensen’s “The Rigid Disk Drive Industry” explores many dynamics of the disk drive manufacturers who are incumbents and new entrants. He comes to a similar conclusion that we discussed in class–new entrants introduce the most disruptive innovations. These new entrants matured and displaced products of industry leaders.

In his paper, Christensen looks at metrics like the cost per megabyte or physical technological changes. I wanted to see if analyzing patent data would match his conclusions–especially that new entrants in the disk industry eventually take over and gain market share.

Methodology

I use patentsview.org data on US granted patents. Mostly, I analyse the g_patent data, which contains information like the patent ID number and the date filed. I merge abstract data so I can filter for disk specific patents, and I also include assignee data to sort by companies that are incumbents or new entrants.

I first do a time series analysis of patents filed by disk manufacturers mentioned in Christensen’s paper. The incumbents I include are IBM, Digital Equipment, Control Data, Data General, Burroughs, Fujitsu, Hitachi, and Univac. For new entrants, I filter for Wang, Prime, NCR, Nixdorf, Apple, Commodore, Compaq, Tandy, Sun Microsystems, Seagate, Conner, and Quantum.

I then filter the data by investigating patent abstracts for the companies above. I use the keywords disk, rigid, drive, storage, and magnetic to filter all patents with abstracts that contain these terms. I do the same time series analysis, counting the number of patents filed by incumbents versus new entrants.

Results and Discussion

This chart supports Christensen’s conclusion that new entrants are more innovative than incumbents, with a huge spike in new entrant patents starting in 2005 to now. We also see an overall decline in incumbent patent filings per year, further suggesting that these disk manufacturers have become less innovative. It is important to note though that new entrants can become incumbents, and disk manufacturers in the 80s may not be disk manufacturers now. Apple is a good example of both of these phenomena–most of their patents now are for iPhones and Macs, and since the 2010s, I would consider them an incumbent.

When filtering the patent abstracts, there seems to be an issue with a lack of data. I think that because I filtered both by company and abstract, the sample sizes become too small to make accurate conclusions. This might be amended by including a larger list of incumbents and new entrants outside of those in Christensen’s paper.

Though most of the companies analyzed started out at disk manufacturers, the ones that stayed successful have kept with the times. Apple, for example, switched to producing different types of hardware. IBM, which has pretty much always been an incumbent, has stayed innovative by embracing AI. In this case, a company’s innovativeness might be less of whether they are an incumbent or a new entrant, and more of how agile they are in acting upon new technology trends.

Code linked here.

[druusun]

This week’s readings on dynamic capabilities and strategic management highlight how firms sustain competitive advantages in rapidly evolving industries. The rigid disk drive industry serves as a case study demonstrating how companies with strong adaptive capabilities thrived while those unable to reconfigure their resources declined. This memo explores how firms navigate technological disruptions and presents a quantitative model illustrating the impact of dynamic capabilities on firm performance.

To quantitatively assess the role of dynamic capabilities, we model firm performance (​​) as a function of adaptation rate and technological disruption:

Where:
r = adaptation rate (higher for firms with strong dynamic capabilities, lower for those lacking adaptability).
D(t) = technological disruption factor, which either boosts or reduces firm performance based on its response strategy.

A graph modeling this equation would show that:

1. Firms with high adaptation rates experience exponential growth and capitalize on disruptions.
2. Firms with low adaptability initially grow but stagnate or decline when a technological shift occurs.
3. The disruption event (marked by a vertical line) creates a clear divergence between adaptive and rigid firms.

Policy and Strategic Implications:

1. Proactive Investment in Innovation
   • Firms must allocate resources toward R&D, workforce upskilling, and exploratory projects to remain resilient.
2. Encouraging Strategic Flexibility
   • Leadership should foster an organizational culture that prioritizes agility and rapid decision-making in response to emerging technologies.
3. Government Support for Innovation Ecosystems
   • Policymakers can facilitate adaptation by providing tax incentives for R&D and fostering public-private collaborations.

[yasminlee]

The electric vehicle industry is undergoing a shift from vertically integrated production models, exemplified by Tesla, to a modular manufacturing approach adopted by firms such as Hyundai/Kia, Volkswagen, Rivian, and General Motors. This memo examines how modular strategies are reshaping the competitive landscape of EV manufacturing, using insights from Christensen’s disruptive innovation theory and Teece’s dynamic capabilities framework. The attached market share trend graph (2022-2024) illustrates Tesla’s declining dominance as modular EV firms gain traction.
Tesla’s success was historically built on a highly integrated supply chain, where the company controls battery production, software development, and charging infrastructure. This strategy enabled early market dominance but has introduced scalability challenges as competition increases. In contrast, modular EV firms such as Hyundai/Kia, Volkswagen, and GM use shared battery platforms (ex: GM’s Ultium platform, Hyundai’s E-GMP architecture) and external suppliers to reduce costs and accelerate production.
Christensen’s disruptive innovation theory suggests that market leaders often struggle to adapt to architectural shifts, allowing new entrants to displace incumbents. Tesla’s vertical integration, while beneficial for quality control, may hinder its ability to scale as modular firms achieve cost efficiencies through specialization and supplier partnerships. This is shown in the graph, as Tesla's market share has decreased in the past few years, whereas other modular firms have grown in market share. GM specifically had the biggest increase in market share, largely because of the success of its Ultium battery platform. These trends highlight how modular approaches allow firms to adapt quickly to market demand, while Tesla’s vertically integrated model faces scalability constraints.
Teece’s dynamic capabilities framework argues that firms must continuously sense, seize, and transform their strategies to maintain competitive advantage. Tesla must decide whether to double down on vertical integration or adopt a hybrid model, outsourcing key components while retaining core software and battery innovations. The long-term market leader will likely be the firm that balances customization with cost efficiency. If Tesla fails to adapt, modular manufacturers could surpass it in global EV production, just as modular PC manufacturers overtook vertically integrated computer giants in the 1990s.

[JaslinAg]

The Price of Progress: A Model to Measure the Relationship between Monopoly Rights and Access or Human Welfare

In Tuesday’s class, Professor Akcigit emphasized the role of flexible markets. We discussed how the free-market design has a built-in mechanism that allocates effectively the system to be profitable. In contrast, planned economies struggle with the reallocation of resources. The choice of a market is a complex decision on the balance of welfare and efficiency. This ties into our discussions on monopoly rights in innovation. These monopoly rights are needed to incentivize firms to innovate, but they create a distortion that lowers the amount of intermediate goods.

To take a deeper look into this, I explored the pharmaceutical industry. Pharmaceutical companies produce a lot of patents, invest a lot of money into R&D, and tend to be highly innovative. Parallel to this increase in innovation, the monopoly rent charged to patriots has skyrocketed. There are many ethical considerations, as highlighted in Pharmaceutical patents: reconciling the human right to health with the incentive to invent. “Tensions arise when the right to health intersects with IP rights (IPR), and protecting the interests of the inventors of scientific creations are balanced against rights to access science and rights to health.”

To visualize this, I propose a simple model:

if ($A_t > A*$) and ($C_t < C*)$:
H = True

Where:
$H =$ happiness, in this scenario a patient is happy if they can access care.
$A*$ represents the needed productivity to have the innovation needed to provide care
$C*$ represents the max a customer can pay to access care

From our model on innovation, we know that based on a firm’s choice of $\mu$ there is a probability $\mu$ that $A_{t+1} = \gamma A_t$
Additionally, $C_t$ depends on the monopoly markup a firm can charge. In our model, this monopoly price is $\frac{1}{\alpha}$. However, in the real world, the markup charged by pharmaceutical companies is related to the quality of the innovation, which is represented by $A_t$ in this model.

An expansion of this model can scale happiness by the difference between $A*, C*$ and $A_t, C_t$.
If $A* < A_t$, the quality of the medication is increased which increases happiness.
If $C* > C_t$ the price of the medication is less than the maximum amount the patient can pay which also increases happiness.

[henrysuchi]

In this memo, I discuss the implications of Akcigit et al (2025) in the context of the Chinese economy. Drawing on data from the World Bank, it is easy to identify from looking at GDP per capita (as I have plotted below) that Chinese growth was dramatic in the 1990s and then slowed down very recently. It is worth noting the timing of this growth: it occurred right when Chinese Communist Party took specific steps to liberalize markets and entered the WTO.

If total economic output is canonically modeled as $Y = (AL)^\alpha K^{1-\alpha}$, then we must look to the available data to determine the causes of the Chinese slowdown. While labor markets are not the strongest in China right now (in fact, it notoriously has a difficult job market for new entrants), the timing of the slowdown in labor growth does not in itself suggest that it is the primary cause for the slowdown. Indeed, while the labor force overall has stopped growing, see the graph, this may very well be endogenous to other forces shaping labor demand. See below the Chinese labor force size over time, also using the World Bank’s data.

Given this, we should also consider the origins of this slowdown. For one, it may be that there has been a crash in capital that has resulted in stagnating labor demand. However, using data from the St. Louis Fed, it seems unlikely that this is the case, as capital stock has actually continued to grow in China, even during the 2008 Recession. It is also unlikely that there is something wrong with the labor supply; China has reversed its one-child policy, and education in China is notoriously (arguably over-, if not at least) well-supplied, meaning the issue is not a human capital deficit either. This therefore raises the point of technology.

Confusingly, patents do not indicate a huge technological slowdown in China. If anything, a quick perusal of publicly available analyses of patents indicates that it is quite the opposite. Analysts from Our World in Data plot the patent applications over time for all countries and find China overwhelmingly beating everyone else. This is intuitive given the high levels of human capital stock and public investment in science in China. However, perhaps there is not enough willingness among Chinese firms to adopt new technologies that diverge from the pack, the key to innovation. This may be interpreted in terms of the differences between market and planned economies discussed by Akcigit et al (2025). The reason why the United States surpassed the Soviet Union in innovativeness by the end of the Cold War was its diffuse economy and its free markets which incentivized innovation without freezing over other possibilities. Given the incredible pressure that China places on businesses and businessmen, this could very well be the cause of its growth slowdown.

[amulya-agrawal]

Industry Disruption & Business Dynamism with Walmart and Amazon

AI has driven significant creative disruptions in supply chain logistics for many companies historically. We saw this in the rigid disk industry -- an industry that evolved rapidly with new product architectures and emerging market segments that led to the displacement of established firms like IBM in the "Rigid Disk Drive Industry" text. Similarly, Walmart -- once the dominant force in retail logistics -- is now playing catch-up against Amazon's AI-driven fulfillment network. This case illustrates how AI-driven operational efficiencies can serve as disruptive innovation -- altering industry leadership and market domination. By integrating disruptive innovation theory and the dynamic capabilities model, this memo explores the strategic responses required for Walmart to compete in an AI-driven logistics landscape.

In the "Dynamic Capabilities and Strategic Management" text, the dynamic capabilities framework explains how firms develop and sustain competitive advantages in fast-changing environments. A large firm like Walmart must be willing to continuously transform, reconfigure, and leverage their resources effectively, but they struggled to adapt when technology and markets evolved -- being a victim of market myopia, as they focused on current customers rather than emerging ones.

For Walmart specifically, much like the incumbent disk drive manufacturers, they focused on incremental improvements to their existing supply chain -- underestimating the transformational power of AI. Meanwhile, Amazon leveraged AI-driven robotics, real-time data analytics, and machine learning algorithms to optimize fulfillment, reduce costs, and redefine retail logistics. All of this was done through understanding their emerging customers' needs -- allowing them to become a market dominator.

Amazon deployed AI-powered Kiva robots in their warehouses -- reducing fulfillment time from 60 minutes to 15 minutes per order. Along with this, AI-driven demand forecasting based on deep learning algorithms reduces overstocking and stockout -- improving efficiency and lowering costs. And, along with these, Amazon utilizes machine learning to analyze delivery routes in real-time -- lowering fuel costs and expediting delivery speeds (such as through their Prime delivery program).

Recently, Walmart invested in AI-powered robotics and autonomous delivery, but they lag behind Amazon in AI adoption. Unlike Amazon's data-driven inventory systems, Walmart's reliance on brick-and-mortar retail creates structural limitations in AI-driven logistics -- mirroring the rigid disk industry where incumbents failed to recognize new architectures until disruptors had already captured market share.

Applying the Dynamic Capabilities Framework, firms sustain competitive advantages by sensing, seizing, and transforming capabilities in response to change. For "sensing," when Amazon recognized AI's opportunity in automated warehousing, demand forecasting, and last-mile delivery early on, Walmart underestimated AI's role, but they later invested in AI-powered inventory tracking. For "seizing," Amazon acquired Kiva Systems (robotics) to develop in-house AI logistics software. Walmart later partnered with AI startups (like Symbolic DroneUp) to introduce automated warehouses and drone delivery. For "transforming," amazon reconfigured their business model by shifting from retail focus to AI-driven logistics leader by positioning AWS to offer AI supply chain solutions to other retailers. Walmart, on the other hand, invested in store-to-home AI-driven fulfillment and autonomous tracking, but they were playing catchup to not just Amazon, but also companies like Instacart that had already done this.

For the custom analytical element, I visualized how AI adoption impacts logistics efficiency over time. Amazon, as an early adopter, has compounded its AI-driven efficiencies, creating a growing competitive gap over Walmart, a late adopter. Thus, firms that fail to understand emerging customer needs and develop in-advance of their competitors will fall behind.

[diegoscanlon]

talking a lot but saying little

On open source
Mainly reconstructing the arguments made by Zuck here. It’s helpful to understand the two parties involved – Meta, developers – and differentiate incentives as such.

Developers – main values seem to be open ecosystem where innovation (product) is not restricted, and application specific models (want more control for better performance). Other values include data protection, efficiency and affordability, standard for the long term.

Meta – mainly value open ecosystem / no restrictions to building (previously restricted by Apple).

One way for Meta to ensure that it can build whatever it wants is to build its own model. But if Meta is the only entity using this model, Zuck is making the assumption that it wouldn’t grow as fast as it could – that opening the development of their models (which started off behind) could allow them to accelerate more quickly (compares the development and incentives of closed-model to closed variants of Unix). This acceleration is caused by the alignment of incentives between developers and Meta – developers want open source for autonomy and quality, and would therefore be willing to help improve the model they use.

One argument may be that open source decreases Meta’s competitive advantage, which implies some impact on profits. From a technical perspective, Zuck expects models to be extremely competitive, which implies any advancements for one model will likely be replicated / matched quickly after by other models; hiding advancements doesn’t give you an advantage for very long. From a business perspective, Zuck does not see selling access to Llama as Meta’s business model, unlike OpenAI or other providers. Open source models are cheaper, and he assumes developers will use cheaper models. Comparability of open and closed models make this assumption stronger.

These two ideas (business model, technical advantages) seem somewhat related: if we believed that technical advantages would lead to higher revenue because other systems aren’t as advanced, then we would want to protect technical advantages. The degree to which we would protect these advantages would increase as our revenue is more dependent on selling access to these closed models. However, if we believe that both technical advantages aren’t real, and revenue is not dependent on selling access to the model (making us indifferent towards marginal model strength), then we might be inclined to capture the benefits of open source – there are many more positives gained from open source than closed source.

Mistral gives us the opportunity to observe the conversion of an open source company to a closed one (and the timeline of events might give some insight into the thesis behind the switch). Mistral launched Mistral Large in Feb 2024, which was their first release of a closed source (unclear if the weights are still open, but broadly seems to categorize as closed source) was closely followed by an investment by Microsoft (the round, more like a bridge round, was only $15m after raising at a $450m valuation by a16z). With the launch of Le Chat, we might interpret Mistral to be putting itself into competition with OpenAI and ChatGPT, implying its business model is dependent on selling access to its models, which might increase its sensitivity to secrecy of technical advantages.

Thus, we might want to differentiate the types of companies that are developing models, categorizing by the relationship between the LLMs and their business models. OpenAI, Mistral seems to make money by selling access to the model. Meta (though unclear) seems to make money by applying their model to other parts of their business (Instagram, glasses, assistant, advertising).

However, there seems to be some differentiation to be made between cost of usage and access to code as it relates to open source; paywall doesn’t necessarily mean a model is closed source. If Meta’s goal is to have developers use the model now, then decreasing costs must lead to increased usage. Charging for usage, even if still open source where developers can benefit from an open ecosystem and custom specificity, might still decrease incentives to build on the platform. Thus, cost and access to code are two different incentives as well. The vagueness of what is open source makes unclear the argument that as the model becomes closer to AGI, Zuck will charge for it (because why would you give people an employee for free), but does that mean it's closed source?

However, it’s also important to differentiate between developers and researchers. Having this non-secrecy (represented by open source), Meta might retain more talent (researchers) – it seems that in the past and present, Google’s secrecy policies has caused researchers to leave the company (some of whom start more non-secrecy companies, like Mistral; others start startups, like character.ai). Maybe something to do with researchers publicly getting acknowledged for their contributions at Meta too.

source