The Modern Flywheel: The Design of Infinite Growth and Competitive Warfare

Introduction: Reinterpreting the Flywheel in 2025

Re-examining the Archetype

The ‘Flywheel’ concept, popularized by Jim Collins, has become the most powerful business model to explain the phenomenal growth of Amazon. The core of this classic model lies in a virtuous cycle of mutually reinforcing elements. Lower prices create a better Customer Experience, which drives more Traffic to the platform. Increased traffic attracts more third-party Sellers, which expands the breadth and depth of product Selection. This entire process lowers the Cost Structure through economies of scale, which in turn enables even lower prices, accelerating the flywheel’s momentum. This model is more than just a growth strategy; it represents the essence of a self-reinforcing system where the business model itself becomes the engine of growth.

The 2025 Proposition: Data is the New Axle

As of September 2025, the concept of the flywheel has undergone a fundamental evolution. The most powerful flywheels today are no longer confined to a virtuous cycle of operational efficiency. The modern flywheel is redefined as an ecosystem with data as its axle, accelerated by artificial intelligence (AI). The momentum of this new model is often amplified by aggressive strategies that push legal and ethical boundaries. This report aims to provide an in-depth analysis of how these modern flywheels operate and the competitive strategies hidden behind them.

A Strategic Redefinition of ‘Foul Play’: Accelerants and Moats

The ‘foul play’ mentioned by users must be understood not as simple unethical behavior, but as part of a sophisticatedly designed competitive strategy. This report redefines it within the framework of ‘Strategic Accelerants and Competitive Moats.’ These are actions intentionally executed for three purposes: (1) to Kickstart the flywheel at an unnatural speed beyond what organic growth allows, (2) to exponentially amplify the flywheel’s rotational speed, and (3) to build insurmountable barriers that competitors cannot enter. Through this analysis, we will uncover the reality of how modern companies pursue sustainable, or near-infinite, growth.

Part I. The AI-Native Data Flywheel: The New Apex Predator

The most powerful flywheel of our time is a self-reinforcing loop of user engagement, data acquisition, AI model enhancement, and service improvement. This model creates a compounding data advantage that is nearly impossible for latecomers to catch up to.

1.1. Dynamics of the AI-Data Virtuous Cycle

The core loop of the AI-Data Flywheel consists of four stages. First, an enhanced user experience, realized through AI-powered hyper-personalization, leads to, second, increased user engagement and growth. This becomes the source for, third, the creation of vast and proprietary datasets. Finally, this data is used for, fourth, the training and enhancement of AI models, which in turn refines the user experience from the first stage, completing the cycle.

The true power of this flywheel lies in its compounding effect. With each rotation, not only does the service improve, but the data gap with competitors—the ‘Data Moat’—widens exponentially. An AI model trained on 1 billion interactions performs on a qualitatively different level than one trained on 1 million, creating a non-linear competitive advantage. As of 2025, this flywheel is the core engine driving the most important business trends, including hyper-personalization, AI agents, predictive analytics, and AI-powered marketplaces.

This structure creates a ‘Data Gravity’ effect. As a platform accumulates more data, it exerts a gravitational pull on users with better services and on partners with richer insights, drawing the entire ecosystem into its orbit. The attempt to artificially generate this gravity is the essence of the ‘strategic accelerants’ we will discuss next.

1.2. Strategic Accelerant: Aggressive Data Acquisition (Data Scraping)

The biggest challenge every AI-Data Flywheel faces is the ‘cold start’ problem—the lack of initial data required for model training. Data scraping is a prime example of a ‘foul play’ tactic used to solve this problem.

Case Study: Yanolja vs. Yeogi Eottae

This legal dispute clearly illustrates how data scraping is used as a competitive weapon. Starting in 2015, Yeogi Eottae used crawling technology to illicitly collect data from its competitor Yanolja, including lists of affiliated accommodations, addresses, and pricing information, for its internal sales operations. The court rulings in this case split in two, offering important insights for strategists.

In the criminal trial, the defendants were acquitted of violating the Information and Communications Network Act. The court ruled that Yeogi Eottae’s actions were not ‘hacking’ that directly infiltrated Yanolja’s servers, but rather the technical collection of information already public to general users. In other words, the act of scraping technically accessible information was not deemed a crime in itself.

However, in the civil trial, a violation of the Unfair Competition Prevention Act was recognized, resulting in a 1 billion KRW damages award. The court found that Yeogi Eottae had gained an unfair advantage by using the database, which Yanolja had built through significant investment and effort, in a manner contrary to fair trade practices. This clarified that systematically misappropriating a competitor’s core, refined data asset for direct commercial use constitutes an ‘act of unfair competition.’

The difference between these two rulings exposes a key legal gray area in the age of data-driven competition. While accessing public data itself may not be a crime, systematically siphoning off a competitor’s curated data asset—built at great expense of time and money—for commercial gain can be considered a clear illegal act. This is because it is an attempt not just to steal information, but to replicate the ‘accumulated momentum’ of a competitor’s flywheel.

1.3. Strategic Accelerant: Algorithmic and Behavioral Manipulation

To artificially accelerate the ‘user engagement’ phase of the data flywheel, platforms often employ algorithmic strategies that manipulate user behavior and perception.

The Dark Side of Engagement Optimization

Recommendation algorithms on platforms like YouTube and TikTok are designed to maximize user watch time. To achieve this, they often tend to amplify politically biased, sensational, or extreme content. Such content effectively drives user engagement by eliciting strong emotional responses, but it has the side effect of deepening ‘Filter Bubbles’ and ‘Confirmation Bias’ in society. This is a strategic choice to increase the flywheel’s speed at the expense of social cost.

The Illusion of Control

What is more insidious is that platforms give users the illusion of control when, in reality, they have none. A 2022 study by the Mozilla Foundation found that YouTube’s ‘Dislike’ or ‘Not Interested’ buttons have almost no effect on the actual recommendation system. This suggests an intentional design where the platform’s goal of maximizing engagement takes precedence over explicit user feedback. This allows the platform to maintain the flywheel’s stable momentum while minimizing the risk of user churn.

‘Shadow Banning’

‘Shadow banning,’ a controversial practice on platforms like TikTok, is another form of algorithmic control. It involves secretly restricting the visibility of a specific creator’s content without their knowledge. This allows the platform to effectively suppress content that does not align with its commercial or policy interests, and to steer the entire ecosystem in a direction most favorable for accelerating its flywheel. This entire process occurs without transparent accountability, strengthening the platform’s absolute power.

1.4. Flywheel Risk: Data Monetization and Ethical Collapse

As powerful as it is, the data flywheel carries fatal risks. Especially in models built on user trust, when that trust is broken, the flywheel can grind to a halt or even spin in reverse.

Case Study: The Fall of 23andMe

The business model of the direct-to-consumer (DTC) genetic testing company 23andMe was a classic example of an AI-Data Flywheel. (1) Affordable genetic testing kits attracted (2) millions of users, who provided (3) vast and valuable genetic and health data. This data was then (4) monetized through new drug development partnerships with pharmaceutical giants like GlaxoSmithKline (GSK).

However, this flywheel collapsed under two fatal shocks. The first was a massive hack in 2023 that resulted in the data breach of about 7 million customers. The second was the saturation of the genetic testing market. The company eventually filed for bankruptcy in March 2025.

The crux of this case is that during the bankruptcy proceedings, the court ruled that user data could be considered a company ‘asset’ and sold off. This was a ruling that shook the very foundation of the data flywheel. Users may have consented to their data being used for a specific company’s ‘research purposes,’ but they never agreed to it being sold like a commodity to the highest bidder in an auction after the company went bankrupt. This case starkly revealed how fragile the ‘implicit contract’ between users and companies is, and how deep the ethical abyss of data ownership runs.

Ultimately, competitive advantage in the AI era comes not from the algorithm itself, but from the proprietary, real-time ‘flow’ of data that continuously trains it. This means the strategic focus must shift from building better algorithms to designing loops that effectively capture user data and drive engagement. It is at this point that ‘foul play’ like behavioral manipulation emerges, and where mishandling that data leads to catastrophic risks, as seen with 23andMe.

Part II. The B2B SaaS Ecosystem Flywheel: The Design of Lock-in

B2B SaaS (Software-as-a-Service) companies move beyond selling single products to building multi-product ecosystems that customers cannot escape, creating a flywheel of exponentially increasing customer lifetime value (LTV).

2.1. Dynamics of the Ecosystem Flywheel

The B2B SaaS Ecosystem Flywheel has a four-stage cycle. It begins with (1) Core Product Adoption that solves a specific key problem. As this product becomes deeply embedded in the customer’s business processes, (2) Deep Workflow Integration occurs, turning the software into an irreplaceable ‘System of Record.’ This creates opportunities for (3) Multi-Product Ecosystem Upsell, where customers adopt additional related solutions from the same vendor. As this process deepens, it ultimately results in (4) Prohibitively High Switching Costs, perfectly ‘locking in’ the customer. The stable revenue generated from this is then reinvested into product development, strengthening the first stage of the flywheel.

The ultimate goal of this strategy is to elevate the software from a selectively used ‘tool’ to an indispensable ‘infrastructure.’ When budgets are cut, single-function solutions are the first to go, but an integrated ecosystem responsible for a company’s core operations remains essential. B2B software purchasing decisions are driven by return on investment (ROI) and operational efficiency. Since using a single integrated platform significantly reduces complexity, data silos, and vendor management costs compared to managing dozens of individual tools, the ecosystem strategy itself has a powerful value proposition.

2.2. Strategic Accelerant: Designed Dependency (Vendor Lock-in)

Vendor lock-in is not an accidental phenomenon but a strategic accelerant intentionally designed to prevent customer churn and ensure long-term revenue.

Technical Lock-in: Vendors use various technical mechanisms to create high switching costs. These include proprietary APIs that are incompatible with other platforms, unique service integration methods, custom data formats that are difficult to export, and configurations optimized only for a specific platform. This makes migrating to a competitor’s product not just a simple software swap, but a complex redesign project requiring enormous cost and time.

Contractual Lock-in: Lock-in is also reinforced on the commercial side. Multi-year contracts with steep discounts, auto-renewal clauses, and tiered pricing models that impose financial penalties for downgrading or churning based on past usage are common examples. These contractual structures make switching financially unfeasible even if a customer finds a better alternative.

Human Lock-in: When an organization’s members become deeply accustomed to a specific tool’s interface and workflows, ‘process and user experience lock-in’ occurs. Switching to a new tool entails a drop in organization-wide productivity and massive retraining costs, leading management to prefer the status quo.

2.3. Strategic Accelerant: Bundling and Tying as a Weapon

Bundling and tying are among the most powerful ‘foul play’ tactics a dominant firm can use to distort a competitive market by leveraging its market position.

Case Study: Microsoft 365 & Teams vs. Slack

This is the clearest example of a tying strategy in modern business. Microsoft leveraged its overwhelming dominance in the office software market (the ‘tying product,’ Office 365) to include its collaboration tool, Teams (the ‘tied product’), in the bundle, effectively offering it for free.

Antitrust regulators argue that this practice deprives competitors like Slack of the opportunity to compete fairly on the merits of their products alone. Many corporate customers choose Teams not because it is superior, but because it is ‘included’ with the essential Office 365 they already have to use. This is an act of unfairly extending monopoly power from one market (office software) to another (collaboration tools).

Microsoft’s main defense is based on ‘economies of scope’ and ‘increased consumer benefit’—that is, by integrating multiple services and offering them at a lower price, they provide greater value to consumers. However, regulators argue that this practice ultimately harms competition, stifles innovation, and reduces consumer choice in the long run.

2.4. Strategic Accelerant: ‘Killer Acquisitions’

‘Killer acquisitions’ are when a dominant firm acquires a nascent competitor not for its immediate revenue, but to eliminate a potential future competitive threat. This is the most direct way to cut off a competitor’s flywheel before it starts spinning in earnest.

Case Study: FTC vs. Meta (Acquisition of Instagram & WhatsApp)

The core of the Federal Trade Commission’s (FTC) antitrust lawsuit against Meta (formerly Facebook) is the claim that Meta’s acquisitions of Instagram and WhatsApp were made with the clear purpose of eliminating significant threats to its social networking monopoly. The argument is that Meta maintained its monopoly not by competing on product merit, but by using its vast capital to buy out competitors.

Such acquisitions remove potential innovators from the market and make venture capitalists hesitant to invest in competing startups (since the most likely exit strategy is to be acquired by a giant). Consequently, this slows down innovation across the entire market and limits consumer choice.

Competition in the modern B2B SaaS market is no longer a battle between individual products, but a war between ecosystems. This makes integration and platformization a strategic imperative. A startup can no longer succeed simply by building a ‘10x better’ product. It must either integrate into an existing ecosystem, like Salesforce’s AppExchange, or have a strategy to build its own defensible ecosystem from the start. This dramatically raises the barrier to entry and creates a structure that is overwhelmingly favorable to established players with massive capital.

Part III. The Creator Economy Flywheel: Monetizing Influence and Community

The creator economy is based on a unique flywheel that grows through the interaction of three pillars: the platform, the creators, and the audience. The platform’s key role here is to facilitate the creation and consumption of content and to monetize the value generated in the process.

3.1. Dynamics of the Content and Community Flywheel

The creator economy flywheel consists of a five-stage virtuous cycle. (1) Attracting talented creators (by offering creation tools and the promise of monetization) leads to (2) the production of diverse, high-quality content. This content then (3) attracts and retains a large, engaged audience. The audience’s attention is converted into (4) platform revenue through advertising, subscriptions, etc., and a portion of this revenue is (5) distributed as creator compensation, which in turn becomes the driving force to attract more creators.

In 2025, the key accelerator for this flywheel is AI. Platforms are investing heavily in AI-powered creation tools that assist with idea generation, video editing, graphic design, and more. This directly strengthens the ‘content production’ stage of the flywheel by lowering the barrier to entry for content creation and simultaneously increasing both output and quality.

3.2. Strategic Accelerant: Asymmetric Power Structures

Behind the creator economy flywheel lies an asymmetric structure where the platform holds all the power. While it appears to be a symbiotic relationship on the surface, it is filled with ‘foul play’ designed to maximize the platform’s profits.

Opaque Algorithms and Revenue Sharing

Platforms have total control over the recommendation algorithms that determine a creator’s success, but the workings of these algorithms are a complete ‘black box.’ Furthermore, revenue-sharing models are complex, opaque, and can be unilaterally changed by the platform, keeping creators in a state of constant dependency.

Case Study: The ‘Pro-rata’ System in Music Streaming

The ‘pro-rata’ system adopted by most music streaming platforms is a prime example of this asymmetric power structure. This method pools all paid subscription fees and distributes revenue based on the share of total streams each song accounts for. This structure inevitably leads to a ‘winner-take-all’ phenomenon where a small number of superstars receive the majority of the revenue.

The bigger problem is that this system is extremely vulnerable to abuse. ‘Streaming parties,’ where fandoms organizationally repeat-stream to boost a specific artist’s rank, or ‘chart manipulation’ through brokers, exploit the loopholes of the pro-rata system to distort charts and disrupt the revenue distribution structure. While this has the effect of increasing total streams and engagement for the platform, the vast majority of independent and niche artists receive negligible earnings, often less than a fraction of a cent per stream, which suffocates the creative ecosystem. The ‘user-centric model,’ proposed as an alternative, would distribute each user’s subscription fee only to the artists they actually listened to, potentially solving this structural problem. However, its adoption is slow as it directly challenges the existing power structure of the platforms.

3.3. Strategic Accelerant: Predatory Self-Preferencing

‘Self-preferencing’ is when a dominant platform uses its control over its core service (e.g., search, social feed) to give an artificial advantage to its other new services or products, putting competitors at a disadvantage.

Case Study: EU vs. Google Shopping

The European Commission (EC) fined Google €2.42 billion for abusing its dominance in the search market to unfairly display its own price comparison service, ‘Google Shopping,’ at the top of search results while demoting competing services. This is a classic example of leveraging a powerful existing flywheel (search traffic) to unfairly kickstart a new one (commerce).

This logic applies directly to creator platforms as well. For example, when Instagram launched ‘Reels,’ it heavily promoted Reels content in the main feed and explore tab. This came at the expense of visibility for traditional photo posts or content shared from competing video platforms. This strategy leveraged Instagram’s massive existing user base to quickly activate the new feature and effectively fend off competitors like TikTok and Snapchat within its own ecosystem.

A creator economy platform is not a neutral marketplace, but a ‘centrally planned economy’ where the platform acts as the government. The platform uses the tools of algorithms and economic policies to orchestrate the behavior of creators and consumers to maximize the platform’s overall ‘Gross Domestic Product’ (GDP), i.e., total revenue. In this structure, a creator’s success is a means to achieve the platform’s goals, not the ultimate goal itself. In the long run, the sustainability of this flywheel depends on how it manages creator churn. If the power imbalance becomes too exploitative, creators will seek to diversify their income streams outside the platform (e.g., newsletters, paid memberships), which weakens the platform’s core asset: its content supply. This creates a strategic dilemma for the platform, which must extract enough value for profitability while leaving enough compensation to prevent a mass exodus of creators.

Part IV. Regulatory Headwinds and Strategic Risks

No flywheel can spin forever without friction. There are external (regulatory) and internal (strategic) forces that can slow down, stop, or even reverse a powerful growth flywheel.

4.1. Regulatory Headwinds: The Counter-Flywheel

As the ‘foul play’ strategies of giant platforms became rampant, regulatory authorities around the world began to activate a powerful ‘counter-flywheel’ to control them.

The European Union’s Digital Markets Act (DMA)

The DMA is a regulation that directly targets the unfair practices of giant platforms designated as ‘gatekeepers.’ The key provisions of this law are designed to neutralize the strategic accelerants discussed earlier.

Ban on Self-Preferencing: Explicitly prohibits practices like the Google Shopping case.
Ban on Tying: Regulates bundling strategies like Microsoft Teams.
Ensuring Interoperability and Data Portability: Prevents gatekeepers from building data silos and mandates that they allow users to easily transfer their data to other services or interact with competing services. This is a direct measure to weaken the data gravity effect and vendor lock-in.

Aggressive Antitrust Enforcement in the U.S.

The U.S. Department of Justice (DOJ) and the Federal Trade Commission (FTC) are also simultaneously pursuing lawsuits against Big Tech companies with a similar problem awareness as the DMA.

DOJ vs. Google: Focuses on monopoly contracts in the search market and tying allegations in the ad tech market.
FTC vs. Meta: Defines the acquisitions of Instagram and WhatsApp as ‘killer acquisitions’ and is demanding corporate divestiture.
FTC vs. Amazon: Takes issue with anti-competitive practices, such as penalizing sellers who offer lower prices on other platforms than on Amazon’s marketplace.

Domestic Platform Regulation Trends in Korea

South Korea’s Fair Trade Commission (KFTC) is also strengthening its investigations into the ‘gapjil’ (abuse of power) practices of domestic platform companies like Coupang and Naver. Key regulatory targets include Coupang’s ‘dynamic pricing’ policy that forces sellers to supply products at prices lower than competitors, unilateral contract terminations, and unfair terms and conditions such as the unauthorized use of seller content.

These regulatory movements act as a flywheel in themselves. As regulatory bodies accumulate successful cases against Big Tech, this leads to legal precedents, political momentum, and institutional expertise. As this ‘regulatory flywheel’ accelerates, future enforcement is likely to become faster, broader, and more powerful. Companies are facing an era where they can no longer dismiss fines as a ‘cost of doing business.‘

4.2. Internal Collapse and the Flywheel’s Reversal

A flywheel can collapse not only from external shocks but also from internal factors.

Negative Network Effects

When a platform grows beyond a certain size, overcrowding can actually degrade the user experience. This happens when social media feeds are filled with ads and spam, or when marketplaces are flooded with low-quality sellers and counterfeit goods. This erodes the ‘positive customer experience,’ the core driver of the flywheel, causing user churn and causing the flywheel to spin in reverse.

Strategic Complacency and Failure to Adapt

A successful flywheel can, paradoxically, become the biggest cause of organizational complacency.

Case Studies: Nokia & Yahoo: In their respective eras, they both possessed formidable flywheels but collapsed because they failed to adapt to fundamental market shifts—smartphones and sophisticated search algorithms. They made the mistake of continuing to push their old flywheels while the world was moving to a new kind of flywheel. This teaches us that a flywheel is only effective when it is synchronized with the fundamental assumptions and dynamics of the market.

Unsustainable Tokenomics

Case Study: Axie Infinity (P2E Game): The flywheel of the Play-to-Earn (P2E) game Axie Infinity was built on a purely speculative structure of ‘Play → Earn Tokens → Cash Out.’ For this model to be sustained, there had to be a constant influx of new users paying more money than existing users were cashing out. When the influx of new users slowed, the token price plummeted, and the entire ecosystem entered a death spiral. This is an important case showing that a flywheel with no intrinsic value creation source other than speculation is bound to collapse.

4.3. Summary of Flywheel Models, Accelerants, and Counter-Strategies

The key points discussed in this report are summarized below. This table provides an at-a-glance view of each flywheel model’s core driver, the ‘foul play’ strategies that accelerate it, and the corresponding risks and regulatory responses.

Flywheel Model	Core Virtuous Cycle	’Foul Play’ Strategic Accelerants	Representative Cases	Key Risks / Failure Factors	Regulatory Response
AI-Native Data	User Engagement → Data Generation → AI Enhancement → Service Improvement	Data Scraping, Algorithmic Manipulation, Opaque Data Monetization	Google (Search), YouTube (Recommendations)	Ethical Collapse / Loss of Trust (23andMe)	Data Privacy Laws (GDPR), Algorithmic Transparency Mandates (DMA)
B2B SaaS Ecosystem	Core Product → Workflow Integration → Ecosystem Expansion → High Switching Costs	Vendor Lock-in (Technical/Contractual), Tying, Killer Acquisitions	Microsoft (Office/Teams), Salesforce (AppExchange)	Disruption by Open Standards / New Architectures	Antitrust Investigations into Tying and Acquisitions (FTC/DOJ, DMA)
Creator Economy	Creator Attraction → Content Production → Audience Growth → Revenue Generation → Creator Compensation	Opaque Algorithms & Revenue Sharing, Predatory Self-Preferencing	YouTube, TikTok, Music Streaming Platforms	Creator Churn / Diversification of Revenue Outside Platform	Ban on Self-Preferencing (DMA), Demands for Revenue Model Transparency

Conclusion: Designing a Defensible and Sustainable Flywheel for the Next Decade

This report has provided an in-depth analysis of the flywheel models at the core of modern corporate growth strategies and the aggressive competitive tactics hidden behind them. The analysis has made it clear that the most powerful flywheels often gain their momentum through ‘strategic accelerants’ that push legal and ethical boundaries.

The Strategist’s Dilemma

Business leaders in 2025 face a significant dilemma. They must balance the immense competitive advantage offered by aggressive, boundary-pushing flywheel strategies against the exponentially increasing regulatory and reputational risks that accompany them. A purely ‘clean’ flywheel may be too slow to survive the competition, while a flywheel that relies solely on ‘foul play’ is a ticking time bomb.

Recommendations for the Future

Therefore, to build a sustainable competitive advantage for the next decade, one must design a flywheel that is not just fast, but also defensible and resilient. The key strategic principles for this are as follows:

Define Your Core Value Loop: You must be ruthlessly clear about the fundamental virtuous cycle that drives your business. Analyze successes and failures to find your organization’s own repeatable formula for success.
Consciously Choose Accelerants and Manage Risks: Consciously decide which ‘strategic accelerants’ to use, and explicitly analyze and prepare for the legal, ethical, and reputational risks of each. Recognize that every strategy comes at a price.
Anticipate Regulatory Headwinds and Incorporate them into Your Design: Proactively internalize the spirit of regulations like the DMA into your business model before being forced to by regulators. Building systems with interoperability and data portability in mind can turn regulatory compliance from a cost into a symbol of trust and a competitive advantage.
Monitor for Signs of Friction and Decay: A flywheel is not a perpetual motion machine that runs on its own after a strong push. You must constantly monitor and eliminate ‘friction’ factors that impede the flywheel’s rotation, such as negative network effects, strategic complacency, and the weakening of core drivers.

The ultimate goal is not just to create a fast-spinning flywheel, but to build one that is made to last—one that continues to create value without stopping, despite external shocks and internal corrosion. This will be the core competency of winning companies in the competitive landscape of 2025 and beyond.

One Person Unicorn