Is AI the Next "Internet-Scale Shift"? Evans' "Three-Stage Strategy for AI Utilization" and Future Predictions
Benedict Evans is a strategic analyst who has been tracking trends in the tech industry for many years. In his 2025 presentation, "AI Eats the World", he centers on the question of whether AI will be the next platform shift in the history of technology or merely an evolution, providing a broad overview that includes historical context, actual usage, corporate strategy, and future predictions.
1. What kind of "platform shift" is AI?
1-1. The historical paradigm of technology
Evans points out that while AI (specifically generative AI/LLMs) shares the same structure as existing innovations, citing past platform shifts like the internet and smartphones, it is unique in that its physical and theoretical limits are unknown. For example, with the internet and mobile, the limits of "how fast can it get" or "how small can it get" were clear, but with AI, the fact that "we don't know how much better it can get" is fundamentally different from past shifts.
This perspective leads to the implication that "AI is not just a product, but can become a new infrastructure."
1-2. The definition of AI and the transition of perception
Regarding the question "What is AI?", Evans observes that in general usage, the term AI is only applied to "new and unknown technologies." In other words, as things like machine learning and databases become widely used, the term "AI" disappears and they become "just features." This is his famous paradox: "AI disappears as a feature."
This is consistent with the history of past technological innovations becoming "taken-for-granted features" (de-thematization).
Example)
Automatic elevators are no longer "AI-powered" but have become "normal elevators"
Databases and automatic backup functions were also novel at first, but are now standard features
2. The reality of AI adoption and its challenges
2-1. Usage gap: The difference between awareness and habit
Evans points out the gap between high awareness and low daily usage as the reality of AI adoption. For example, even though ChatGPT has 800 to 900 million weekly active users worldwide, only a small fraction of people use AI as an "indispensable tool" in their daily lives.
This provides an important insight into the adoption pattern of new technology.
Many people have "tried it"
However, it has not become established in actual work or daily routines
In other words, there is a reality where AI is widespread, but the UX and integration needed to find daily value have not progressed.
2-2. The reality of corporate adoption
He also points out that while there is much talk about "AI adoption projects" at large companies and consulting firms, the definition of what constitutes AI adoption is ambiguous, and in many cases, it is not linked to actual business value. In fact, even in the cases of Accenture and BCG, it is said that many "AI projects" are actually just extensions of traditional automation and analysis.
From this perspective, he warns against the simplistic argument that "AI strategy means just implementing AI," and provides the insight that true AI adoption should be considered in tandem with the redesign of business processes.
3. The Structure of Value Creation through AI
3-1. The Commoditization of Technology and the Location of Value
Evans points out that current AI models (LLMs) may be rapidly commoditizing.
In other words, the advantages of research results and large-scale models can become obsolete in a short period.
On the other hand, he suggests that the layers that actually generate value lie outside of these models, such as:
Integration via UI/UX(Ease of use)
Embedding into existing workflows
Industry-specific solutions (e.g., laptop specialization)
These are common to the "value differentiated on a platform" seen in the traditional software industry.
3-2. A Three-Stage Model of Industrial Transformation
Evans organizes AI adoption and value creation into three stages:
Absorb
→ Embedding AI as a function into existing processesInnovate
→ Creating new products and servicesDisrupt
→ Transformation that redesigns the industrial structure itself
Currently, many companies remain in the "Absorb" stage, and it is suggested that new thinking and organizational design are required to move on to the next stages of "Innovate" and "Disrupt."
4. The Future of AI: Predictions and Strategic Implications
4-1. Physical and Theoretical Limits of AI
Evans cites the fact that "physical and theoretical limits are unknown" as a key point in evaluating the future of AI.
In traditional computing (modems and processors), there were physical constraints such as latency and size, but with current AI, there is a structural uncertainty in that there is no axis for predicting how far it will evolve.
This uncertainty fuels debates like "When will AI become AGI?" and obscures the focus of investment and strategy.
4-2. Strategic Implications: Productization and Integration
The strategic implications derived from Evans' analysis can be summarized as follows:
Do not treat AI as a standalone product
→ Integrate it as a feature and link it to existing value creationDefine the user's 'Jobs to be Done'
→ Design AI starting from 'what they want to accomplish'Identify areas for transformation
→ Move beyond simple automation toward creating new user experiences and new markets
These points provide a roadmap for companies to connect AI to actual business value, rather than ending up with mere technology implementation when drafting their AI strategies.
Conclusion
Benedict Evans' 'AI Eats the World' is an analysis that calmly evaluates the possibility of AI being the 'next big thing' in technological history, while shedding light on current challenges and future strategic design.
By comparing AI not just as a technological innovation, but against historical platform shifts, and by organizing changes in user behavior, corporate value creation structures, and industrial transformation mechanisms, it provides practical insights and strategies.
While technological evolution is certainly progressing, the core message of this presentation is that how that value connects to society and business depends on future product design and implementation strategies.