The Value Network Shift

In the [YOU] on AI Field Guide

The cycle’s account of the Trivandrum training week is, among other things, an account of a value network shift experienced as a lived event. The senior engineer who discovered that if implementation could be handled by a tool, the remaining twenty percent of his work—the judgment about what to build, the architectural instinct about what would break, the taste that separated a feature users loved from one they tolerated—was not a remnant but the core, was experiencing the new network’s premium structure before the language existed to describe it.

The cycle documents the broader shift through the emergence of what it calls vector pods: small groups whose purpose is not to build but to decide what should be built, and to direct AI tools toward building it. Five years earlier, such a structure would have been incoherent in a software organization. Who directs without building? The new value network provides the answer: the person who can see across domains, assess technical feasibility, understand user needs, and exercise the taste required to determine whether a feature serves its purpose is more valuable than the person who can implement what a specification describes.

The value network shift is also the framework that explains the SaaSpocalypse more precisely than purely financial analysis can. The trillion dollars of market value that evaporated from software companies was not a repricing of earnings. It was a repricing of where in the value stack the competitive advantage resided. Companies whose value was above the code layer—in data, integrations, institutional trust, workflow knowledge—retained their defensible position because the value network shift left those layers intact. Companies whose primary value was functionality—features that performed specific tasks that AI could replicate in conversation—were exposed because the shift moved the premium away from the layer they occupied.

Origin

Christensen developed the value network concept to address a limitation of his initial disruption research. The first-generation framework explained why incumbent firms failed to respond to disruptive threats; the value network concept explained the mechanism more precisely: incumbents were not merely slow or resistant. They were organized to create value in a network that the disruption was in the process of replacing, and the organizational structures, incentive systems, and capability investments that made them excellent in the old network made them poorly positioned for the new one.

The empirical foundation was the same disk drive research that established the disruption pattern. Each generation of disk drive did not merely produce smaller drives; it shifted the entire ecosystem of suppliers, customers, and complementors. The manufacturers who excelled in the 14-inch market were organized around the customers, supply chains, and performance metrics of that market. The 8-inch market had different customers (minicomputer manufacturers rather than mainframe manufacturers), different supply chains, different performance metrics. Excellence in the old network was not transferable to the new one because the networks were structurally different, not merely smaller versions of each other.

Key Ideas

Execution to Judgment. The defining axis of the AI-era value network shift. The pre-AI network rewarded execution—the ability to translate human intention into machine instruction. The post-AI network rewards judgment—the ability to determine what should be translated, evaluate whether the translation serves its purpose, and exercise the taste required to assess quality under conditions of uncertainty. This is not a marginal shift in emphasis. It is the replacement of one bottleneck with another, and the organizational structures that served the execution bottleneck actively impede the judgment bottleneck because they were designed to manage the wrong constraint.

Peripheral Becomes Central. The capabilities that commanded the premium in the old network—implementation skill, domain-specific technical knowledge, the ability to navigate programming languages and frameworks—become commodity inputs in the new network. The capabilities that were peripheral—architectural instinct, cross-domain synthesis, the taste that distinguishes lasting value from superficial competence—become the scarce resources around which organizational structures and compensation models must be rebuilt. This inversion is structurally predictable from Christensen’s framework but experientially disorienting for practitioners who built their professional identities around the old network’s premium structure.

The Ecosystem, Not the Code. In the new value network, competitive advantage migrates from the code layer to the layers that code cannot replicate: proprietary data accumulated through years of deployment, integration ecosystems that connect platforms to each other, institutional trust earned through reliable service, and the accumulated understanding of customer workflows that no AI-generated alternative can acquire through conversation. The companies that survive the value network shift are those whose foundation was always above the code layer; those whose foundation was functionality are exposed.

Incentives Determine Outcomes. The future of the AI transition—as the Christensen Institute’s research argues—will be determined not by the capability of the technology or the intentions of its developers but by the incentive structures of the new value network: who funds AI companies, who their customers are, how competition shapes what trade-offs are rewarded. The value network shift is not merely an analytical concept. It is the mechanism by which the AI transition’s consequences will be determined, because the network determines the incentives and the incentives determine the outcomes.