Skip to main content
The “new oil” metaphor has evolved through three phases — data, identity, and AI — each pointing at something real while hiding a problem the industry was not ready to confront.

​
What this piece is

This paper traces the evolution of the “new oil” metaphor across three phases: data, identity, and AI. Each phase pointed at something real. Each also hid a problem the industry was not ready to confront. The paper examines what broke in each era and what that means for the next decade of technology strategy. Originally published on Medium, February 2026.

​
What is the central argument?

The “new oil” label is not the point. The point is what each label was trying to fix. Data gave leverage but created liability. Identity forced permission but risked commodification. AI provides automation and speed but amplifies the need for trust, control, and accountability. The next dominant constraint will not be a single thing — it will be whatever hurts most at the time: trust, human connection, compute, provenance, distribution, or agency.

​
What are the three phases?

Phase 1: Data is the new oil. Digital businesses built feedback loops where data improved products, products attracted users, and users produced more data. The phrase gave leaders permission to invest. But data in isolation is rarely useful — it needs context, rights, and trust infrastructure to create value. The costs of treating data as a free commodity turned out to be social, legal, and reputational. Phase 2: Identity is the new oil. Data became abundant; the bottleneck shifted to whether you could trust the data and act on it safely. Identity answers who is this person, is it really them, what are they allowed to do. Open banking made this obvious — it was an identity and consent problem dressed up as a data problem. But identity is not a commodity to extract. It is a relationship to earn and maintain. Phase 3: AI is the new oil. AI is a general capability that sits on top of most workflows. It lets software meet humans where they are. But it still depends on data, distribution, and trust. The new constraints are compute costs, governance, security, IP, and reliability. AI does not remove the need for identity and trust. It amplifies it. What comes next. The paper evaluates six candidates for the next constraint: trust, human connection, compute, provenance, distribution, and agency. Each has conditions under which it would dominate the narrative.

​
How does this connect to the wiki’s knowledge areas?

This paper connects three knowledge areas documented in this wiki. The data phase connects to Open banking and identity and the consent infrastructure work done through DirectID. The identity phase draws directly from the Lessons from a Fintech project and the miiCard-to-DirectID pivot. The AI phase connects to AI in regulated markets and the governance-first approach documented in the AI Maturity Roadmap.