Scope: Global · Macro outlook and the physical economics of AI · DOI: 10.5281/zenodo.23107280
Santiago Sainz · Economics and Social Affairs Department · ISDO
Abstract:
This report compiles and compares 2027 economic projections published by the International Monetary Fund, the OECD, the World Bank, and the central banks of the United States, the eurozone, the United Kingdom, and Japan, and sets that macro outlook alongside a second, equally weighted line of inquiry: the physical and regulatory economics of the artificial intelligence infrastructure now reshaping how, and where, economic growth itself is generated.
On the macro side, the report’s central finding is that institutional forecasts for 2027 diverge genuinely rather than converging on a technical consensus. The IMF’s 3.4% world growth projection and the OECD’s 3.0% trace back to different explicit assumptions about how long the 2026 Middle East conflict will last. On the AI infrastructure side, this edition adds a dedicated analysis built around four questions. How large is data center electricity demand actually projected to grow? How are governments across the EU, Ireland, the Netherlands, Singapore, China, and the United States responding, and why do their regulatory models differ so sharply? What does the environmental cost of unused corporate data, “dark data,” reveal about a gap none of those regulatory models currently addresses? And which efficiency technologies are already closing part of that gap in practice, regardless of what regulation requires?
A connecting pattern runs through both halves of this report: the same AI-driven investment the IMF credits with sustaining the 2027 growth recovery moves in close parallel with the electricity demand documented in Section 7, meaning the macro forecast and the infrastructure reckoning are not two separate stories but two measurements of one story, examined at different levels of resolution. This report treats that relationship as a pattern worth tracking together rather than a quantified causal chain.