Citrini Research modeled what happens if AI actually works as promised. The results are terrifying

Citrini Research has published a provocative thought experiment titled 'Macro Memo from 2028,' modeling the second-order economic effects if AI agents deliver on their productivity promises. The scenario begins with a step-function improvement in agentic coding tools by late 2025, enabling a single competent developer to replicate mid-market SaaS applications in mere weeks. This triggers a rational but devastating chain reaction: CIOs renegotiate enterprise software contracts at 30% discounts, long-tail SaaS companies are hit hardest, and companies like ServiceNow face revenue collapse as their clients' AI-driven headcount reductions directly cancel software licenses. The paper argues each corporate decision is individually logical, but the aggregate result is systemic failure.

The analysis traces the mechanics of 'intermediation collapse,' where AI agents lack brand loyalty and relentlessly optimize for cost, dismantling moats like DoorDash's home-screen advantage. As companies capture AI productivity gains through layoffs, displaced high-earning tech workers (who drive 65% of discretionary spending) cut back, weakening consumer-facing businesses. This creates a feedback loop: AI improvements fuel cost cuts, which fund more AI adoption, leading to more cuts and reduced consumer demand. The scenario warns of private credit defaults on PE-backed software deals underwritten on 'recurring' revenue that suddenly stops recurring. Unlike typical AI doom scenarios, this paper focuses on the financial contagion and the absence of a natural economic brake, presenting a sobering stress test for investors and tech leaders betting on perpetual software growth.