The 2028 Global Intelligence Crisis: Citrini Research’s fictional macro scenario

Summary

The article is framed as a scenario exercise: “what if our AI bullishness keeps being right… and what if that’s actually bearish?”. It is written as a “CitriniResearch Macro Memo” dated June 30, 2028, reconstructing a pre-crisis economy and its fallout, while reminding readers at the end that it’s February 2026 and “the canary is still alive”.

In the memo’s world, unemployment prints 10.2% and the S&P 500 is down 38% from its October 2026 highs. Early on, AI-driven layoffs look market-friendly: margins expand, earnings beat, and corporate profits get recycled into AI compute. Productivity surges and nominal GDP still looks healthy, yet real wage growth collapses as white-collar workers are displaced into lower-paying roles.

The piece labels the gap between measured output and household income as “Ghost GDP”: output that shows up in national accounts but does not circulate through the human consumer economy. From there, the memo outlines a self-reinforcing spiral: better AI reduces labor needs, spending softens, firms face pressure and shift more OpEx from payroll into AI, AI improves further, and another round of cuts becomes possible.

In practice

The memo breaks the crisis into three reinforcing transmission channels.

First is operational, starting in software. In late 2025, agentic coding tools take a step-change: with Claude Code or Codex, a competent developer can replicate the core functionality of a mid-market SaaS product in weeks. That changes renewal negotiations and makes “build vs buy” newly credible. The narrative includes a procurement anecdote where the threat of replacing a vendor via AI-enabled “forward deployed” engineers forces steep discounts. Disruption hits the long tail first (e.g., Monday.com, Zapier, Asana) but reaches incumbents.

A key mechanism is reflexivity: what is rational at the firm level becomes destructive in aggregate. Seat-based revenue models get mechanically hit as customers cut headcount, and incumbents—under board pressure and with stocks down—respond by cutting people and funding the technology that is disrupting them, accelerating the loop.

Second is “friction going to zero” in consumer behavior and intermediation. By early 2027, LLM usage is depicted as default, and Qwen’s open-source agentic shopper becomes a catalyst for agents handling purchases. Agents run in the background, continuously optimizing price and terms, compressing subscription economics built on inertia (passive renewals, post-trial price jumps). The memo lists categories where the value proposition was “I’ll handle the tedious complexity”: travel booking, insurance renewals, financial advice, tax prep, routine legal work. Even real estate commissions compress sharply as AI agents replicate knowledge and reduce information asymmetry.

Once agents control transactions, they target the biggest recurring cost: fees. The scenario describes agents routing around card interchange (2–3%) by using stablecoins on Solana or Ethereum L2s, turning agentic commerce from a product story into a payments-plumbing story. Models reliant on friction and habitual app loyalty (with DoorDash as an example) lose moats when agents price-check everything, every time.

Third is financial. The memo argues the risk becomes systemic because the U.S. is fundamentally a white-collar services economy. As incomes weaken and consumption falls, stress appears first in private credit exposed to software/tech and then in the “Mortgage Question”: prime mortgages that were “good on day one” become fragile because the underlying income assumptions are structurally impaired.

Why it matters

• It highlights a specific macro risk: productivity can rise while household income and demand fall (“Ghost GDP”).
• It shows how agents can attack business models built on intermediation, fees, and user inertia.
• It links tech disruption to financial amplifiers (private credit, insurers, mortgages/MBS) in a chain.
• It frames the gap between fast-moving capabilities and slow-moving institutions as a core vulnerability.