The 2028 Global Intelligence Crisis

A speculative scenario modeling what happens if AI advancement continues to exceed expectations — and why that might be economically catastrophic.

What Is It

A 10,000+ word scenario analysis written as a macro research memo from "June 2028" looking back at how rapid AI advancement triggered a Global Intelligence Crisis — mass white-collar unemployment, economic collapse, mortgage crisis, and government paralysis. Written in February 2026, this is explicitly not a prediction, but a thought experiment exploring underexplored left-tail risks.

"What if our AI bullishness continues to be right…and what if that's actually bearish?"

Key Premise

Human intelligence has always been the scarce input. Every economic institution — labor markets, mortgage underwriting, tax systems — was designed assuming human intelligence would remain scarce and valuable.

What happens when machine intelligence becomes a competent substitute?

The scenario models the "Intelligence Displacement Spiral":

AI capabilities improve
Companies lay off white-collar workers
Displaced workers spend less
Companies invest savings into more AI
AI capabilities improve (repeat)

"A feedback loop with no natural brake."

Timeline (from the scenario)

2025-2026: The Buildup

Late 2025: Agentic coding tools (Claude Code, Codex) take step-function jump in capability
2026: Initial white-collar layoffs boost corporate margins, stocks rally (S&P → 8000)
Oct 2026: ServiceNow reports first major SaaS slowdown; stock drops 18%
Markets treat it as "sector-specific" disruption

2027: The Acceleration

Q1 2027: Agentic commerce agents go mainstream; Mastercard sees transaction volume slow
Q2-Q3 2027: Private credit defaults in PE-backed software deals
Sept 2027: Zendesk ($5B LBO) misses covenants — largest private credit software default
Nov 2027: Market crash as "daisy chain of correlated bets on white-collar productivity growth" unravels

2028: The Crisis

June 2028: Unemployment at 10.2%, S&P down 38% from Oct 2026 highs
Prime mortgages in San Francisco, Seattle, Austin showing stress
Federal receipts down 12% as labor's share of GDP collapses from 56% → 46%
Government debates "Transition Economy Act" while social unrest escalates

Key Mechanisms

1. "Ghost GDP"

"Output that shows up in the national accounts but never circulates through the real economy."

A GPU cluster in North Dakota generates output of 10,000 Manhattan workers — but machines spend zero dollars on discretionary goods. GDP grows, velocity of money collapses.

2. The Intermediation Collapse

Any business model based on human limitations dies:

Subscription economy: Agents auto-cancel forgotten subscriptions
Travel booking: Agents assemble itineraries faster/cheaper than platforms
Insurance renewals: Agents re-shop annually (15-20% of premiums vanish)
Real estate commissions: Compress from 5-6% → under 1%
DoorDash/Uber Eats: Agents route to lowest fee; competitors emerge overnight

"Machines optimizing for price and fit do not care about your favorite app or the websites you've been habitually opening for the last four years."

3. The Private Credit Implosion

PE-backed SaaS deals assumed mid-teens revenue growth in perpetuity. Those assumptions died with agentic coding.

Zendesk case: $10.2B LBO (2022) backed by$ 5B direct lending facility. By 2027, AI agents handle customer service autonomously — ARR is no longer "recurring." Loan defaults, becomes largest private credit software default on record.

The "permanent capital" myth: Alternative asset managers (Apollo/Athene, KKR/Global Atlantic) had acquired life insurers and invested policyholder money into private credit. When loans defaulted, regulators cracked down — the "locked-up capital" was actually Main Street annuity savings.

4. The Mortgage Question

"Are prime mortgages money good?"

White-collar workers with 780+ credit scores, 20% down, stable employment records. Loans were good on day one. The world changed after.

Product manager making $180K → driving Uber for$ 45K. Can still make mortgage payment by draining savings and stopping all discretionary spending. "Technically current on their mortgage, but just one more shock away from distress."

Early delinquencies spike in SF, Seattle, Manhattan, Austin — the tech/finance heavy metros.

5. Government Revenue Crisis

Federal government's revenue base is a tax on human time. As labor's share of GDP collapses (64% in 1974 → 56% in 2024 → 46% in 2028), receipts fall 12% below projections.

"The output is still there. But it's no longer routing through households on the way back to firms, which means it's no longer routing through the IRS either."

Key Quotes

On the Negative Feedback Loop

"AI got better and cheaper. Companies laid off workers, then used the savings to buy more AI capability, which let them lay off more workers. Displaced workers spent less. Companies that sell things to consumers sold fewer of them, weakened, and invested more in AI to protect margins. AI got better and cheaper."

On Why This Time Is Different

"Every new job, however, required a human to perform it. AI is now a general intelligence that improves at the very tasks humans would redeploy to. Displaced coders cannot simply move to 'AI management' because AI is already capable of that."

On the Speed of Change

"AI capability is evolving faster than institutions can adapt. The policy response is moving at the pace of ideology, not reality."

On the Structural Nature

"In a normal recession, the cause eventually self-corrects. This cycle's cause was not cyclical."

On Economic Assumptions

"For the entirety of modern economic history, human intelligence has been the scarce input. We are now experiencing the unwind of that premium."

The S&P 500 Projection (in the scenario)

Oct 2026 peak: 8000
June 2028: Down 38% → ~5000
Projected bottom (if mortgage crisis hits): 3500 (57% drawdown, matching GFC)
- That would be November 2022 levels — month before ChatGPT

Policy Paralysis

Proposals Debated (in the scenario)

"Transition Economy Act": Direct transfers to displaced workers funded by deficit + tax on AI inference compute
"Shared AI Prosperity Act": Public claim on AI infrastructure returns (sovereign wealth fund model)

Political Gridlock

Right: Calls it Marxism, warns taxing compute helps China
Left: Warns of regulatory capture by incumbents
Fiscal hawks: Unsustainable deficits
Doves: Don't repeat post-GFC austerity mistakes

Meanwhile: Occupy Silicon Valley blockades Anthropic and OpenAI offices for three weeks.

What Makes This Scenario Compelling

1. Plausible Mechanisms

Each link in the chain is individually believable:

Agentic coding displacing SaaS revenue
Agents optimizing away intermediation layers
PE-backed software defaults
High-earner displacement hitting discretionary spending

2. Second-Order Effects

The piece traces how sector-specific disruption metastasizes:

Software → consulting → insurance → real estate → gig economy
Job losses → spending cuts → margin pressure → more AI investment
Private credit → insurance balance sheets → municipal bonds

3. The "OpEx Substitution" Insight

"A company that had been spending $100M a year on employees and$ 5M on AI now spent $70M on employees and$ 20M on AI. AI investment increased by multiples, but it occurred as a reduction in total operating costs."

Traditional recessions see CapEx cuts. Here, AI spending increases as overall costs fall.

4. The Speed Differential

"Two years. That's all it took to get from 'contained' and 'sector-specific' to an economy that no longer resembles the one any of us grew up in."

5. No Natural Floor

Unlike traditional automation (which required humans to build, deploy, maintain), AI agents improve themselves. The displacement accelerates rather than stabilizing.

Limitations & Critiques

Explicitly Not a Prediction

The authors are clear:

"This is a scenario, not a prediction. This isn't bear porn or AI doomer fan-fiction. The sole intent of this piece is modeling a scenario that's been relatively underexplored."

Assumes No Adaptation

Government could act faster/better than modeled
New job categories could emerge faster than displacement
Social safety nets could adjust
AI development could plateau (multiple sigmoid attempts shown)

Convex to Infrastructure Winners

The scenario notes that AI infrastructure (NVDA, TSM, hyperscalers) kept performing throughout. Taiwan and Korea outperformed. The crisis is concentrated in consumption-heavy economies.

Omits AI Productivity Gains to New Industries

Focuses heavily on displacement, less on potential for AI to enable entirely new categories of goods/services that don't exist yet.

Why This Matters

For Risk Assessment

Even if probability is low, this scenario has:

High impact (economic collapse)
Plausible transmission mechanisms
Fast timeline (2 years from "contained" to crisis)

For Portfolio Construction

"As investors, we still have time to assess how much of our portfolios are built upon assumptions that won't survive the decade."

Questions the piece forces:

Which businesses depend on human limitations?
Which loans are underwritten against white-collar income assumptions?
Which equities are long consumer spending vs. infrastructure?

For Policy Planning

"As a society, we still have time to be proactive. The canary is still alive."

The piece argues for thinking through policy frameworks before the crisis, not during.

harness-engineering-openai — the AI capabilities driving the scenario
cognitive-debt-ai — related risk of losing understanding of systems
anthropic-distillation-attacks — competitive dynamics accelerating AI development
openai-mission-evolution — mission drift as economic pressures mount
dmux, autonomous-dogfooding-skill — tools accelerating the displacement

Added: February 25, 2026