A viral "thought experiment from the future" is forcing Wall Street to confront AI's darkest economic scenario — and Michael Burry thinks it's not bearish enough
Executive Summary
- Citrini Research's "2028 Global Intelligence Crisis" memo — a fictional postmortem written from June 2028 — has gone viral across Wall Street, endorsed by Michael Burry and covered by Fortune, TechCrunch, and Business Insider within 48 hours of publication
- The core thesis: AI-driven corporate profits create "Ghost GDP" — output that inflates national accounts but never circulates through the real economy, because machines spend zero dollars on discretionary goods
- The scenario: S&P 500 crashes 38% from October 2026 highs, unemployment hits 10.2%, and the $13 trillion mortgage market fractures as white-collar incomes are structurally impaired
Chapter 1: The Memo That Shook Wall Street
On February 22, 2026 — amid a historic blizzard paralyzing the US Northeast, a DHS shutdown entering its ninth day, and markets still reeling from the Supreme Court's IEEPA tariff ruling — an analyst named James Van Geelen published a 4,000-word memo on Substack that would, within hours, become the most-discussed document in finance.
Van Geelen is the founder of Citrini Research and Substack's top-ranked finance writer. A former Los Angeles paramedic with degrees in biology and psychology, he built his reputation on what he calls "second-order thinking" — looking past immediate headlines to anticipate what must happen next. His real-world investment portfolio has reportedly surged over 200% since May 2023, driven by early calls on Nvidia and GLP-1 drugs.
The memo, titled "The 2028 Global Intelligence Crisis," is structured as a postmortem from June 2028 — a fictional macro research note looking back at how the AI boom destroyed the very economy it was supposed to supercharge. Co-authored with Alap Shah of the University of California's technology management program, it describes a scenario where the S&P 500 has crashed 38% from its October 2026 peak of 8,000, unemployment has printed at 10.2%, and the American consumer economy has entered what the authors call a "human intelligence displacement spiral."
Within hours, Michael Burry — the investor who famously predicted the 2008 housing crisis, immortalized in "The Big Short" — shared the memo on X with the comment: "And you think I'm bearish." Fortune, TechCrunch, and Business Insider all published analyses within a day. By Monday morning, the memo had become required reading in trading desks from New York to London.
Chapter 2: The Ghost GDP Thesis
At the heart of Citrini's scenario is a deceptively simple concept: Ghost GDP.
For the entirety of modern economic history, human intelligence has been the economy's scarce input. Capital is abundant. Natural resources are finite but replaceable. But the unique human ability to analyze, decide, create, persuade, and coordinate — this was "the thing that could not be replicated at scale." The entire structure of economic life is built around this scarcity. Citrini calls it "friction."
AI destroys friction. And that is the problem.
In Citrini's scenario, the initial wave of AI-driven layoffs in early 2026 does exactly what layoffs are supposed to do: margins expand, earnings beat expectations, stocks rally. By October 2026, the S&P 500 flirts with 8,000 and the Nasdaq breaks above 30,000. Record-setting corporate profits are funneled back into AI compute. Nominal GDP prints mid-to-high single-digit annualized growth. Productivity booms.
But there is a ghost in this machine. A single GPU cluster in North Dakota generating the output previously attributed to 10,000 white-collar workers in midtown Manhattan creates "output that shows up in the national accounts but never circulates through the real economy." The velocity of money flatlines. The 70% of GDP driven by the human-centric consumer economy withers. Because machines spend zero dollars on discretionary goods.
The negative feedback loop has no natural brake:
- AI capabilities improve → companies need fewer workers
- White-collar layoffs increase → displaced workers spend less
- Consumer spending contracts → margin pressure rises
- Companies invest more in AI to cut costs → AI capabilities improve further
This is the human intelligence displacement spiral.
Chapter 3: The Death of Friction
Citrini's most provocative insight concerns what happens when AI agents begin operating 24/7 to optimize consumer decisions. Every business model built on "habitual intermediation" — the monetization of human friction — faces existential threat.
The memo traces a specific sequence of disruption:
Phase 1: Travel (Q4 2026). AI agents assemble complete travel itineraries faster and cheaper than any booking platform. Expedia, Booking.com, and their peers see transaction volumes collapse as agents negotiate directly with airlines and hotels, routing around platform fees.
Phase 2: Financial services (2027). Insurance renewals, financial advice, tax preparation — any service whose value proposition is "I will navigate complexity that you find tedious" — gets disrupted. "The agents found nothing tedious," the memo notes.
Phase 3: Transaction fees (2027-2028). AI agents ruthlessly route around the 2-3% interchange fees charged by credit card networks. Visa and Mastercard's moats, built on friction, erode as friction goes to zero.
"Turns out that a lot of what people called relationships was simply friction with a friendly face," Van Geelen writes.
| Sector | Revenue Model | Friction Dependency | AI Vulnerability |
|---|---|---|---|
| Travel booking | Platform fees (15-25%) | Search complexity | Extreme |
| Insurance brokerage | Commission (5-15%) | Comparison difficulty | High |
| Financial advisory | AUM fees (0.5-1.5%) | Portfolio complexity | High |
| SaaS mid-market | Per-seat licensing | Workflow integration | Extreme |
| Credit card networks | Interchange (2-3%) | Payment infrastructure | Moderate-High |
Chapter 4: The SaaS Reflexivity Trap
The most detailed section of Citrini's scenario involves the software industry — a thesis that aligns with the "SaaSpocalypse" already underway in real-world markets since Anthropic's Claude Cowork launch triggered a $285 billion SaaS market cap destruction in late January 2026.
In Citrini's telling, the mechanism of destruction is reflexivity — a self-reinforcing loop where each company's rational response collectively produces catastrophic outcomes.
The trigger: by mid-2026, agentic coding tools took a step-function jump in capability. A competent developer working with Claude Code or Codex could replicate the core functionality of a mid-market SaaS product in weeks. Not perfectly, but well enough that CIOs reviewing $500,000 annual renewals started asking: "What if we just built this ourselves?"
The fictional ServiceNow Q3 2026 earnings report crystallizes the trap: net new ACV growth decelerates from 23% to 14%, the company announces a 15% workforce reduction, and shares fall 18%. The interconnected nature of these systems reveals itself. ServiceNow sells seats. When Fortune 500 clients cut 15% of their workforce, they cancel 15% of their licenses. The same AI-driven headcount reductions boosting margins at their customers mechanically destroy ServiceNow's revenue base.
The historical disruption model said incumbents resist new technology, lose share to nimble entrants, and die slowly. That's what happened to Kodak, Blockbuster, BlackBerry. What Citrini describes for 2026-2028 is fundamentally different: the incumbents don't resist because they can't afford to.
With stocks down 40-60% and boards demanding answers, AI-threatened companies cut headcount, redeploy savings into AI tools, and use those tools to maintain output at lower cost. Each company's individual response is rational. The collective result is catastrophic. Every dollar saved on headcount flows into AI capability that makes the next round of job cuts possible.
Chapter 5: The $13 Trillion Mortgage Time Bomb
Citrini's scenario extends far beyond software into the heart of the American financial system.
White-collar workers currently make up approximately 50% of US employment and drive roughly 75% of the nation's discretionary spending. In the fictional 2028 scenario, these workers see their earnings power structurally and permanently impaired. High-earning professionals are forced to downshift into gig economy roles, flooding the labor supply and pushing down wages economy-wide.
The implications for the $13 trillion residential mortgage market are severe. Prime borrowers with 780 FICO scores — historically the safest segment of the housing market — see their incomes structurally impaired. Mortgage underwriters must reassess whether prime mortgages are still money-good.
Simultaneously, the $3 trillion private credit market faces reckoning. Seventeen years without a real default cycle left private credit bloated with PE-backed software deals that assumed annual recurring revenue would remain recurring. When AI disruption triggers the first wave of defaults in mid-2027, the daisy chain unravels.
Historical Precedent: The 2008 Analogy
| Feature | 2008 Financial Crisis | Citrini's 2028 Scenario |
|---|---|---|
| Trigger | Subprime mortgage defaults | White-collar income impairment |
| Transmission | CDO/MBS correlation | SaaS→PE→private credit chain |
| Affected borrowers | Subprime (low FICO) | Prime (high FICO, AI-displaced) |
| GDP impact | -4.3% peak-to-trough | "Ghost GDP" masks true decline |
| Market crash | S&P -57% | S&P -38% (scenario) |
| Unemployment peak | 10.0% (Oct 2009) | 10.2% (fictional June 2028) |
| Policy response | QE, TARP, fiscal stimulus | "Lack of comprehensive plan" |
The critical difference: in 2008, the problem was identifiable and addressable — remove toxic assets, recapitalize banks, restore lending. In Citrini's 2028, the problem is structural. You cannot restore incomes destroyed by a technology that keeps improving at the exact tasks displaced workers would pivot toward.
Chapter 6: The Counterarguments
To be clear, Citrini describes this as "a scenario, not a prediction" and explicitly disclaims it as "bear porn or AI doomer fan-fiction." Several significant counterarguments exist:
1. Historical productivity gains reallocate value rather than destroy it. When AI drives costs down, goods and services become cheaper, effectively raising real purchasing power even for lower-income workers. The Industrial Revolution displaced craftsmen but created factory workers. The internet destroyed travel agents but created an entirely new digital economy.
2. The timeline may be too aggressive. Enterprise software adoption typically takes years, not months. Procurement cycles, integration requirements, regulatory constraints, and organizational inertia all slow the pace of disruption. The "competent developer replicating a SaaS product in weeks" scenario requires technical talent that remains scarce.
3. Government response is not helpless. Universal basic income, retraining programs, fiscal stimulus, and regulatory intervention could mitigate the displacement spiral. The New Deal responded to the Great Depression. The CARES Act responded to COVID. Policy lag is real but not infinite.
4. AI creates new categories of work. The history of technology suggests that the most important jobs created by a new technology are ones we cannot imagine beforehand. Nobody predicted "social media manager" in 1990 or "prompt engineer" in 2020.
5. Corporate self-restraint. If AI-driven layoffs destroy the consumer base, companies face declining revenues. Rational CEOs should recognize the collective action problem and slow the pace of displacement. But as Citrini notes, this is a classic prisoner's dilemma — no individual firm can afford to be the one that doesn't cut costs.
Chapter 7: Scenario Analysis
Scenario A: Controlled Transition (30%)
Thesis: AI displacement occurs gradually, policy responds adequately, new job categories emerge.
Prerequisites:
- AI capability improvements slow from exponential to linear
- Federal government implements UBI or major retraining programs by 2027
- Enterprise adoption constrained by integration complexity
- New human-AI collaborative roles emerge at scale
Historical precedent: The 1990s internet transition — significant displacement in retail, media, and travel, but net job creation over a 10-year horizon. US unemployment peaked at 6.3% (2003) before declining to 4.4% (2007).
Market outcome: S&P 500 experiences 15-20% correction, recovers within 18 months. Sectoral rotation from software/services to infrastructure/energy.
Scenario B: Citrini's Displacement Spiral (40%)
Thesis: AI adoption outpaces societal adaptation, creating a deflationary consumer recession masked by Ghost GDP.
Prerequisites:
- Agentic AI capabilities continue step-function improvements through 2027
- White-collar layoffs accelerate beyond 5% of total employment
- Private credit defaults trigger financial contagion
- Policy response lags by 12-18 months
Historical precedent: The 1930s Great Depression — technological productivity gains (electrification, assembly lines) coincided with insufficient demand and deflationary spiral. Unemployment peaked at 24.9% (1933).
Market outcome: S&P 500 drops 30-40% from peak. Gold and real assets outperform. Private credit market seizes. Recovery takes 3-5 years.
Why 40%: The current trajectory most closely matches this scenario. The SaaSpocalypse is already underway ($285B destroyed), private credit is already showing stress ($3T first credit cycle test), and AI capabilities are improving faster than consensus expects. The prisoner's dilemma dynamics are visible in real-time — every company cutting headcount to fund AI investment.
Scenario C: AI Supercycle Continues (30%)
Thesis: Productivity gains translate into sufficient new demand, Solow Paradox resolves favorably, and AI creates more value than it destroys.
Prerequisites:
- AI-driven cost reductions create massive consumer surplus
- New trillion-dollar industries emerge (autonomous vehicles, personalized medicine, space economy)
- Displaced workers upskill faster than jobs disappear
- AI infrastructure spending sustains employment through multiplier effects
Historical precedent: The post-WWII boom — massive technological displacement (wartime automation) followed by unprecedented prosperity driven by new industries (suburbs, consumer electronics, aviation).
Market outcome: S&P 500 reaches 10,000+ by 2028. New tech giants emerge. Income inequality widens but absolute living standards improve.
Chapter 8: Investment Implications
Immediate (0-6 months)
- Nvidia earnings (Feb 25) become critical signal — any guidance weakness validates Citrini's thesis that AI investment is creating Ghost GDP rather than sustainable growth
- Private credit exposure warrants urgent review — BDCs with heavy software/SaaS lending face first-mover risk
- Volatility positioning — VIX at ~21 dramatically underprices the tail risks described in the memo
Medium-term (6-18 months)
- Long real assets, short financial intermediation — the friction-to-zero thesis implies structural headwinds for payment networks, insurance brokers, and financial advisors
- Gold and commodities benefit from both Ghost GDP fears and the ongoing central bank de-dollarization trend ($5,000 gold)
- Short SaaS long-tail — companies with per-seat pricing models and high customer concentration face the reflexivity trap
Long-term (18+ months)
- Infrastructure over software — physical assets (energy, mining, defense) are harder to replicate than digital services
- AI infrastructure beneficiaries — TSMC, energy utilities, data center REITs benefit regardless of which economic scenario plays out
- Government bonds — in Scenario B, deflationary recession drives Treasury yields sharply lower
Conclusion
The Citrini memo's power lies not in its specific predictions — which the authors themselves call a scenario, not a forecast — but in its articulation of a mechanism that markets have not priced. The negative feedback loop between AI-driven productivity gains and consumer spending destruction is not speculative. It is a logical consequence of replacing the economy's scarce input — human intelligence — with an abundant one.
Michael Burry recognized the housing crisis because he understood that the financial system was making a correlated bet it didn't realize it was making. Citrini's memo makes a similar argument: the entire economy is one long daisy chain of correlated bets on white-collar productivity growth. When that chain breaks, the contagion is not financial — it is structural.
The question is not whether AI will displace white-collar workers. That process has already begun. The question is whether the displacement will be gradual enough for society to adapt, or rapid enough to trigger the deflationary spiral Citrini describes. The answer likely depends on a variable that no economic model can forecast: the pace of AI capability improvement over the next 24 months.
As Van Geelen wrote in April 2025, there is a "sword of Damocles" hanging over the white-collar employee. The Citrini memo has given that sword a name, a timeline, and a mechanism. Wall Street is now debating whether it will fall.
Sources: Citrini Research "2028 Global Intelligence Crisis" (Feb 22, 2026), Fortune, TechCrunch, Business Insider, Bloomberg, Seeking Alpha
Leave a Reply