On 24 February 2026 a speculative Substack note from Citrini Research triggered fresh volatility in US equities, warning of widespread disruption from autonomous AI agents. The post — framed by its authors as a “scenario, not a prediction” — sketches a timeline to June 2028 in which white‑collar unemployment tops 10%, a private‑credit shock spreads to mortgages, and markets enter a deep slump. Investors reacted quickly: the S&P 500 fell more than 1% the following Monday and several software and payments names lost roughly 4–6%. The episode has revived debates about how quickly AI could reshape jobs, credit and financial stability.
Key takeaways
- Citrini Research published a Substack scenario that projects a chain of events running through June 2028, including US unemployment cresting over 10%.
- Market moves followed the post: the S&P 500 dropped over 1% on the Monday after publication; several companies named in the piece — Uber, American Express, Mastercard and DoorDash — fell about 4–6%.
- The scenario argues AI agents will destroy business models that monetise “friction,” hitting SaaS firms and forcing pricing pressure on legacy vendors.
- Citrini cites the 2022 Zendesk buyout (Hellman & Friedman and Permira, $10.2bn) as an example of private‑credit exposure vulnerable to revenue shocks.
- The report describes a cascading failure: software defaults lead to private‑credit losses, mortgage stress and a late‑2027 crash that erases 57% of the S&P in the scenario.
- Protests dubbed “Occupy Silicon Valley” and the concept of “ghost GDP”—output that does not circulate in the real economy—are central social consequences in the narrative.
- Market strategists and academics differ: some call the note a wake‑up call about structural risk; others say current AI capabilities remain far from the scenario’s assumptions.
Background
Over the last two years, rapid advances in large‑language models and agentic tools have amplified investor interest and anxiety about automation. Firms such as OpenAI and Anthropic have released developer tools that extend coding and workflow automation capabilities, prompting fresh debate about productivity gains versus displacement risks. Independent research newsletters and Substack essays have become influential, sometimes moving prices when a widely read note synthesises technology, credit and labour risks into a single narrative.
Private credit grew as an asset class after the Global Financial Crisis, with non‑bank lenders financing leveraged buyouts and growth deals across software and technology sectors. These structures often rely on projected, multi‑year revenue streams; the Citrini scenario highlights how those revenue assumptions can break down if AI fundamentally reduces product demand. At the same time, the US fiscal system remains heavily linked to labour income, complicating government responses if mass job losses occur.
Main event
Citrini’s core narrative begins with an assumed rapid improvement in agent capabilities, citing recent model releases that enhance code generation and task automation. In the scenario, businesses adopt personal and enterprise AI agents that perform complex coordination tasks, eroding demand for many current SaaS offerings and platforms that capture frictional income. The report argues that firms relying on subscription and middleman models — travel agencies, delivery platforms, payments networks — face margin compression as agents route around intermediaries.
As white‑collar roles contract, the scenario envisions displaced workers moving into precarious gig work, depressing wages and consumer spending. Reduced household income then feeds back into lower demand for goods and services, prompting firms to substitute further investment in automation rather than hiring. Citrini labels this dynamic “a feedback loop with no natural brake,” where lower demand and rising automation reinforce one another.
The note then traces financial channels: private‑credit loans to software companies, structured on optimistic revenue forecasts, begin to default as revenues fall. Using Zendesk’s 2022 take‑private by Hellman & Friedman and Permira for $10.2bn as an illustrative case, the scenario describes a large private‑credit software default that strains asset managers holding diversified balance sheets. Simultaneously, mortgage delinquencies rise as affected households struggle to service loans, culminating in a severe market contraction in late 2027 in the scenario.
Finally, Citrini projects political and social fallout: government revenues fall as earnings decline, while large AI firms continue to report strong headline output, producing “ghost GDP” that obscures underlying distress. Public protests against AI firms, described as an “Occupy Silicon Valley” movement, are presented as a plausible social response to growing inequality and perceived corporate immunity.
Analysis & implications
The scenario packs multiple transmission channels — labour market displacement, firm revenue declines, private‑credit defaults and mortgage stress — into a single cascade. Each channel is individually plausible under an extreme set of assumptions, but the probability that they synchronise exactly as described is uncertain. What matters for policy and markets is not only whether the scenario occurs but how much tail risk it exposes: even a smaller frictional shift in labour demand could produce concentrated credit losses and regional housing stress.
For investors, the episode underlines how narratives can amplify short‑term volatility. The immediate reaction — a >1% S&P decline and selective weakness in software and payments names — reflects how concentrated market positions and headline risk interact with algorithmic trading and risk models. Portfolio managers monitoring exposure to long‑duration tech earnings and to private credit should reassess shock scenarios that compress those future cash flows.
On policy, the scenario challenges conventional macro tools. Central banks influence aggregate demand through interest rates and liquidity, but cannot directly halt a structural substitution of human labour by automation. If displacement is rapid and concentrated, fiscal responses — retraining, income support, and targeted credit backstops — would be the main lever, yet time and political constraints could limit their effectiveness. The scenario therefore highlights the need to pre‑position stabilisers and regulatory frameworks for high‑impact automation.
Internationally, a US‑centric labour shock would have spillovers through financial channels, given the dollar’s centrality and the global footprint of US tech firms. Countries with closer trade and financial links to US consumer demand or with significant exposures to US private credit could face second‑order effects, forcing multilateral coordination on financial stability and labour policy.
Comparison & data
| Metric | Citrini scenario | Market reference / near‑term move |
|---|---|---|
| US unemployment | Peaks >10% by June 2028 | Historical pre‑2026 levels varied; scenario projects a structural jump |
| S&P 500 | 57% wipeout at peak of crisis (late 2027) | Index fell >1% the Monday after the Substack note |
| Company moves | Uber, AmEx, Mastercard, DoorDash cited; each lost ~4–6% | Software component hit lowest level since April tariff shock |
| Private deal example | Zendesk take‑private financing cited ($10.2bn, 2022) | Used to illustrate private‑credit exposure to SaaS revenues |
The table presents scenario projections alongside the immediate market responses reported after the note circulated. It is not a forecast but a mapping of claims to observed short‑term price moves. Readers should treat the projected magnitudes — particularly the 57% S&P loss and the timing to mid‑2028 — as part of a constructed stress case rather than a baseline expectation.
Reactions & quotes
Market commentators offered rapid, contrasting takes. Some described the piece as sensational but useful as a warning about structural change; others warned that current AI systems remain far from the all‑consuming agents the scenario requires. The following excerpts capture a range of viewpoints and are presented with context.
“It reads like doomsday sensationalism, but it does force investors to think about how different today’s economy is compared with a few years ago.”
Neil Wilson, Saxo Capital Markets (market analyst)
Wilson’s remark frames the public reaction: even if the scenario is implausible in full, it can alter risk perceptions by highlighting structural change already underway.
“This is a scenario, not a prediction — intended to map a plausible chain of economic and financial outcomes if agentic AI adoption accelerates.”
Citrini Research (Substack)
Citrini emphasized the hypothetical nature of its narrative while laying out linkages across labour, corporate revenue and credit markets, describing the exercise as stress testing systemic consequences.
“A widely circulated Substack thought piece is enough to knock the market sideways, showing how narrative risk now moves prices alongside fundamentals.”
Stephen Innes, SPI Asset Management (asset manager)
Innes’ comment reflects how swiftly market sentiment can shift when a coherent story ties together technology, credit and macro risks.
Unconfirmed
- The exact timeline to June 2028 and the magnitude of a 57% S&P loss are projections from the Citrini scenario and are not independently verified.
- Claims that AI agents will universally replace all white‑collar tasks and cause mass, immediate redeployment into gig work lack direct empirical support at present.
- The scenario’s assertion that agents will shift most transactions to cryptocurrencies as a default payment rail is speculative and not supported by current payment adoption trends.
Bottom line
The Citrini Research note is a tightly constructed stress scenario that links rapid AI adoption to labour displacement, private‑credit losses and housing stress. Its shock value lies in chaining plausible mechanisms into a concentrated calamity; that chaining, rather than any single claim, was sufficient to unsettle markets. Investors and policymakers should treat the episode as a reminder that narrative risk can move prices and that structural technological change may create novel tail risks.
Practical steps follow from this diagnosis: risk managers should review exposures to long‑duration tech earnings and private‑credit vehicles; regulators and fiscal authorities should evaluate contingency tools for concentrated labour shocks; and researchers should prioritise empirical work on the speed and distribution of automation risk. Whether the extreme outcome unfolds or not, the note underscores the urgency of building frameworks that can respond to fast, system‑wide shifts in how value is created and distributed.
Sources
- The Guardian (media report summarising the Citrini scenario)
- Citrini Research (Substack; primary scenario publication)
- OpenAI (official blog; reference for recent model releases)
- Anthropic (company site; reference for model developments)
- S&P Dow Jones Indices (index provider; market data reference)