James van Geelen’s ‘Citrini’ Thesis: Why the AI Capex Cycle Faces a Reckoning
Key Takeaways
- James van Geelen, the strategist behind the influential Citrini research, warns that the massive capital expenditure fueling the AI boom is reaching a precarious tipping point.
- His 'AI Doom Scenario' suggests that a failure to realize immediate productivity gains could trigger a systemic market correction as investors re-evaluate tech valuations.
Key Intelligence
Key Facts
- 1The 'Citrini Selloff' thesis warns of a systemic re-rating of AI-linked equities due to unsustainable capex.
- 2Hyperscaler capital expenditure is projected to exceed $200 billion annually without clear short-term ROI.
- 3Market concentration in the top 5 tech stocks is at historic highs, increasing the risk of a liquidity vacuum.
- 4James van Geelen compares the current AI build-out to the fiber-optic over-provisioning of the late 1990s.
- 5The scenario suggests a potential 20-30% correction if AI monetization fails to meet institutional expectations.
Who's Affected
Analysis
The financial markets are currently grappling with a paradox: record-breaking capital expenditure by tech giants and a growing skepticism regarding the immediate return on that investment. James van Geelen, the mind behind the 'Citrini' research brand, has emerged as a leading voice in articulating this tension. His 'AI Doom Scenario,' which has recently gained viral traction across institutional trading desks, posits that the market is dangerously close to a 'Minsky Moment' for artificial intelligence—a point where the debt and capital used to fund the build-out can no longer be supported by the actual cash flows generated by the technology.
At the heart of van Geelen’s thesis is the 'Citrini Selloff'—a projected sequence of events where the primary drivers of the S&P 500’s recent gains undergo a violent re-rating. For the past two years, the market has been sustained by a virtuous cycle: hyperscalers like Microsoft, Alphabet, and Meta spend tens of billions on Nvidia’s H100 and Blackwell chips, which in turn fuels record earnings for the semiconductor sector and drives index-level momentum. However, van Geelen argues that this cycle is 'cannibalistic' if the end-users of AI software do not see a corresponding surge in productivity or revenue. If the hyperscalers do not see a path to monetization that justifies $200 billion in annual capex, they will eventually be forced to pull back. When that happens, the 'air pocket' in demand for semiconductors could lead to a collapse in the most crowded trades in the market.
If the hyperscalers do not see a path to monetization that justifies $200 billion in annual capex, they will eventually be forced to pull back.
This scenario draws uncomfortable parallels to the fiber-optic build-out of the late 1990s. During that era, companies laid thousands of miles of cable in anticipation of internet traffic that didn't materialize for another decade, leading to a decade of stagnation for the hardware providers. Van Geelen suggests we are seeing a similar 'over-provisioning' phase in AI. While the long-term utility of Large Language Models (LLMs) is rarely questioned, the short-term valuation of the companies building them assumes a level of immediate adoption that may be physically and organizationally impossible for most enterprises to achieve in the next 18 months.
What to Watch
Furthermore, the 'Citrini' outlook highlights the risk of market concentration. With a handful of stocks accounting for the vast majority of the market's year-to-date returns, any fundamental shift in the AI narrative isn't just a sector problem—it’s a systemic risk. Van Geelen points out that institutional positioning is so heavily skewed toward 'AI winners' that a shift in sentiment could trigger a liquidity vacuum. In this 'doom scenario,' the selloff isn't just a correction; it's a forced deleveraging of the AI trade that could spill over into broader credit markets and consumer confidence.
Investors should watch for the 'Capex-to-Revenue' ratio in upcoming quarterly reports from the 'Magnificent Seven.' If capital spending continues to climb while AI-attributed revenue remains a rounding error, the 'Citrini Selloff' moves from a theoretical scenario to a market reality. Van Geelen’s perspective serves as a critical counterweight to the prevailing 'AI-at-any-price' mentality, suggesting that the next phase of the market will be defined not by who can build the biggest model, but by who can actually turn a profit from it.
Sources
Sources
Based on 2 source articles- BloombergOdd Lots: James van Geelen on His AI Doom Scenario (Podcast)Feb 28, 2026
- BloombergJames van Geelen on His Viral AI Doom ScenarioFeb 28, 2026
How we covered this story
Every story in our finance coverage is assembled from multiple primary sources, cross-referenced for factual consistency, and scored along three independent dimensions: sentiment, operational impact, and source-cluster confidence. Single-source rumors and unverifiable claims do not pass our editorial gate. When a story shows "Verified by N sources" with N≥2, the development is independently corroborated; when N=1, we mark it explicitly so readers can weigh the signal accordingly.
Impact scoring uses a 1-10 scale weighted toward regulatory, financial, and operational consequence rather than coverage volume. A topic that runs in every outlet but moves no real decisions ranks lower than a niche regulatory filing that reshapes how operators in the finance space have to behave. Read our full methodology for the scoring rubric, our glossary for term definitions, and our trends index for the longitudinal view across the beat.
| Signal on this page | What it tells you |
|---|---|
| Verified by N sources | Independent corroboration count. N≥2 is our confidence floor; N=1 is marked explicitly. |
| Impact score (1-10) | Regulatory + financial + operational weight. 8+ signals an experienced-operator action item. |
| Sentiment | Five-tier classification trained on labeled finance-specific corpora. |
| Timeline | Where applicable, the related-events sequence that contextualizes today's development. |