James van Geelen’s AI Doom Thesis Triggers Market 'Citrini Selloff'
Key Takeaways
- Analyst James van Geelen has detailed a catastrophic 'doom scenario' for the AI sector, arguing that unsustainable capital expenditures and lagging returns are priming the market for a massive correction.
- The thesis, popularized via the Citrini platform, has sparked a significant market reaction now known as the 'Citrini Selloff.'
Key Intelligence
Key Facts
- 1James van Geelen's 'AI Doom Scenario' argues that AI capital expenditure is currently unsustainable relative to revenue.
- 2The 'Citrini Selloff' refers to the market volatility triggered by the viral spread of this bearish thesis.
- 3Hyperscalers have invested an estimated $150B+ in AI infrastructure in 2025 alone, creating a high-stakes 'ROI gap'.
- 4The thesis suggests a potential 'Minsky Moment' where the AI investment bubble bursts due to diminishing returns on scaling.
- 5Bloomberg's Odd Lots podcast featured van Geelen to discuss the systemic risks of the current AI hardware boom.
Who's Affected
Analysis
The AI industry is currently grappling with a profound narrative shift as James van Geelen’s 'AI Doom Scenario' moves from the fringes of financial research to the center of global market discourse. For the past three years, the dominant market logic has been one of 'infinite demand' for compute, driven by the belief that generative AI would rapidly transform every sector of the global economy. However, van Geelen’s thesis, amplified through Bloomberg’s Odd Lots and his own Citrini platform, suggests that we are approaching a 'Minsky Moment' for artificial intelligence—a point where the debt-fueled investment in infrastructure far outpaces the actual utility and revenue generated by the technology.
At the heart of the 'Citrini Selloff' is the concept of the AI Capex Trap. Since 2023, hyperscalers like Microsoft, Alphabet, and Meta have collectively poured hundreds of billions of dollars into NVIDIA-powered data centers. Van Geelen argues that this investment cycle is fundamentally different from the internet boom of the late 1990s. While the build-out of fiber optics eventually enabled the modern web, the current build-out of GPU clusters carries massive depreciation costs and energy requirements that demand immediate, high-margin returns to remain viable. If enterprise software companies cannot prove that AI is driving significant productivity gains or new revenue streams within the next 12 to 18 months, the funding for this infrastructure will likely evaporate, leading to a systemic collapse of the hardware supply chain.
The AI industry is currently grappling with a profound narrative shift as James van Geelen’s 'AI Doom Scenario' moves from the fringes of financial research to the center of global market discourse.
The market impact of this thesis has been swift and severe. The 'Citrini Selloff' reflects a growing skepticism among institutional investors who are no longer satisfied with 'AI-integrated' product announcements and are instead demanding hard data on ROI. This shift in sentiment is particularly damaging to mid-tier AI startups and hardware providers that have relied on the assumption of perpetual growth in compute demand. Van Geelen’s scenario posits that as the 'hype premium' is stripped away, we could see a valuation reset comparable to the 2000 dot-com crash, where even fundamentally sound companies saw their stock prices decimated by the broader industry contagion.
What to Watch
However, van Geelen’s perspective is not merely a call for a market correction; it is a fundamental critique of the current AI research trajectory. He suggests that the industry’s obsession with 'scaling laws'—the idea that more data and more compute will inevitably lead to more capable models—may be hitting a point of diminishing returns. If the cost of training the next generation of models (e.g., GPT-5 or its equivalents) increases exponentially while the performance gains are only incremental, the economic justification for the entire AI ecosystem begins to crumble. This 'research wall' is a critical component of the doom scenario, as it implies that the massive infrastructure being built today may become obsolete before it ever pays for itself.
Looking forward, the industry must now contend with a more disciplined investment environment. The 'Citrini Selloff' serves as a wake-up call for AI researchers and executives to pivot from 'capability at any cost' to 'efficiency and economic viability.' The coming quarters will be a litmus test for the industry: either companies will begin to show the tangible productivity gains that justify the $100 billion-plus annual spend, or van Geelen’s doom scenario will transition from a viral thesis to a permanent market reality. Analysts should watch for signs of Capex rationalization among the 'Magnificent Seven' and a potential consolidation in the LLM provider space as the 'easy money' era of AI development comes to an abrupt end.
Timeline
Timeline
The Capex Boom
Massive investment in NVIDIA H100 and B200 GPUs by hyperscalers and AI startups.
The ROI Question
Institutional investors begin questioning the lack of AI-driven revenue in enterprise software earnings.
Citrini Report Viral
James van Geelen publishes his detailed AI Doom Scenario, gaining traction on social media and financial forums.
Odd Lots Feature
Bloomberg's Odd Lots podcast interviews van Geelen, mainstreaming the 'Citrini Selloff' narrative.
Market Reaction
Significant volatility in semiconductor and cloud computing stocks as the 'Doom Scenario' is priced in.
How we covered this story
Every story in our ai 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 ai 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 ai-specific corpora. |
| Timeline | Where applicable, the related-events sequence that contextualizes today's development. |