Understand this story
Citrini Research founder James van Geelen offered a judgment, arguing that changes in the financing structure of AI infrastructure have made market sentiment part of the fundamentals. Speaking on the Monetary Matters podcast, he said that incremental funding for AI infrastructure has shifted from cloud providers' own cash flow to equity and debt financing, and that a downturn in sentiment would directly cut off the funding chain. He estimates that supporting current infrastructure valuations would require AI lab ARR to reach about $1 trillion by the end of 2028, versus a current base of about $13.6 billion; he also cited DeepSeek's internal inference framework data, saying it serves 155 trillion tokens per day, implying an inference gross margin of about 80%. On positions, he is long Meta and short Google, and said the condition for returning to chips is sustained growth in paid usage and a re-acceleration in the second derivative of ARR growth. Citrini Research, once among the earliest bulls on AI hardware, has substantially cut its semiconductor holdings and holds some short positions.
AI-generated from public reports
Developments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 26
Wallstreetcn
Citrini, the earliest bull on AI hardware, has significantly cut its semiconductor positions, saying changes in financing structure have made sentiment a fundamental
Citrini Research, the chip research firm that was once the earliest bull on AI hardware, has significantly cut its semiconductor positions and holds some short positions. Founder James van Geelen said on the Monetary Matters podcast that incremental funding for AI infrastructure has shifted from cloud providers' own cash flow to equity and debt financing, and that a downturn in sentiment would directly cut off the funding chain. He estimates that supporting current infrastructure valuations would require AI lab ARR to reach about $1 trillion by the end of 2028, versus a current base of about $13.6 billion, and cited DeepSeek's internal inference framework data as saying it serves 155 trillion tokens per day, implying an inference gross margin of about 80%. In terms of positions, he is long Meta and short Google, and the conditions for returning to chips are continued growth in paid usage and a re-acceleration in the second derivative of ARR growth.