Because artificial intelligence chips produce massive amounts of heat, standard air cooling cannot keep pace. The data center cooling market is projected to reach $40–45 billion by 2030, with liquid cooling claiming the largest share.
Product used as an collateral. CPU Central Purpose Chips - General Purpose Chips
GPU Graphics-Processing units used to train and run AI Models ---- being used for loans by neocloud firms
Target USD 700bn by the five 'hyperscale giants'
2024 Macquairie investment bank USD 500m loan to Lambda / Fluidstack. April 2025 backed by GPU chips from Nvidia
Corweave neocloud company runs 49 data centres with a compute related power capacity of one gigawatt and used chips as part of the collateral for a series of loans form alternative-asset manager Black Stone / Magnetar Capital.
Rental prices of the H100 GPU Nvidia
2xfunds leverage / COMPUTE becomes the currency of the AI era
The AI infrastructure boom is driving a historic shift in capital deployment. Infrastructure and corporate investors are rapidly pouring billions into liquid cooling solutions and dedicated power generation, as traditional air cooling is insufficient for high-density servers and AI chips
1. The Liquid Cooling Investment Wave
Because artificial intelligence chips produce massive amounts of heat, standard air cooling cannot keep pace. The data center cooling market is projected to reach $40–45 billion by 2030, with liquid cooling claiming the largest share.
Energy Infrastructure: Energy developers and clean power providers are prime targets, capitalizing on the massive premiums placed on delivering immediate electricity to high-density server campuses.