Login to access
Want to subscribe?
This article is part of: Edge Insights
To find out more about how to join or access this report please contact us
As AI workloads push rack densities higher and liquid cooling becomes essential infrastructure, a fragmented vendor landscape is ripe for consolidation.
Welcome to our ‘Coffee break reading’ series. These pieces are designed to be short and easy to read, presenting our opinion on a hot topic and challenging you to think about it too.
As liquid cooling moves from specialist HPC deployments into the core of AI data centre design, CDUs are emerging as a strategic control point in the infrastructure stack. What was once a fragmented ecosystem of niche vendors is quickly becoming a critical battleground for infrastructure OEMs, HVAC providers and liquid cooling specialists seeking to capture a larger share of AI infrastructure spend. As deployments scale and reliability requirements rise, operators increasingly favour partners that can deliver integrated, globally deployable cooling platforms rather than standalone components.
The result is a market likely to consolidate around a smaller number of vendors capable of combining liquid cooling expertise with manufacturing scale, global supply chains and service capabilities. For investors, operators and technology suppliers alike, the next phase of the liquid cooling market will be defined not just by innovation in cooling technologies, but by which companies can assemble the platforms and partnerships needed to support the rapid expansion of AI infrastructure worldwide.
Our insight covers the following key points:
- Why has liquid cooling become so pivotal?
- CDU architectures: Three form factors and how operators deploy them
- In-rack CDUs
- In-row CDUs
- Bank/wall-mounted CDUs
- The liquid cooling vendor ecosystem
- Recent transactions illustrating consolidation
- Why more consolidation is inevitable
- Implications for the ecosystem
Related research
- Network edge data centre forecast: Building the base
- Telco AI data centre forecast: Where will the compute live in the AI era?
