Direct-to-chip liquid cooling is becoming the standard approach for hyperscalers as rack densities continue to rise to support AI infrastructure. But immersion cooling still has a place, particularly at the edge and far edge, where the requirements can look very different. The two technologies are often discussed as competing approaches, despite serving different needs. For operators, the choice will depend on where cooling is being deployed, the infrastructure available and the workloads it needs to support. This session will look at where DLC and immersion make sense, and why the industry may need both rather than a single approach. This episode will discuss:
- Why hyperscalers are moving towards direct-to-chip cooling
- Where immersion cooling could make sense at the edge and far edge
- What drives the choice between DLC and immersion for different deployments
- How both technologies could find their place as liquid cooling adoption grows
