When considering server delivery, there is often an expectation that AI is nearing its point of operation. In reality, high density AI requires a coordinated cooling solution. Whilst the server hall may be complete with all electrical distribution, rack, network and compute components, the liquid cooling loop may not have the required components to make it operable. From a commercial perspective, you are now investing capital into your equipment which isn’t going to produce the planned amount of computing resources. When considering your project timelines, you want to know when your servers will be on site and when your cooling solution is going to be able to reliably reject heat from your compute.