Lambda, the GPU cloud company, is deploying Vertiv liquid cooling systems across its AI data centres, the companies said in coverage by Data Center Dynamics, standardising on direct liquid cooling as its rack densities climb past what air can serve.
The announcement is one of the week’s least surprising and most telling. Liquid cooling has moved in two years from exotic option to default specification for AI halls, and the vendor relationships being locked in now, between the neoclouds who rent the compute and the thermal companies who keep it alive, will structure the industry’s cost base for a decade. Vertiv’s position in that layer is why its order book is read as an AI capex proxy in its own right.
For Lambda the deployment is capacity strategy: cooling headroom is sellable megawatts, and in a market where customers queue for high density racks, the operator who can cool them fastest takes the contract.