Nutanix is betting $20m that it can slash AI costs by running four-fifths of its workloads on open-weight models instead of costly frontier systems.
Andrew Brinded, the company's chief revenue officer, said Nutanix is investing the sum in its own GPU infrastructure with a view to running cheaper models on in-house racks for more routine tasks, while reserving blue-chip LLMs for heavier-duty operations.
