Latham Dives into AI Infrastructure with Own GPU Servers

US law firm Latham & Watkins has taken an unconventional approach to AI adoption. Instead of relying on cloud infrastructure and third-party legal AI platforms, the firm has invested in Nvidia GPU servers and is customizing its own AI models.

The decision was driven by a desire for greater control over sensitive client data, which Latham wants to keep entirely within its own systems. As Rene Mendoza, the firm's chief information officer, explained, 'Sometimes we may have information that is so sensitive... we don't want to put it to any cloud vendor.'

The move comes as vendors begin to shift from subsidized token use to a consumption basis, which could lead to significant costs for firms using AI. By owning its own GPU infrastructure, Latham aims to gain flexibility and potentially reduce costs in the long run.

However, this approach also requires substantial capital investment and specialist technical expertise, placing responsibility for maintaining and securing the infrastructure squarely on the firm.