US multinational law firm Latham & Watkins has bought Nvidia GPU servers and started customising its own AI models, the Financial Times revealed yesterday (11 September).
Latham, the US’s second-largest law firm with $8.3bn in revenue last year, has in recent years purchased several servers, each of which holds multiple GPUs, the powerful computer chips that run AI models, the FT said.
Advertisement
One of the driving forces is the ability to keep clients’ sensitive data entirely within its own systems.
“Sometimes we may have information that is so sensitive, client information that we really want to protect, we don’t want to put it to any cloud vendor,” Rene Mendoza, the firm’s chief information officer, told the FT.He added that “with a lot of consumption costs coming” for how AI usage is priced, the firm wanted to have “flexibility”.
Advertisement
The approach is markedly different from that taken by most major law firms, which have largely relied on cloud infrastructure and third-party legal AI platforms such as Harvey and Legora, or built their own applications while continuing to consume underlying models and compute through external providers.
Latham is going further down the technology stack. By owning GPU infrastructure and customising open-weight models itself, the firm can potentially run certain AI workloads without sending client data to either a legal AI vendor, model provider or hyperscaler.
That gives Latham greater control over confidentiality, model choice and potentially cost, particularly as vendors begin to move away from subsidised token use to a consumption basis. It does come with a significant trade-off. Running GPU infrastructure and models in-house requires substantial capital investment and specialist technical expertise, and places responsibility for maintaining and securing the infrastructure squarely on the firm.
Keep an eye out for more details.