US giant ramps up in-house tech to build bespoke AI models
Latham & Watkins has splashed out on its own Nvidia-powered servers as the US giant looks to build and customise AI models in-house, in a move thought to be a BigLaw first.
The firm has bought several Nvidia servers, each fitted with multiple GPUs that provide the computing power needed to run hefty AI models. Rather than relying solely on outside tech providers, Latham’s own engineers are now using the hardware to tailor existing AI models to the firm’s needs.
For the non-techies, these models are known as “open-weight” AI models, which can be downloaded, customised and run on Latham’s own hardware. This gives the firm an in-house alternative while it continues to use familiar tools such as ChatGPT, Claude and Gemini.
Bringing the technology in-house also means particularly sensitive client information can stay within Latham’s own systems, rather than being sent to an external cloud provider.
Speaking to the Financial Times (£), Latham chief information officer Rene Mendoza said: “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.”
The set-up also gives Latham more flexibility over which AI providers it relies on, with Mendoza adding: “We are not hitching our wagon to one particular company.”
The move puts Latham out in front among the legal heavyweights, becoming the first major law firm publicly known to have bought its own AI hardware and fine-tuned models. Others have taken different routes, with Kirkland & Ellis working with Palantir, A&O Shearman teaming up with Harvey and Freshfields announcing a partnership with Anthropic earlier this year.
And it is no small tech project. Latham employs more than 900 technology specialists, while its servers are housed in locked data-centre space accessible only by the firm’s own staff. The firm has not disclosed how much it has spent, although a set-up of this scale, including the hardware and specialist staff needed to run it, can cost tens of millions of dollars a year.
The tech push is also changing the make-up of the firm itself. Michael Rubin, a Latham partner and chair of its AI strategy committee, said the number of technology specialists, innovation lawyers and AI-focused roles has increased significantly in recent years.