Latham is buying its own Nvidia GPUs, standing up private infrastructure, and fine-tuning open AI models – while keeping every major AI vendor on speed dial.
They don’t have to know whether OpenAI, Anthropic, Google, Nvidia, or someone we haven’t heard of yet wins. It can use the winner, build around them, or walk away.
This isn’t an AI purchase. It’s a leverage play. And at Latham’s size, leverage gets very profitable very quickly.
1. Optionality Is the Real Weapon
This shrewdness isn’t better AI. It’s never getting trapped by one vendor’s limitations, security requirements, prices, or roadmap. If a new model crushes today’s leader? Latham can move. When an open model becomes good enough? Latham can run more of it privately.
Maybe more importantly – when a client doesn’t want its most sensitive information touching an outside model? Latham has another place to put the work. The firm doesn’t need to build everything itself. It only needs enough capability to offer the right tool and approach.
Latham’s scale makes this bet possible in a way most firms simply can’t match. Expensive infrastructure and engineering talent can be spread across thousands of lawyers, enormous clients, and mountains of premium work. While the amounts invested are undisclosed, sources estimate the investment to be at least $45 to $55 million annually – rivalling Kirkland as a percent of revenue.
Most firms will become increasingly dependent on AI vendors. Latham is spending money to make AI vendors compete for Latham. This is strategic leverage.
2. Confidentiality Becomes the Pitch
Latham may have found a way to turn one of AI’s biggest client anxieties into a reason to hire Latham.
The most valuable matters often contain the biggest secrets: the undisclosed deal, board investigation, activist attack, CEO crisis, government investigation, or bet-the-company litigation. These are precisely the matters where clients can be most nervous about AI touching their information.
Major AI providers already offer serious enterprise security. Running the model itself gives Latham more control. Latham can bring AI firepower to clients’ most sensitive problems while running selected workloads inside infrastructure they control. It isn’t an IT feature – it’s business development.
The more sensitive the matter, the more valuable control becomes. The more valuable control becomes, the more Latham differentiates itself on exactly the work every elite law firm wants most.
3. The Margin Play Everyone’s Sleeping On
This may be the biggest economic prize. Clients don’t pay Latham millions because a matter requires 7,000 lawyer hours. They pay because the acquisition is worth $15 billion, the investigation threatens the company, or the litigation can change its future.
AI can attack the cost of producing the answer without attacking the value of the answer. Imagine Latham delivering a $2 million matter with 25% less lawyer time while producing an equal or better outcome. And faster.
The matter doesn’t suddenly become 25% less valuable to the client. Clients are happier – they got their outcome faster – preserving the fee. The difference becomes an enormous economic opportunity: produce at AI economics while pricing at elite-law-firm value.
Clients will eventually demand their share of the productivity dividend. They always do. But Latham doesn’t need to capture all of it. Capturing even a sliver becomes enormous money.
What Has to Go Right
None of this works just because Latham bought Nvidia servers. The models have to perform, the infrastructure has to deliver, and Latham has to keep elite technical talent while competing with technology companies for the same people. Latham has a history of taking on more risk and knowing what to do to make it happen.
The Bottom Line
Everyone is watching the horses: which model wins, who has the smartest AI, and who signs with whom. Latham is playing a different game – building enough infrastructure and expertise to use whichever horse wins, keep its most sensitive workloads private when it chooses, negotiate with vendors from strength, and capture more of the enormous economic upside AI offers.
And this is what AI is ultimately about.
MBR
Forwarded to you? Get your own copy every week! Subscribe below.
Stay ahead of the market with BTI's latest research
