Large law firms divide on building or buying artificial intelligence as new technology spending remains low

Roughly two-thirds of large law firms spent less than 2% of their budgets on new technology in 2024, according to data shared with Bloomberg Law's Leading Law Firms survey. Smaller firms moved faster, with nearly 30% of those under 500 attorneys allocating more than 5% of their budgets to new tools. Industrywide AI spending is forecast to steeply accelerate in 2026 and beyond, said Gretta Rusanow, managing director and head of advisory services at Citi's law firm group. Sixty-six firms of all sizes broke out spending in both new technology and recurring subscription categories. At firms with more than 1,000 lawyers, outlays on technology subscriptions dwarfed new-tech investment. One in three firms with fewer than 500 attorneys, by contrast, said they spent more than 8% of their 2025 budgets on technology subscriptions and recurring costs.

The build-or-buy fork

As legal AI adoption spreads, a cluster of top-tier firms is directing money into proprietary tools rather than off-the-shelf subscriptions. Kirkland & Ellis recently announced plans to spend $500 million on its own AI model. DLA Piper is "moving towards an enterprise solution" that uses scale and proprietary data, though outside AI tools will still likely play a role, said Loren Brown, the firm's US vice chair. "We spend a lot of time thinking about to what extent are we going to build our own independent AI platform" versus a more flexible "open architecture" strategy, said Jane Rogers, a member of Ropes & Gray's management committee. Other large firms are building internal AI platforms that integrate proprietary workflows and let lawyers swap out large language models, according to a chief innovation officer who spoke on condition of anonymity. For that firm, subscription tools like Harvey and Legora serve as "a bridge."

Why some firms are buying

Rogers said the build-or-buy call is as much a business decision as a technology one. Ropes & Gray has chosen to buy. "The technology is changing so quickly now that it doesn't appear to us at this point that doing something proprietary makes sense," said Marsha Stein, the firm's chief information officer. "Right now, the best option, we think, is to actually buy." Harvey co-founder Gabriel Pereyra pushed back on the binary framing. He sees a false choice that wrongly assumes "either firms are going to build everything themselves, or a company like us or a model provider is going to build everything themselves." The debate is unfolding as firms prioritize the ability to swap AI tools for the latest technology.

Small firms are priced out

For most firms, building proprietary AI is not an option. "The build-versus-buy selection or strategy is not applicable to anyone" outside the top 50 firms, said Offit Kurman Chief Innovation Officer Alex Finkel. The numbers bear that out. Roughly 67% of large firms reported spending under 2% of their 2025 budgets on building or buying new technology. Without economies of scale, smaller firms pour a higher share of spending into recurring subscriptions. The market is becoming more segmented, putting well-capitalized firms in a position to extend their advantage - even if they choose not to build.

Why this matters for legal professionals

The spending data reveals a legal market splitting into tiers. Partners and in-house teams at large firms should watch how proprietary AI tools affect pricing models, client expectations, and data security. Lawyers at midsize and boutique firms, where budgets skew toward subscriptions, need to evaluate whether those tools deliver measurable productivity gains or simply add to overhead. For law firm leaders mapping out technology roadmaps, the decision to build or buy is already shaping competitive positioning - and the gap could widen quickly as AI investment accelerates next year.