$500M and Counting: Inside Big Law's AI Budget War

Kirkland & Ellis has committed $500 million to building a proprietary AI platform, and now the firm is putting real weight behind that number. The world’s highest-grossing law firm, with $10.6 billion in annual revenue, has already deployed $100 million of the commitment this year, alongside a multiyear partnership with Palantir Technologies to build tools for its private equity fund formation practice.

“We are building this revolutionary fund formation engine to help clients navigate that complexity,” said Erica Berthou, a Kirkland partner and member of the firm’s global executive committee. Palantir’s chief legal officer, Ryan Taylor, was more direct about what the partnership signals: “Kirkland is defining what the next generation of professional services will look like.”

The firm’s AI and innovation team has grown to more than 180 AI engineers and data scientists, according to Bloomberg Law’s reporting on the buildout, along with more than 250 Kirkland attorneys, including about 100 equity partners, directly embedded in the build through internal “AI Pods.”

More than 1,000 Kirkland lawyers are expected to use the resulting system, which is designed to link funds, transaction history and investor obligations into a single operational platform rather than function as a standalone research tool. Few technology commitments of this size have come from a law firm. Fewer still have this much lawyer buy-in baked into the build itself.

Kirkland is the most visible signal of where 2027 budget dollars are headed, but it’s far from the only firm making that bet. This move sits at the leading edge of an industrywide reallocation that’s still small in percentage terms and compounding fast.

Speaking to Bloomberg Law, a Kirkland spokesperson framed the buildout as a talent strategy as much as a technology one: the firms that succeed won’t just win the fight for top lawyers; they’ll combine those lawyers with engineering, data infrastructure and operational scale.

The Industry Trajectory, in the Data

Firms spent 0.25% of total revenue on AI at the end of 2025, according to Citi’s law firm advisory group, more than double the 0.11% of revenue AI spending accounted for at the end of 2024. “Firms continue to expect that spend to accelerate,” said Daniel Greenfield, a director in Citi’s group.

The percentage still looks modest next to the rest of a firm’s budget, but the doubling rate is the real story: a line item that small, compounding at that pace, stops being a rounding error within a few budget cycles, not many.

The acceleration is recent. Citi Hildebrandt’s year-end advisory for 2025 measured demand growth at just 1.9%, with revenue up 11.3%, in a report the firm and Hildebrandt Consulting released that December. Demand growth more than doubled within the first half of 2026 alone.

That AI spending is coming from a position of strength, not desperation. Firms saw revenue grow 11.7% and billable-hour demand climb 4.2% in the first half of 2026, well above the historical 1.5% to 2% demand average, according to Citi’s law firm advisory group. Growth ran even higher, at 13.1%, among the 50 largest firms by revenue, with overhead expenses at those firms climbing 12.6%. Total expenses across the industry rose 9.7% over the same period, with compensation up 8.9% and overhead up 10.5%.

Demand growth wasn’t evenly distributed, either. Am Law 50 firms posted 4.6% demand growth, well ahead of the 3.6% among the rest of the Am Law 100 and the 3.9% at Am Law Second Hundred firms, according to Above the Law’s coverage of the same Citi data.

Firms have also been raising rates to fund the buildout. Worked rates, the actual rates clients agree to pay, climbed 7.1% in the second quarter of 2026 alone.

The Build-vs-Buy Fork

Industrywide AI spending is expected to steeply accelerate in 2026 and beyond, Gretta Rusanow, managing director and head of advisory services at Citi’s law firm group, told Bloomberg Law. The strategic question firms face now is whether to build that capacity themselves or buy it from somewhere else.

Kirkland is the clearest example of building. DLA Piper is taking a middle path by “moving towards an enterprise solution” that leverages its scale and proprietary data, said Loren Brown, the firm’s U.S. vice chair, though he expects DLA Piper will still rely on commercial tools alongside whatever proprietary system it builds.

Jane Rogers, a member of Ropes & Gray’s management committee, described a third option: her firm is weighing a fully proprietary platform against a more flexible “open architecture” strategy that swaps underlying language models in and out rather than committing to one build.

For most firms, the choice isn’t really a choice. Beyond the largest 50 firms, “the build-versus-buy selection or strategy is not applicable to anyone,” said Alex Finkel, chief innovation officer at Offit Kurman.

In the same report, roughly 67% of large firms reported spending under 2% of their 2025 budgets on new technology, a level nowhere near what building a proprietary platform requires. Smaller firms spend differently, but not necessarily less: nearly 30% of firms with fewer than 500 attorneys spent more than 5% of their budget on new technology last year, and 1 in 3 sub-500-attorney firms spent more than 8% of their budget on technology subscriptions and recurring costs.

Scale determines strategy here. Large firms build, and now have the partnerships and headcount to prove it. Smaller firms subscribe, and pay a proportionally higher share of their budget to do so.

But building isn’t automatically the safer bet even for firms that can afford it. Kirkland’s scale and its Palantir partnership are designed specifically to manage that risk, pairing external engineering expertise with internal legal judgment instead of building in isolation.

That combination, not the dollar figure alone, is what smaller firms lack when attempting to build on their own.

The Real Per-Attorney Cost

The costs get more concrete once they’re translated into dollars per lawyer. Firms are spending $50 to $350 per attorney each month on AI tools, Greg Lambert, chief innovation officer at Jackson Walker, told Bloomberg Law, with AI adding roughly 30% to the cost of legal-research and practice tools firms already had before the AI wave hit.

Lambert added that, once the AI-driven cost increases embedded in existing tools are counted alongside dedicated AI spending, firms may actually be spending closer to 0.5% of revenue on AI, roughly double Citi’s headline 0.25% figure.

The gap between those two numbers isn’t a contradiction. It’s a reflection of how hard the true figure is to pin down when so much AI spending arrives disguised as a price increase on tools firms were already paying for, rather than as a clean, separately budgeted line item.

One data point puts those industry percentages in sharp relief. McDermott Will & Schulte chairman Ira Coleman has estimated the firm’s AI spending at about $30 million a year. If that holds, that would equal roughly 1.3% of the firm’s 2024 revenue, more than 10 times what Citi’s survey found the typical firm spending on AI, according to the same Bloomberg Law reporting on Lambert’s estimates.

Some of that gap may not last. In the same Bloomberg Law piece, chief innovation officers and legal technology consultants said that many firms are testing AI subscription tools on a trial basis, with limited licenses or discounted rates from vendors eager to market the association. Those trials and sweetheart deals are unlikely to hold once vendors move firms onto standard pricing.

Why Firms Are Scaling Up via M&A

Not every firm is scaling AI through budget lines alone. Some are getting there through mergers, and 2026 has been an active year for exactly that kind of consolidation across Big Law. It has also shown the downside: Weil, Gotshal & Manges is winding down after its merger talks with Willkie Farr ended.

Coleman said part of the firm’s rationale for its 2025 merger with Schulte Roth & Zabel, which created a 1,750-lawyer, nearly $3 billion firm, was gaining the scale needed to pay for AI. “You can’t do that if you’re a four or five hundred lawyer firm,” Coleman said during an episode of the “On the Merits” podcast.

That’s a notable reframing of a familiar Big Law growth strategy. Law firm mergers have traditionally been sold on practice-area coverage or geographic reach. But Coleman describes AI infrastructure spending as a reason to merge in its own right, on par with those older rationales, and it’s hard logic for a firm outside the top ranks to argue with.

If that reasoning spreads across the current wave of consolidation, AI budget season and merger season stop being separate stories, and firms without the appetite to build or the revenue to license at scale may find themselves looking for a merger partner instead of a vendor contract.

What 2027 Budgets Need to Answer

The big, unresolved question heading into next year isn’t really about the percentage. It’s about what the spending is actually being used for. Firms need to decide whether AI investment is a client-facing differentiator they can point to in pitches, the way Kirkland’s Palantir partnership already functions as one, or whether it primarily remains an internal efficiency play that never shows up in a pitch deck at all.

Rusanow herself has called the current spend figure “surprisingly low” relative to AI’s expected impact on the industry, suggesting that most of the ramp is still ahead of firms and not behind them. That leaves 2027 as the year firms have to show their work.

Kirkland has the most money and the most public commitment riding on the answer. Whether the firm’s Palantir-built platform delivers a client-facing edge that a general counsel can point to when choosing the next firm will be the test of 2027 budgets.