Battle Moves From Hype to Burnout, Billing and Platforms
Mega firms like Kirkland & Ellis and Latham & Watkins are turning their AI arms race into a battle pitching associates and the onset of AI into a knife-edge fight that directly challenges the 2000 hour grind, the traditional apprenticeship model and identifying what top-of-the-market salaries mean for associates these days.
Kirkland & Ellis has earmarked US$500 million to build a proprietary AI platform trained on its own deal history and partner judgment, betting that “owning the brain” of the firm will deliver faster, smarter work and a new edge over rivals while firms like Latham & Watkins push equally hard on AI‑native training and ecosystem tools to stay in the game.
As Kirkland boss Jon Ballis (pictured) said, “The idea is that we’re going to take the collective intelligence of our institution and be able to deploy that throughout our firm.”
Kirkland & Ellis’s US$500 AI bet is the clearest sign yet that the BigLaw AI arms race has moved beyond vendor hype into hard questions about burnout, billable hours and what it means to be an associate in 2026.
While Latham & Watkins is positioning itself as an AI‑heavy “ecosystem player” with AI‑native lawyers and client‑facing AI advisory practices, Kirkland is doubling down on owning both the platform and the data that underpin its core legal work.
Kirkland: build the platform, train it on your deals
Kirkland idea to commit US$500 million of its own revenue over the next three to four years to build a proprietary AI platform trained on the firm’s drafting history, workflows and partner input has been framed by firm bosses as a way to “take the collective intelligence of our institution and deploy that throughout the firm”.
The firm expects to spend more than US$100 million in year one, with hundreds of millions more to follow, while continuing to license external AI tools where they don’t create long‑term differentiation.
Kirkland have around 250 lawyers, including 100 partners, feeding detailed information about how they actually work into the system.
The platform is intended to run across entire mandates, not as a series of disconnected point solutions, which is a notable shift from the “copilot everywhere” approach many peers have taken.
We covered this story in May, but the way firms are approaching the AI issues continues to evolve rapidly – and interestingly for observers of the AI race.
Latham: build AI‑native lawyers and client‑facing AI practices
Latham, by contrast, has become one of the most AI‑public firms, and have tasken a different approach.
They emphasise AI‑native lawyers, a dedicated global Artificial Intelligence practice and mandatory AI training for new associates. The firm runs a multidisciplinary AI team advising innovators, investors and corporates across the “AI stack”, effectively turning AI itself into a high‑value practice area.
Inside the firm, Latham has invested in structured AI education and AI academy that includes large‑scale training sessions for hundreds of young lawyers with a clear message to use the tools, but “keep your hands on the wheel” and personally review outputs.
Its global M&A co‑chair recently told Bloomberg that AI is already being integrated into “every transaction” to enhance speed and intelligence, underlining how deeply embedded these tools have become in Latham’s deal work.news.
Burnout, billable hours and the 2,000‑hour problem
The core tension for associates is simple – AI removes hours, but firms still expect 1,800–2,000+ billable hours per year. As AI speeds up routine tasks and makes them more predictable, work that used to support classic billable structures is drifting towards flat, subscription and hybrid pricing models, while complex high‑uncertainty matters remain hourly.
That shift collides directly with burnout. Survey work across mid‑sized firms shows that when AI is deployed thoughtfully, cutting cognitive load and tedious volume, nearly half of legal professionals report being more likely to stay.
But when AI is layered on top of unchanged billing targets, it can feel like intensification with more pressure to move faster, without real relief on hours, as legal media have reported.
And so some firms are experimenting with how best to handle the issue. One high‑profile AI programme has allowed first‑year associates to spend hundreds of hours of their annual quota on AI experimentation and training, rather than client‑billable work, directly addressing the “2,000‑hour problem”.
Kirkland’s leadership has gone a step further rhetorically, openly linking its AI investment to a shift away from traditional billable hours and toward value‑based pricing, signalling a willingness to “lean into” models that monetise outcomes rather than time.
Legal AI platforms: vendor stack versus proprietary spine
The Kirkland–Latham contrast is particularly stark around legal AI platforms. Kirkland is building an internal spine as a firm‑specific platform informed by hundreds of lawyers and supported by more than 180 AI engineers and data scientists, including dozens of specialist AI engineers working via alliances and in‑house roles.
The objective is to make the firm’s own matter history, drafting patterns and decision‑making the core asset, and to treat vendor tools as interchangeable layers above it.
Latham, on the other hand, has leaned into an ecosystem approach, integrating multiple external platforms, emphasising AI‑native lawyers and building out an AI practice that advises on regulation, deals and disputes across industries.
In that model, the firm’s competitive edge is less about owning the platform and more about having the best people advising on, and orchestrating AI technologies for clients.
For associates, these two approaches feel different day‑to‑day. In a Kirkland‑style proprietary environment, junior lawyers may spend more time co‑designing workflows, testing outputs and imposing their own practice experience into the platform. They are effectively acting as internal product managers and AI supervisors.
In a Latham‑style ecosystem, associates are more likely to rotate between tools, matter types and industries, building AI literacy alongside traditional black‑letter practice.
We recently reported on Willkie Farr use their ‘Willkie Works’ to deploy AI systems through their own ecosystem within the firm.
Associate billing, retention and the flight‑risk question
The question now is does AI make associates more or less likely to stay with a firm doing this sort of implementation with AI? Evidence from broader legal industry research suggests AI can cut burnout, reduce repetitive strain and increase perceived autonomy, but only if firms explicitly bake AI training and experimentation into the billing model.
Where firms formalise AI‑specific hours, invest in AI academies and give juniors visibility into how AI shapes their careers, retention indicators improve.
Kirkland’s big money investment in its AI platform is a double‑edged sword in that respect. On one side, it offers associates access to a uniquely rich, firm‑trained system and the chance to work at the frontier of AI‑enabled deal and disputes work.
On the other, if that platform is seen also as a way to shrink junior classes and squeeze more throughput from fewer bodies, as Harvard Law School have observed, without meaningfully changing hours expectations, mid‑levels will have ample incentive to jump to boutiques, ALSPs or in‑house roles that promise deeper product exposure and more control over AI strategy.
Latham’s visible investment in AI training and AI‑native practice areas is one way to keep that narrative positive, signalling to associates that AI is a career accelerator rather than just a margin tool.
Whether Kirkland and other ultra‑profitable firms match that with similarly explicit billing and training reforms will go a long way to determining where the next wave of BigLaw talent chooses to build its AI‑age careers.lw+8
