Lilly Bets Up to $2.25 Billion on Profluent's AI-Designed Recombinases, Pushing Gene Editing Beyond CRISPR Into Kilobase-Scale Territory
Eli Lilly's deal with Bezos-backed Profluent funds AI-designed recombinases that aim to insert entire genes, going beyond CRISPR's localized cuts.
Overview
Eli Lilly has signed a strategic research collaboration with AI biotech Profluent worth up to $2.25 billion in milestone payments, betting that artificial intelligence can design a class of gene-editing enzymes capable of inserting entire genes — a feat that current CRISPR-based tools cannot reliably achieve. The deal, announced on April 28, 2026, focuses on AI-designed site-specific recombinases and marks Lilly’s second recombinase pact in four months.
What We Know
Under the agreement reported by STAT, Lilly will pay Profluent up to $2.25 billion in milestone payments if every program in the multi-program collaboration succeeds. The Indianapolis pharma is also providing an undisclosed upfront payment, committed research and development funding, and tiered royalties on net sales, according to Yahoo Finance. In exchange, Lilly receives an exclusive license to develop the selected enzymes and commercialize any resulting medicines.
The technology centers on recombinases — enzymes that cut and rejoin DNA at programmable target sites. Profluent will use its AI models to design and optimize site-specific recombinases for multiple genomic targets, Yahoo Finance reported, with the explicit goal of enabling larger and more accurate DNA modifications than current gene-editing tools allow. Where CRISPR/Cas9 typically introduces small edits at a single locus, the Profluent platform aims to insert kilobase-scale stretches of DNA — long enough, in principle, to swap in entire functional genes rather than correct individual letters.
STAT framed the deal as Lilly’s bid to move “beyond CRISPR” into territory where therapies could replace faulty genes wholesale. The same report noted that Lilly’s appetite for genetic medicine is being fueled by record revenues from its obesity and diabetes franchises, and that the company has opened a new genetic medicine center in Boston while acquiring a series of gene-editing and gene-therapy companies in recent years.
A Second Recombinase Bet
The Profluent partnership is not Lilly’s first wager on recombinase technology. In January 2026, the company struck a deal worth more than $1.12 billion with Germany-based Seamless Therapeutics to develop hearing-loss treatments using the startup’s recombinase platform, as BioPharma Dive reported. Seamless, which launched in 2023 with $12.5 million in seed funding led by Wellington Management and Forbion, engineers recombinases to “precisely pinpoint and modify specific areas of the genome,” and CEO Albert Seymour told BioPharma Dive at the time that the agreement validated the platform’s broad disease-modifying potential.
Where Seamless brings a wet-lab engineering approach, Profluent contributes a generative AI stack. The two deals together suggest Lilly is hedging across the recombinase landscape rather than committing to a single technical path.
What Recombinases Could Unlock
The appeal of recombinase-based editing — and the reason Lilly is willing to backload billions in milestones — lies in the limits of existing tools. CRISPR-Cas9 and base editing are precise but local: they correct single mutations or short stretches. Many inherited diseases, however, are caused by a heterogeneous mix of mutations spread across a gene, meaning a single CRISPR therapy may only help patients carrying one specific variant.
A recombinase capable of inserting a fully functional copy of a gene would, in theory, treat every patient with a given disease regardless of their specific mutation. That is the bet underpinning the STAT-reported framing of recombinases as a path “beyond CRISPR.” The trade-off is technical difficulty: designing enzymes that target a chosen genomic site with high specificity has historically required either painstaking directed evolution or natural recombinases with limited target flexibility. Profluent’s pitch is that AI models trained on large protein datasets can generate custom recombinases on demand for any chosen site.
What We Don’t Know
Neither company disclosed the upfront payment, the size of the committed research funding, or the specific disease targets selected for the collaboration. The number of programs covered by the agreement was not made public, and no clinical timelines were announced — the pact is structured as a research collaboration, with milestone payments tied to development and commercial progress that may take years to materialize.
It is also unclear how Lilly will sequence its two recombinase bets. The Seamless deal is locked to hearing-loss indications, but the Profluent agreement covers “multiple genomic targets” without further specification, leaving open whether the two platforms will compete internally or address distinct disease areas. Neither STAT nor Yahoo Finance reported quotes from named Lilly executives on how the company sees the platforms fitting together.
Finally, no human recombinase therapy has yet entered late-stage clinical trials, and the technology’s safety profile in vivo — particularly off-target insertion risk at kilobase scale — remains an open question that no amount of milestone funding can resolve until patients are dosed.