Basecamp Research published a preprint and reported that its evolutionary AI family, EDEN, can design tools for programmable gene insertion across thousands of human genomic sites. Using ~9.7 trillion biological tokens from BaseData, EDEN reportedly designed insertions at more than 10,000 disease-related loci and produced CAR T‑cells with >90% tumor‑cell clearance in lab assays. The work was developed with collaborators including Nvidia and received an undisclosed investment from Nvidia’s NVentures. Basecamp says EDEN expands CRISPR-style editing beyond small edits to place larger therapeutic DNA payloads at precise genomic locations, and that designs showed high functional success rates in peptide and antimicrobial screens. The result, presented in a company preprint, is not yet peer-reviewed; nevertheless the approach highlights how large biological datasets plus generative models are being positioned to solve longstanding delivery and design bottlenecks in gene and cell therapies.