TL;DR

London startup Basecamp Research has built EDEN — described as one of the largest biological AI models ever trained — using a genomics dataset drawn from millions of species collected from deep-sea volcanoes, remote rainforests and Antarctic ice caps. Backed by Nvidia, the company is now demonstrating therapeutic enzyme design in its Boston lab.

From ice caps to algorithms

Co-founders Oliver Vince and Glen Gowers started with a realisation during their Oxford and Imperial College PhDs: designing sophisticated new medicines would require computers that genuinely understood biology, and that meant collecting DNA at a scale nobody had attempted.

In 2019, they ran a proof-of-concept on an Icelandic ice cap with a solar-powered, sledge-mounted mobile DNA sequencing lab. That field-first approach has since expanded to sampling from shipwrecks, rubbish dumps, active volcanoes and remote ecosystems worldwide.

The resulting dataset, BaseData, feeds EDEN — a cluster of AI models trained on DNA sequences and evolutionary patterns. Nvidia provides the computing backbone for what Vince calls “the thing that understands biology the best.”

What EDEN can do

In January 2026, Basecamp announced it had designed a therapeutic enzyme capable of inserting large amounts of DNA at a precise point in the human genome to treat specific diseases. The approach resembles CRISPR gene editing but with larger genome modifications and without the DNA damage associated with existing techniques.

The applications span cancer (inserting DNA into the immune system to produce cancer-fighting cells), rare genetic diseases (repairing faulty genes), and antimicrobial resistance (designing novel peptides to combat superbugs). The company’s Boston lab is currently running demonstrations.

The business model

Basecamp operates at the intersection of fieldwork, data and AI. Medical researchers can input a patient and disease profile and have EDEN reason out a targeted therapeutic approach. Current drug development can take years and cost millions; AI-driven design aims to compress that timeline, particularly for rare genetic disorders where patients cannot wait for traditional processes.

Looking forward

For the UK’s life sciences sector, Basecamp represents a different kind of AI company — one whose competitive advantage sits not in model architecture but in proprietary biological data collected through years of physical fieldwork. Whether EDEN’s laboratory demonstrations translate into clinical therapies will be the test, but the model of pairing real-world data collection with AI-driven drug design offers a template for UK biotech.