Basecamp Research announced the Trillion Gene Atlas, a coalition with Anthropic, Ultima Genomics and PacBio to generate enormous environmental and organismal sequence datasets and feed them to AI models for therapeutic design. The initiative will combine short‑ and long‑read sequencing and NVIDIA computing to scale evolutionary diversity by roughly 100× over current public resources, the companies said at SXSW and GTC. Basecamp emphasized that the expanded training corpus will enable foundation models (EDEN) to learn from evolutionary fitness and design biologics and therapeutic candidates directly from disease prompts. Partners plan deep sequencing across tens of thousands of samples and integration into large‑scale AI pipelines. The project signals intensifying convergence of biodata generation and generative AI for drug discovery. If successful, the atlas would materially alter training data availability for biological LLMs and could compress early discovery timelines; it also raises questions about data governance, conservation sampling and equitable access to derived therapeutics.