An Interior Department partnership with de extinction company Colossal Biosciences will build the world's largest genomic archive for species protected under the Endangered Species Act, and put a private biotech in charge of the platform holding the data.
The federal government is trying to do something that has never been done at this scale: build a single genomic archive containing reference DNA, tissue, and reproductive cells for every species protected under the US Endangered Species Act. The architecture of that effort, more than the science, is the story worth following. An administration that has spent its term narrowing the Act's regulatory reach is now the public owner of the most comprehensive endangered-species genetic resource ever attempted, and the platform holding it will be operated by a private biotech whose flagship work has drawn sharp questions from independent geneticists.
The partnership was announced this month between the US Fish and Wildlife Service, the Interior Department, and Colossal Biosciences, a five-year-old Dallas-based startup valued above $10 billion. Under the agreement, Colossal's Dallas lab will house physical samples (DNA, tissue, and reproductive cells) for roughly 2,300 species protected under the Act, while the federal government retains ownership of those samples. The company is fronting the cost; Wired reports that CEO Ben Lamm has put the figure "in the tens of millions of dollars" range. The work is the US arm of a larger Colossal Foundation project called the BioVault, which began with a "Colossal 100" priority list of endangered vertebrates and has separately drawn a $60 million investment from the government of the United Arab Emirates.
The structural question is not whether the archive is useful. It clearly is. A reference genome is the baseline dataset conservation geneticists need to identify populations, design assisted reproduction, and run any future gene-editing rescue. The more than 2,300 species protected under the ESA include vertebrates, invertebrates, and plants, and the technical and legal value of having sequenced reference data for all of them is not in dispute. The question is who controls the platform underneath that archive, and on what terms.
That question is sharpened by the contradictions the partnership now sits on top of. According to Ars Technica's reporting on the program, the same administration that has narrowed the Act's reach and limited the circumstances under which species can be listed is now the public owner of the largest genomic archive ever attempted for ESA-protected species. The samples will be public property. The sequencing know-how, the lab infrastructure, and the licensing terms on the platform's underlying technology will not be.
The credibility question is the other layer. Colossal's most public scientific achievement to date is its 2025 announcement of "dire wolf" pups, produced through gene editing and reproductive cloning. Independent geneticists have pointed out that the resulting animals were grey wolves with a small number of edits, and have publicly contested how the company is defining a "species" in its de-extinction work. That credibility question now attaches to the company that will be the US government's primary genomics partner for the next several years. The BioVault is not the dire wolf project, since the work is reference sequencing and cryopreservation, not editing, but the same firm is doing it.
That distinction matters. A reference genome is, scientifically, a foundational resource: it is the digital backbone against which any future genetic analysis of a species is calibrated. The cryopreserved tissue and reproductive cells that Colossal will store are the physical insurance policy for assisted reproduction, gene-editing rescue, or reintroduction efforts that may or may not happen. Gizmodo describes the project as a "DNA Noah's Ark," and the metaphor is fair at the level of storage. The mechanism, sample in and viable cells out decades later, has real long-term conservation value independent of who builds it.
What is worth watching now is governance, not the science. Three things, in order. First, who decides access: the federal government owns the samples, but the access protocols, the cost-recovery structure, and the licensing terms on the underlying sequencing platform are not fully public. Second, whether the archive survives administration changes. A reference genome is a multi-decade resource, and the contract terms that bind a private platform to a public-good mission across political cycles are the part that actually matters. Third, what "species restoration" ends up meaning in the BioVault era, now that the same company doing the sequencing is also the loudest voice in the room on what counts as bringing one back.
The archive is the right thing to build. The partnership is the part that needs scrutiny, not because the science is wrong, but because the platform underneath it is private, the public-good scaffolding around the Endangered Species Act is being cut back at the same time, and the company writing the rules on "de-extinction" is the same one the government has just made its primary genomics contractor.