Loading
Old version — revision 1
This is a fixed snapshot of Colossal Biosciences, saved by Import as part of the initial corpus import. It is not edited and it is not updated; the article may have changed since.
Edit summary: Initial import of content/colossal-biosciences.md — the filesystem corpus, unchanged. Not an edit.
A Texas biotechnology company founded in 2021 to recreate extinct species by editing the genomes of living relatives, and to spin out conservation technology from that work.
Colossal Biosciences is a privately held biotechnology company founded in 2021 by the entrepreneur Ben Lamm and the geneticist George Church to pursue De-extinction: producing animals resembling extinct species by editing the genomes of their closest living relatives. Its stated flagship targets are the woolly mammoth, the thylacine and the dodo. It is the best-funded organization ever to work on the problem and also the most criticized, chiefly over whether an edited proxy animal can properly be described as a de-extincted species.
Colossal does not clone extinct animals: no viable cell has ever been recovered from a species that was already extinct, and the only extinct animal ever cloned, the bucardo, came from tissue banked while its last individual was still alive, as De-extinction describes. Its method is comparative: sequence ancient DNA to identify the genetic differences between the extinct species and a living relative, select a subset of those differences judged to underlie the phenotype of interest, introduce them into the relative's cells using CRISPR–Cas9 and related tools including Base editing, and produce an animal by somatic cell nuclear transfer — the technique behind Dolly and her successors — or by germline transmission. For the mammoth the living relative is the Asian elephant; for the thylacine, the fat-tailed dunnart; for the dodo, the Nicobar pigeon. Church had set out the general programme years before the company existed, alongside proposals for recoded genomes and other large-scale genome engineering.1
The company also frames itself as a conservation technology business, arguing that the tools built for de-extinction — genome assembly from degraded samples, multiplex editing, cloning, artificial gestation — apply directly to endangered species. That claim is the pivot on which most of the debate about Colossal turns.
Lamm, a serial software entrepreneur, and Church, whose Harvard laboratory had been editing elephant cells with mammoth variants since the mid-2010s, incorporated the company in 2021. It raised successive rounds at rapidly rising valuations, reported at over $10 billion in early 2025, funded by venture investors and by individuals with an interest in conservation and in the publicity the projects attract.
Beth Shapiro, an ancient-DNA researcher and author of a well-known skeptical book on de-extinction, joined as chief science officer in 2024 — a hire that gave the company scientific credibility and drew criticism from colleagues who read her earlier work as inconsistent with the company's marketing.2
The mammoth project aims to produce an Asian elephant carrying a set of edits associated with cold adaptation — hair, subcutaneous fat, haemoglobin oxygen affinity at low temperature, and temperature-sensing channels. Church's laboratory reported editing elephant cells with mammoth variants before the company existed. The company has stated a target of first calves in the late 2020s.
Three obstacles are severe. Asian elephants are endangered, and using them as surrogates for a 22-month pregnancy raises welfare and regulatory problems that Colossal proposes to solve with an ex utero gestation system that does not exist for any large mammal; the Ectogenesis literature suggests the gap between supporting a late-gestation lamb and gestating an elephant from implantation is not incremental. Elephant cloning has never been demonstrated. And the number of edits required for a genuinely mammoth-like animal is far larger than the number so far attempted, with each additional edit carrying the off-target and structural-rearrangement risks that multiplex editing accumulates.
The thylacine, extinct since 1936, has a comparatively recent and well-sequenced genome, and its living relatives are small marsupials with short gestation. Marsupial reproduction is in some respects more tractable than placental: young are born at an early developmental stage and complete development in the pouch. The genetic distance between the thylacine and the dunnart is nonetheless very large, considerably larger than between mammoth and elephant.
The dodo and moa programmes face a different obstacle. Birds cannot be cloned by nuclear transfer, so edits must be made in primordial germ cells and transmitted through a chimeric host. The technique works in chickens and has not been established in pigeons or ratites.
In 2025 Colossal announced three pups, named Romulus, Remus and Khaleesi, produced from grey wolf cells carrying, on the company's account, roughly twenty edits across fourteen genes, and described them as dire wolves. Scientific objection was immediate and near-uniform: dire wolves belong to a lineage that diverged from grey wolves several million years ago and differ at a very large number of loci, so an animal with twenty engineered changes is a grey wolf with a modified phenotype. Critics also noted that some edits were chosen to produce a white coat, a trait for which there is no evidence in the fossil record and which appears to reflect popular culture rather than palaeontology.
The company's defence rests on a functional species concept: if an animal fills the ecological role and displays the relevant traits, the label is defensible. That position has few adherents among systematists, and the episode is now the standard illustration of what critics mean when they say de-extinction is a marketing category.
What "de-extinction" deliversNo extinct species has been restored. What exists are living animals of extant species carrying a small number of edits derived from an extinct relative's genome. These animals lack the extinct species' full genome, its gut microbiome, its learned behaviour and its ecological context. Whether they are useful is a separate question from whether the name is accurate.
Colossal is venture funded, with reported valuations rising sharply across successive rounds. It has spun out two companies: Form Bio, a computational biology platform, and Breaking, which works on enzymatic degradation of plastics. The spin-outs are cited by the company as evidence that the de-extinction work generates transferable technology, and by critics as evidence that the de-extinction work is a fundraising vehicle for other businesses.
The company reports a conservation portfolio that includes cloning of red wolves from cell lines derived from a hybrid coyote population, work on cane-toad-toxin resistance in the northern quoll, and genetic rescue projects in several other species. This work is closer to established conservation genetics than the flagship programmes and has attracted less objection.
Four distinct objections recur.
Taxonomic. An edited proxy is not the extinct species. Species are defined by ancestry and by whole genomes, not by a handful of traits, and calling the result a mammoth or a dire wolf misstates what has been achieved. The IUCN's own guidance on the subject uses the term "proxy" throughout and is explicit that the products of such work are not the extinct species restored.3
Conservation opportunity cost. Ecologists have argued that resources spent on de-extinction would save more biodiversity if spent on habitat protection, and that the existence of a claimed technological fix weakens the political case for prevention. It is the same moral-hazard structure that the Precautionary principle literature identifies in other domains. The counter-argument is that Colossal's funding comes from investors who would not otherwise fund conservation at all.
Animal welfare. Elephant surrogacy, repeated cloning attempts with high embryonic loss rates, and the creation of animals with no conspecifics or natural habitat all raise welfare questions that the company has addressed in general terms.
Communication. The company announces results through press releases and documentary partnerships ahead of peer review, which makes independent evaluation slow and shapes public understanding in ways the scientific record later has to correct. This is the criticism that has done the most damage to its standing among researchers, including some sympathetic to the underlying science.
Colossal's technical contributions are real and separable from its claims. Multiplex editing at scale in mammalian cells, ancient genome assembly from heavily degraded material, and cloning in difficult species are all genuine capabilities, and the same toolkit underlies Xenotransplantation pig engineering, parts of Gene drives research, and the governance debates about who may release an engineered organism into a shared environment. Whether an elephant calf carrying mammoth edits is born in the late 2020s is the concrete test the company has set itself, and gestation remains the binding constraint.
The deeper question the company has forced is definitional and will outlast it. If a species is a lineage, no amount of editing restores one, and the entire enterprise is misnamed. If a species is a set of functions in an ecosystem, then editing living relatives is a legitimate route to restoring something valuable, and the objection is semantic. Conservation biology has not settled which it means, and Colossal has made that unsettled question expensive.
bookChurch, G. and Regis, E. Regenesis: How Synthetic Biology Will Reinvent Nature and Ourselves. Basic Books, 2012. ↩
bookShapiro, B. How to Clone a Mammoth: The Science of De-Extinction. Princeton University Press, 2015.↩Written nine years before its author became the company's chief science officer, and argues that edited proxies are not restored species.
reportIUCN Species Survival Commission. IUCN SSC Guiding Principles on Creating Proxy Species for Conservation Benefit. International Union for Conservation of Nature, 2016.↩Guidance from a conservation body with no regulatory authority; it uses "proxy" throughout and sets a conservation-benefit test rather than a taxonomic one.