In a world where science fiction keeps turning into reality, imagine stepping into a world where the ice age’s giants walk again. That is the promise made by Colossal Biosciences, a high‑profile synthetic biology company based in Austin, Texas that pursues ambitious de-extinction and conservation goals. The company was founded in 2021 by Ben Lamm, a serial tech entrepreneur, and Dr. George Church, a renowned Harvard geneticist often called one of the pioneers of modern genomics. The company brings together teams in molecular biology, genomics, reproductive physiology, and computational biology, and it is partners with academic labs and conservation organisations, with an aim to ‘restore lost biodiversity’.
Colossal’s most publicised aim is its flagship project, the de-extinction of the woolly mammoth. Using preserved ancient DNA and sophisticated gene-editing tools like CRISPR-Cas9, they intend to create an elephant–mammoth hybrid; essentially an Asian elephant modified with mammoth genes to reintroduce cold resistance, thick fur, and fat layers suited for Arctic survival. Beyond the wholly mammoth project, Colossal has announced projects targeting dodo bird, the Tasmanian tiger, and the bluebuck antelope, added most recently to their list in 2026, for revival efforts.
Colossal’s general technical approach combines several pillars. First, they study DNA and genes of the extinct creature. Scientists extract DNA from fossils, bones, or museum specimens. Next, they find the closest living relative by comparing the reconstructed extinct creature’s gene sequences to those of living species to see which one is most similar. Then they identify important differences by looking at the genes and genetic regions that likely explain the extinct animal’s distinctive traits, such as size, fur, jaw shape, cold tolerance, etc. These are candidate genes that they edit in the DNA of the living relative, using tools like CRISPR to match the extinct species’ versions of those genes. The edited embryos are developed in the lab (in vitro) and then implanted into surrogate mothers, which, if successful, gives birth to an animal that carries the edited traits. Finally, they raise and observe the offspring, how it grows, behaves, and functions. They check whether the traits are expressed as intended, and whether the animals can reproduce and pass on the traits.
In April 2025, Colossal made headlines with their effort to de-extinct the dire wolf, a large Pleistocene predator that disappeared roughly 10,000–13,000 years ago. After studying ancient DNA extracted from fossils, Colossal compared the dire wolf genome to modern canids and identified the gray wolf as the closest living relative, reporting approximately 95% genomic similarity. The company then, described to have engineered roughly 20 edits across 14 key genes thought to influence growth, cranial morphology, coat patterning, and other traits tied to the dire wolf phenotype. Colossal reported that those edits yielded three wolf cubs, six-month-old males Romulus and Remus and two-month-old female Khaleesi, that outwardly resemble dire wolves in size, skull shape, coat characteristics, and some behaviours. The company hailed it as their first successful de-extinction event, claiming that it has effectively de-extinct dire wolves as a species.
However, the announcement has sparked intense debate and criticism from independent biologists, conservationists, and ethicists. Many scientists have warned that although a 5% genetic difference may sound small, it still means hundreds of millions of base-pair differences and likely thousands of functionally important variants. Only twenty edits across fourteen genes alters only a vanishingly small fraction of the whole genome and may only reproduce certain external traits without restoring the organism’s deeper biology. Experts also highlighted that fossil and museum DNA are often fragmented, contaminated, and incomplete, leaving gaps in haplotypes, regulatory sequences, and epigenetic information, making reconstruction of a fully accurate direwolf genome technically impossible at present. Other criticisms focused on ecological, welfare, and ethical concerns, arguing that even morphologically similar edited animals may not fill the same ecological niches or restore lost ecosystem functions because environments and biological communities have changed since the original species became extinct. Hence, these skeptics conclude that Colossal only created genetically modified wolves rather than truly resurrected direwolves.
Facing mounting criticism, Colossal responded with an important rhetorical and conceptual shift that the company itself described as a ‘major departure from earlier statements. The company clarified that their goal isn’t to recreate exact genetic replicas of extinct animals, but rather functional proxies. By functional proxy, Colossal means an engineered organism, typically derived from a close living relative, that reproduces the essential functions, phenotypes, or ecological effects of an extinct species. Practically, this means prioritising traits and behaviours that matter for ecosystem dynamics or scientific insight, while accepting that many genetic and developmental details will differ.
Framing the project this way has major strengths because it accepts scientific limits. Recreating a completely identical genome is technically and scientifically impossible. So, the goal is creating proxy animals that can perform similar ecological roles or provide valuable scientific understanding. The dire‑wolf example illustrates both promise and restraint. The three cubs that outwardly resemble dire wolves are a technical triumph in directed phenotype engineering, showing how targeted edits can shift morphology and behaviour. But that same result also underscores why it’s right to be cautious about language.
Now, can we call the dire wolf ‘virtually de‑extinct’? If the term means creating a living organism that restores key traits and ecological functions of an extinct species, then yes. Under carefully specified criteria, a functional proxy that reliably expresses target traits and can be ethically managed could qualify as virtual de‑extinction. But if it requires full genetic identity, historical continuity, and complete ecological equivalence, then no. The better approach is transparent criteria: clearly define important traits and functions, prove proxies reproduce them across generations, confirm health and welfare, and test ecological impacts. When those boxes are checked, the label ‘virtual de‑extinction’ is defensible as a pragmatic, conservation‑oriented category. If Colossal maintains this proxy-focused, scientifically transparent, conservation-based approach, it could make meaningful contributions to biodiversity restoration.
The writer is a student, North South University. Email: [email protected]
