Colossal Biosciences announced on April 7, 2025, that it had achieved what it calls the world’s first de-extinction through the birth of three dire wolves named Romulus, Remus, and Khaleesi. The same company now says it plans to produce a woolly mammoth calf by 2028, a timeline that depends on scaling gene-editing techniques from mice to elephants, a leap that independent scientists have publicly questioned. Whether the dire wolf milestone actually validates the technical path to a mammoth birth is the central tension behind both claims.
Dire wolf births and the mammoth clock
Colossal’s dire wolf announcement described a process built on ancient-genome sequencing and CRISPR editing of modern canid cells. In its Business Wire release, the company characterized the births as proof that extinct species can be functionally restored through genetic engineering and framed the event as the first confirmed de-extinction birth in history. The three pups, presented as viable, behaviorally normal animals, are meant to demonstrate that a reconstructed genome can be instantiated in living bodies, not just in cell lines or embryos.
The mammoth project runs on a parallel but far more difficult track. Colossal’s stated goal is to produce a cold-adapted elephant–mammoth hybrid calf by 2028. The company’s public evidence for mammoth-related progress centers on a March 2025 bioRxiv preprint describing multiplex CRISPR edits in mice that produced hair phenotypes resembling woolly mammoth fur. That preprint, which the company has promoted as a “woolly mouse” experiment, details specific gene targets and experimental results behind the fur phenotype, including edits to keratin and hair-growth regulators. Colossal presents this as a modular toolkit: identify mammoth traits, map them to genes, and install those edits into a living organism.
The gap between editing hair genes in a mouse and gestating a viable mammoth–elephant chimera in a surrogate elephant is enormous. Mice have short gestation periods, well-understood reproductive biology, and decades of established protocols for embryo manipulation and germline editing. Asian elephants, the closest living relatives to the woolly mammoth, have gestation periods of roughly 22 months, produce oocytes that are difficult to harvest, and have never been successfully cloned. Every step from nuclear transfer to embryo implantation to surrogate health monitoring remains unproven at scale in elephants. Translating a gene-editing workflow from a standard laboratory model to a long-lived, endangered megafaunal species is not a simple matter of scaling up lab equipment.
Independent scientists challenge the woolly mouse as mammoth evidence
A Nature News analysis of the woolly mouse preprint quoted independent researchers who expressed skepticism that the mouse data represent meaningful progress toward a living mammoth. In that Nature coverage, scientists noted that producing altered hair in a rodent does not demonstrate the ability to create viable mammoth–elephant hybrid embryos, sustain them through blastocyst development, or carry them to term in a surrogate. They emphasized that while the preprint shows successful gene editing for specific phenotypic traits, it does not address the reproductive and developmental biology challenges that define the actual bottleneck.
Critics also pointed out that the woolly mouse work focuses on visible, comparatively simple traits-hair length, density, and curvature-rather than complex mammalian adaptations such as fat deposition, thermoregulation, or immune compatibility between species. Editing a handful of loci in mice does not guarantee that dozens of coordinated edits in elephants will avoid deleterious interactions. The more edits required to approximate a mammoth-like phenotype, the higher the risk of off-target effects, developmental failures, or unanticipated health problems later in life.
A separate bioRxiv preprint on dire wolf ancestry and evolution, posted with DOI 10.1101/2025.04.09.647074, provides genomic context for the dire wolf project but also raises questions about hybrid-cell viability. The dire wolf lineage diverged significantly from modern wolves, and the preprint’s methods section describes the genomic reconstruction process without disclosing full raw sequence reads or edit validation tables in public archives. Without those data, it is difficult for outside researchers to evaluate how complete the reconstructed genome is, or how heavily it relies on modern canid scaffolds to fill gaps in ancient DNA.
If hybrid-cell viability problems appear during the dire wolf work’s scaling phase, similar or worse issues could emerge when attempting mammoth–elephant chimeras at the blastocyst stage, where species divergence is even greater. Cellular incompatibilities might manifest as early embryonic arrest, placental abnormalities, or immune rejection in the uterus. These risks are not visible from photographs of healthy dire wolf pups or from fur phenotypes in mice, yet they are precisely the problems that must be solved for a 2028 mammoth birth to be plausible.
An EMBO Reports commentary titled “How to clone a Dire Wolf?” described the claimed workflow of sequencing ancient DNA, applying CRISPR edits, and performing embryo work. That peer-reviewed piece traced which scientific artifacts exist in the public record and which remain undisclosed, noting that many key steps are documented only through conference talks or company statements rather than detailed methods papers. Notably, no primary data on mammoth oocyte sourcing, nuclear transfer success rates, or surrogate health protocols have been published outside Colossal’s own descriptions, leaving a large gap between the narrative of steady progress and the evidence available for independent scrutiny.
Missing data and the 2028 deadline’s weak points
Three specific data gaps stand between Colossal’s public evidence and its 2028 mammoth target. First, the company has not deposited full raw sequence reads or comprehensive edit tables from the dire wolf animals in public genomic archives. Without independent verification of the edits, the scientific community cannot confirm how closely the engineered animals match the extinct dire wolf genome or assess off-target effects that could complicate future projects. Peer review of such datasets would also help clarify whether the dire wolves are best described as genomic reconstructions, high-percentage hybrids, or something in between.
Second, no peer-reviewed or preprint data exist on elephant oocyte harvesting, nuclear transfer efficiency, or embryo culture protocols for mammoth–elephant hybrids. These are not minor technical details. They represent the core biological steps that must succeed for any large-mammal de-extinction attempt to produce a live birth. The woolly mouse preprint demonstrates gene-editing capability in a model organism, but it does not address any of these reproductive hurdles. Even incremental progress-such as reporting on in vitro embryo development from edited elephant cells-would provide a more concrete foundation for the 2028 goal than trait-level work in rodents.
Third, surrogate elephant health records and gestational monitoring protocols remain entirely internal to the company. Elephant pregnancies carry significant veterinary risks even under normal circumstances, including stillbirths, dystocia, and long postpartum recovery periods. Layering experimental embryos, extensive anesthesia, and repeated reproductive interventions onto endangered animals raises ethical and regulatory questions that cannot be evaluated without transparent reporting. Data on hormone monitoring, ultrasound criteria, and thresholds for terminating high-risk pregnancies would be essential for assessing whether the project can proceed without unacceptable welfare trade-offs.
These missing datasets collectively weaken the 2028 timeline. With only three years between the dire wolf announcement and the mammoth deadline, Colossal would need to solve multiple unsolved problems in parallel: perfecting multiplex editing in elephant cells, achieving reliable nuclear transfer, establishing embryo culture systems that support development to implantation-ready stages, and managing high-risk pregnancies in large, scarce surrogates. Each of these milestones typically requires years of optimization even in well-studied livestock species.
Supporters of the project argue that ambitious deadlines can catalyze innovation and attract funding to neglected areas of reproductive biology. They note that early-stage secrecy is common in biotech and that companies often delay releasing detailed methods until intellectual property is secured. Yet the de-extinction framing, and the use of endangered elephants as experimental surrogates, place Colossal’s work closer to a public conservation intervention than a conventional private R&D program. That framing strengthens calls from independent scientists for preregistered protocols, open data on animal welfare outcomes, and clear criteria for success or termination.
For now, the dire wolf pups and the woolly mouse stand as powerful symbols rather than definitive proof that a mammoth calf is imminent. They show that ancient genomes can guide modern gene editing and that some mammoth traits can be recreated in living organisms. They do not, by themselves, demonstrate that a cross-species pregnancy between an endangered elephant and a heavily edited embryo can be carried safely to term. Until Colossal fills in the missing data on genome fidelity, reproductive techniques, and surrogate health, its 2028 mammoth deadline will remain less a forecast than a statement of intent-and one that the wider scientific community is not yet ready to endorse.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
