Forecast
Complete artificial gestation from embryo to birth is demonstrated by 2126.
Last revised Sep 11, 2026.
Current estimate
probability this has happened by each date
A live human birth follows development entirely outside a human uterus, from fertilisation through implantation, placentation and gestation, under documented medical oversight. Use of any uterine tissue as a substrate does not disqualify the result; a pregnancy established in a body at any stage does.
Resolution criteria
- 2126
- A live human birth follows development entirely outside a human uterus, from fertilisation through implantation, placentation and gestation, under documented medical oversight. Use of any uterine tissue as a substrate does not disqualify the result; a pregnancy established in a body at any stage does.
Probability history
| Revised | State | p(2035) | p(2050) | p(2100) |
|---|---|---|---|---|
hf-forecast-0025 | researched | — | — | — |
hf-forecast-0006 | initial prior | — | — | — |
Strongest evidence for
- The two ends genuinely abut rather than being far apart: one stops at E5.5 and the other starts there, in the same species (hf-evidence-0047).
- Real implantation with genuine trophoblast invasion has been achieved outside an animal at over 90 percent efficiency (hf-evidence-0047).
- The regulatory ceiling is loosening. ISSCR replaced the flat 14-day prohibition with case-by-case review, and two national bodies have recommended 28 days (hf-evidence-0050).
- Stem-cell embryo models reach roughly day 21 and bypass implantation altogether, which is a second and structurally different route (hf-evidence-0050).
Strongest evidence against
- Implantation-side culture stops at E5.5 after 96 hours, the exact stage where organogenesis-side culture begins, and no system has carried one conceptus across that boundary in any species (hf-evidence-0047, hf-evidence-0048).
- The earliest extra-uterine support attempted in any species is roughly 22 to 25 weeks human-equivalent, leaving about twenty weeks in which nothing has been tried (hf-evidence-0048).
- That lower bound is mechanical rather than a matter of refinement: no umbilical vessel is large enough to cannulate earlier (hf-evidence-0051).
- The middle requires the histiotrophic to hemotrophic transition, remodelling of the diffusion barrier from about 50 to 5 micrometres, and active maternal immune tolerance. None has an analogue in growing cells longer or in cannulating and oxygenating (hf-evidence-0051).
- Only humans and non-human primates share a hemochorial placenta, so the mouse and sheep work does not transfer (hf-evidence-0051).
- No jurisdiction permits culturing a real human embryo past 14 days or to viability (hf-evidence-0050).
Present uncertainties
- Whether placentation must be reproduced at all, or whether a stem-cell route that never implants can reach viability, which would make most of the evidence above irrelevant.
Next discriminating observation
Any single continuous conceptus taken through implantation and into organogenesis outside a uterus, in any species; and any government acting on the 28-day recommendations.
Sources
- An ex vivo uterine system captures implantation, embryogenesis, and trophoblast invasion via maternal-embryonic signaling
Combining real mouse embryos with real uterine tissue in an air-liquid-interface device, more than 90 percent of embryos attached rigidly to the endometrium at 24 hours and about 85 percent remained attached at 48 hours at stage E5.25, with genuine trophoblast invasion breaking the epithelial barrier by 36 hours and eliminating uterine luminal epithelium by 48. Development then stopped: the authors state that embryos did not grow beyond the E5.5 stage even after 96 hours.
Limit: Mouse, not human, and mouse placentation differs structurally from human. It uses real uterine tissue rather than an engineered substitute, so it does not show that placentation can be engineered at all, only that it can be sustained briefly outside the animal.
Fetched and confirmed directly by the master agent.
- The uncovered span between embryo culture and extra-uterine support
Implantation-side culture stops at E5.5, which is precisely where post-implantation organogenesis culture begins, and no published system has carried one continuous conceptus across that boundary in any species. On the other side, the earliest gestational age any extra-uterine support system has been attempted, across both major groups and any species, is roughly 22 to 25 weeks human-equivalent. Between the human legal culture limit of 14 days and that entry point lies an uncovered span of roughly twenty weeks in which nothing has been attempted.
Limit: An absence of published attempts is a negative search result rather than proof that none exists, and the span is derived by joining two literatures rather than stated in one source.
Fetched and confirmed directly by the master agent.
- Artificial placenta support of extremely preterm ovine fetuses at the border of viability for up to 336 hours
Six lambs at 93 to 94 days gestation, about 600 grams, were supported for a mean of 250 plus or minus 72 hours with a maximum of 336. Only one of six completed the protocol; three were euthanised after bleeding from the bladder and two after umbilical artery vasoconstriction. Against in-utero controls, lung weight was 37.0 versus 45.7 grams, liver 53.0 versus 75.25, kidney 9.5 versus 12.9, and humerus growth 2.9 versus 4.8 centimetres, all statistically significant. IGF-1, a principal fetal growth hormone, was 17 plus or minus 7 micrograms per litre against 140 plus or minus 18 in controls.
Limit: One study, six animals, one species and one device configuration. It does not establish that every artificial placenta approach impairs growth, and first-generation device deficits are the kind of problem engineering usually reduces.
Fetched and confirmed directly by the master agent.
- The 14-day limit, and what has and has not moved
No jurisdiction permits culturing a real human embryo beyond 14 days, and none permits culture to viability. The ISSCR's 2021 revision removed the flat prohibition in favour of case-by-case review, and the UK HFEA and the Netherlands Health Council have each recommended extending the limit to 28 days, with no government having acted. Stem-cell-derived embryo models, which are legally distinct objects, have been cultured to roughly day 21.
Limit: Because no real human embryo has ever legally been cultured past 14 days, whether one could survive and develop coherently past that point is untested rather than known in either direction. The day 21 figure belongs to a synthetic model, which is a materially different object.
Retrieved by a delegated source worker and not independently re-fetched.
- Why the middle is not the same engineering problem as either end
Bridging the gap requires the histiotrophic to hemotrophic transition, continuous remodelling of the maternal-fetal diffusion barrier from roughly 50 to 5 micrometres, and active maternal immune tolerance of invading trophoblast. None of these has an analogue in growing cells for longer, which is the early-end problem, or in cannulating and oxygenating, which is the late-end problem. Only humans and non-human primates share an identical hemochorial placenta, so sheep and mouse results do not transfer directly. The late-end lower bound is itself mechanical: no umbilical vessel is large enough to cannulate before roughly 22 to 24 weeks in any species yet attempted.
Limit: No source claims placentation is theoretically irreproducible. A century admits approaches nobody currently proposes, such as lab-grown decidual and immune tissue rather than a purely mechanical circuit.
Retrieved by a delegated source worker and not independently re-fetched.
- FDA advisory review and commercial investment in extra-uterine support
The FDA's Pediatric Advisory Committee reviewed artificial womb technology on 19 September 2023 and named prerequisites for a human trial under the paediatric research regulations, including a proposed initial enrolment restricted to very high risk neonates. Vitara Biomedical, spun out of the group behind the original lamb work, raised a 50 million dollar Series B in November 2024. FDA-presented piglet data showed heart failure within hours to days, materially worse than the lamb results, and FDA staff concluded the technology is likely to be more difficult at younger gestations and in smaller patients.
Limit: An advisory committee review is not an approval, no human trial has been authorised, and the piglet results show the outcome depends on species and size rather than on the concept alone.
Retrieved by a delegated source worker and not independently re-fetched.