The American startup Nucleus Genomics, funded by Founders Fund, linked to investor Peter Thiel, offers a service that allows comparing in vitro fertilization (IVF) embryos according to estimates of intelligence, height and genetic predisposition to disease. The company claims its models can calculate significant differences between embryos even before a pregnancy begins.
The Nucleus Embryo platform was launched in June 2025. In September 2026, the company presented the Vitruvian models, which use genetic analysis and machine learning to calculate scores associated with hereditary traits.
The service, publicized with the slogan “Have your best baby” (“Have your best baby”), allows prospective parents to consult the results of different embryos and use that information in deciding which one will be transferred to the uterus.
The technology does not modify DNA or use CRISPR gene editing. It analyzes existing embryos produced during fertility treatment.
Estimated 14.3-point IQ difference does not represent a proven gain
In a technical document published in September, Nucleus presented a model that analyzes approximately 7 million genetic variants and estimates differences between embryos from the same family.
According to the company's calculations, a set of ten embryos would show an expected average difference of 14.3 IQ points between the one with the highest predicted genetic score and the one with the lowest score. For five embryos, the estimated difference drops to 10.8 points.
There is, however, an important distinction: the 14.3 points represent the estimated distance between the two extremes, not the expected increase in intelligence when selecting the highest-ranked embryo.
According to Nucleus's own methodology, the projected average advantage of the embryo with the highest score compared with an embryo with an average score would be approximately 7.2 points in the ten-embryo scenario.
These results derive from mathematical models, not from monitoring children born after this type of selection. The company evaluated its predictions using data from approximately 40,000 siblings of European ancestry from the UK Biobank. Its intelligence model explained about 19% of the cognitive variation considered in the evaluation among family members.
The accuracy of predictions remains limited by genetic factors not captured by the models and by the influence of environment on cognitive development.
Organizations question the service's accuracy and advertising
On September 11, 2026, the National Advertising Division (NAD), a U.S. advertising self-regulatory body, recommended that Nucleus modify or discontinue certain claims about the Nucleus IVF+ service.
The organization concluded that the evidence presented did not demonstrate the ability to accurately identify which embryo would have the highest IQ in adulthood. It also questioned advertisements that suggested reliable predictions of height, eye color, hair color and individual differences in risk for certain diseases.
Nucleus disagreed with the conclusions and announced that it would appeal the decision.
The American Society for Reproductive Medicine (ASRM) also adopted a restrictive position. In an opinion issued in December 2025, the organization stated that polygenic screening of embryos still lacks sufficient evidence of efficacy and safety to be offered as a clinical service.
The opinion emphasizes that selection for nonmedical traits, such as intelligence and height, is outside the scope of reproductive medicine.
Despite scientific reservations and advertising challenges, Nucleus already markets tools that present genetic predictions to prospective parents. The service turns statistical scores into information used in reproductive decisions, while its ability to predict the actual outcomes of those choices remains without sufficient clinical proof.
