Two new genomics startups have released consumer services claiming to allow parents to screen and select embryos based on polygenic scores for IQ. Both are already subject to considerable criticism along scientific, bioethical and political lines. The two startups are also both connected to the emergence of “educational genomics” as a domain of research over recent years. Although educational genomics and the “geneticization of education” it represents has always been ethically and scientifically contested, is it now slipping into an even more controversial form as commercialized services for prospective parents seeking to invest in their children’s educational success?

Superbaby startups

Nucleus Genomics announced its embryo screening and selection service in June 2025, with a press release calling it “the first genetic optimization software that lets parents see and understand a complete genetic profile to select an embryo.” It was followed in late July by Herasight, whose founders announced they were “building the highest quality genetic tests for families” with an “Embryo Tool” for “complex disease prediction” and prediction of “cognitive traits” such as intelligence.

When Nucleus first announced it was building a genetic IQ testing platform a year beforehand, it was criticized as “snake oil” by academic scientists who questioned its scientific basis and credibility. That hasn’t stopped it getting significant media attention and funding from investors including Peter Thiel and Alexis Ohanian, the co-founder of Reddit.

Criticism of its latest embryo testing product launch, however, has been intense. The bioethicists Arthur Caplan and James Tabery argue Nucleus is misleading hopeful parents-to-be with the new embryo screening service, particularly its promises about “genetic optimization.”

They liken Nucleus to Theranos, the fraudulent blood testing startup founded by Elizabeth Holmes, arguing that “like Holmes’ Theranos, … Nucleus Embryo starts from existing technology, and uses that reliable foundation to then leap into the realm of fantastic claims that may entice venture capitalists and wealthy but naive customers but don’t hold up to scrutiny when you start seriously poking around.”

Herasight’s founders have sought to distance themselves from Nucleus in documentation supporting their new embryo screening service, ultimately accusing Nucleus—and other genomics startups—of scientific malpractice (specifically of inflating their results). Clearly these startups are in a competitive race for investor funding, scientific reputation and potential customers of “superbabies,” whether the technology really works as marketed or not.

Both have also been viewed critically as “eugenics projects” that aim to “repackage the highly contentious topic of selective human breeding.” In a thread on Bluesky, the statistical geneticist Sasha Gusev pointed out that several members of the Herasight team subscribe to pro-eugenic and race science views. Population geneticist Graham Coop noted that one Herasight founder was excited by Elon Musk describing their product as “cool” on twitter, while geneticist Kevin Bird highlighted how they responded favourably to a tweet praising their “IQ predictor” from a renowned scientific racist.

While the idea of genetic IQ tests is certainly not new, and has previously been characterized as “eugenics 2.0,” the current political and technological context is clearly more conducive to the sales pitch of these startups. Tech firms and investors share perspectives with race scientists and pro-eugenicists, whose views are amplified in the current US political mainstream. Formerly secretive efforts to build “superbaby factories” for high-IQ kids, as the anti-fascist organization Hope Not Hate has characterized them, are now becoming publicly marketed products pitched hard to aspiring parents.

Polygenic scoring and screening

Researchers who have worked on educational genomics studies over the last decade or so have frequently sought to differentiate their work from consumer genomics companies. They have sought to head off bioethical objections by pre-emptively rejecting eugenicist and racist interpretations of their findings. The bioethical management of educational genomics has included Q&A documents for media and public consumption, responsible research and communication frameworks, as well as calls for pre-emptive regulation of genetic embryo screening for cognitive abilities.

However, in at least two ways, educational genomics and consumer embryo screening are closely related. The first connection is to do with shared methodology; the second about shared personnel and expertise.

Methodologically, both educational genomics studies of the genetic bases of educational outcomes and embryo screening for predicted IQ depend on the technique of polygenic scoring. Polygenic scoring is a complex statistical process that involves calculating thousands of minute molecular differences related to a behavioural or cognitive trait or a social outcome into a single summed score for an individual.

In educational genomics, polygenic scores for educational attainment (years of school and college) are the focus for a number of very large sample studies. The general argument of such research is that polygenic scores can be used to estimate outcomes like attainment, as well as other outcomes such as exam scores, which it is often claimed could be useful to inform educational practitioners and policymakers. These studies have become highly influential, catalyzing a rapid international rise in research exploring the genetic bases of educationally relevant behaviours and outcomes.

There are also many studies calculating polygenic scores for IQ. In some cases, educational attainment is viewed as a proxy for cognitive ability and IQ, with intelligence sometimes even correlated with “brain size” and “cranial volume,” as a recent paper indicates:

[T]he association between brain size and intelligence appears to be almost entirely due to genetics. …  Educational attainment (EA), defined as the number of full-time years of education an individual receives, is a useful proxy trait for cognitive ability…. [T]his is one of the first studies investigating the associations between the genes for education and brain anatomy using a polygenic scoring approach… [which] aimed to assess the association between the genes related to education (as a proxy for general cognitive ability) and the morphometry of specific cortical regions … [and] assessed whether the established association between an EA-PGS and IQ scores is mediated by identified brain structures.

Polygenic scores for educational attainment, then, are often taken as proxy measures of intelligence, with genetic predictors of educational outcomes understood to correlate with IQ scores, with both sharing the same “neural substrates.”

These claims persist despite methodological critiques showing that DNA has almost no direct effect on educational outcomes and that the direct genetic contribution to intelligence is minimal once other personal, cultural and social factors are fully taken into account. Nonetheless, it remains clear that the results of polygenic educational genomics studies and genetic IQ measurement are methodologically symmetrical and often overlapping, with education attainment even treated as a convenient proxy of intelligence.

Commercializing educational genomics

The methodological symmetries between consumer genetic IQ testing and the polygenic prediction of educational outcomes are why some scientists conducting educational genomics studies have looked to consumer genomics technologies as the basis for “predicting educational achievement from DNA.”

It is notable, for example, that one of the most well-known behavioural geneticists involved in educational genomics research, Robert Plomin, has repeatedly claimed consumer polygenic testing of babies could generate data for use in education, and has a role as a listed “collaborator” on the Nucleus Genomics website. In a 2021 paper, Plomin noted the potential use of the non-profit service impute.me, which computed polygenic scores for intelligence for its users; the service was subsequently acquired and commercialized by Nucleus as the basis for its polygenic IQ predictor.

The connection between embryo testing startups and educational genomics, then, is more than just methodological. It also involves the mutually-beneficial exchange of expertise and technologies, with scientists advising startups and startups building new testing technologies for scientists to include in their research designs.

Likewise, one of Herasight’s founders is a research scientist with a position on the steering group of the Social Science Genetic Association Consortium, the group responsible for many of the most influential large-sample studies of the genetic bases of educational attainment. Alexander Young was one of the lead authors of the SSGAC’s 2022 study of “polygenic prediction of educational attainment” based on the genomic data of 3 million people, before establishing Herasight to enact polygenic prediction on pre-implantation embryos as a commercial venture.

Educational genomics has always depended to a significant extent on consumer genomics companies. The large-sample SSGAC studies were only ever possible due to a data access agreement with 23andme, the genetic ancestry firm that amassed the world’s largest private biobank of personal genetic information.

But Nucleus and Herasight are distinct from 23andme because they propose direct testing of embryos for purposes of intelligence prediction, rather than adult customers volunteering their data. The potential here is for polygenic embryo IQ technologies to become part of the expanding infrastucture of technologies and methods that underpins educational genomics research and claims of its relevance to policy and practice.

There is already growing consensus amongst many scientists involved in educational genomics that very young children could be genetically assessed and screened in order to inform schools’ differentiation strategies. Nucleus and Herasight are now positioned to fulfil such a role.

Indeed, the educational genomics studies published over the last decade have set the stage for such startups by producing claims that educational outcomes are significantly genetic. When that assumption is accepted, then services promising to maximize educational outcomes–whether through selection or differentiation–may appear more feasible and desirable.

The prospect here, then, is of commercial genomics startups becoming increasingly integral to the conduct of educational genomics research, and even to the potential integration of genomic data into educational practice and policy. Ultimately, new embryo screening services appear to anticipate educational genomics becoming a commercialized enterprise, even though scientific evidence that these technologies work as claimed remains seriously lacking. Such services are also fraught with ethical problems.

Bioethical controversies

At a recent event on polygenic scores, education and ethics hosted by the Nuffield Council on Bioethics (in which I participated), it was noted that the use of direct-to-consumer genetic testing in educational contexts “could lead to pressure on schools, teachers and policymakers to respond to genomic information without thoroughly weighing up the evidence and associated ethical concerns.” This could mean schools and teachers being put under pressure regarding children’s genetically-predicted potential, resulting in “genomically-driven entitlement.”

Concerns discussed at the workshop included genetic determinism and labelling, with children potentially facing limited expectations about their academic prospects based on genetic data. Discussion also centred on the explanatory limits of genetic data for the complex factors that contribute to educational outcomes, and on the almost negligible role of direct genetic effects on educational outcomes compared to social and environmental factors like parental income.

Polygenic embryo screening and selection technologies like those promised by Nucleus and Herasight could amplify these bioethical concerns and controversies. Given the costs of polygenic embryo selection services (the Nucleus service is marketed at $5,999), it could exacerbate existing inequalities by enabling wealthy families to pay for the privilege of having smarter children. Nucleus even markets its services as “investing in your family’s future.” Though the scientific validity of the service for genetic optimization remains highly disputed, customers could more simply demand from schools extra attention or resources based on belief in the polygenic predictions of academic achievement they have purchased.

Moreover, these startup services play into a contemporary form of eugenics that emphasizes families policing their own gene pools. As Emily Klancher Merchant has argued, there is growing enthusiasm for “genetically engineered children” among pro-natalist groups and Silicon Valley technology entrepreneurs. Intelligence, Klancher Merchant argues, “has always been at the center of American eugenics.” Eugenics attributes socioeconomic inequality to different degrees of intelligence, which in turn is defined “as a biological quality shaped largely by our DNA.”  

With its marketing talk of “genetic optimization” and a price point attainable only for the wealthiest families, services like that of Nucleus serve well the eugenic desire to improve one’s child’s prospects through genetic selection based on intelligence rating–or what Klancher Merchant terms “breeeding for IQ.” But, as many critics of such startups and the methods underlying their platforms have argued repeatedly, these services remain profoundly misleading since polygenic variation has very little discernible direct effect on outcomes like school achievement.

Nonetheless, the risk remains that children’s cognitive abilities and educational achievements may be increasingly imputed to genetic determinants. Once available, polygenic data could be called up by parents, educators and decision-makers to support choices about the allocation of pedagogic provision and resources. What is a child “genetically entitled” to once they are believed to have hardwired high IQ as determined by an expensive scientific procedure undertaken by a startup?

Although the science does not support polygenic IQ embryo screening and selection, as Klancher Merchant argues, the danger is “widespread adoption of the belief that intelligence is primarily genetic and that it can be altered”:

Eugenics has never worked primarily by producing more intelligent babies. Instead, it has worked by creating the illusion that intelligence is primarily genetic, absolving governments of responsibility for ameliorating social inequality.

Similarly, the real danger of polygenic embryo selection, and its potential integration into commercialized educational genomics programs, is the adoption of a belief that educational outcomes are significantly genetically determined, or could even be genetically edited. This might then put pressure on schools to adjust their expectations, pedagogies and allocation of resources based on flawed genetic predictors of cognitive ability, rather than by considering the complex social and personal factors that shape and reproduce educational inequalities.