Nucleus Genomics claims its Vitruvian model can add 14 IQ points to embryo selection
Nucleus Genomics says its new Vitruvian model can optimize embryo DNA for 14 IQ points, nearly a full standard deviation. The company's own whitepaper tells a smaller story, and the science behind embryo intelligence scoring remains deeply contested.
Kian Sadeghi posted the announcement on X himself. "AI is rapidly getting smarter," he wrote. "Now, humanity can too." His company, Nucleus Genomics, says Vitruvian, its newest set of genetic optimization models, can rank embryos from an IVF cycle and push the intelligence gap between the best and worst options to 14 IQ points, close to a full standard deviation. The models were trained on more than a million people, validated across 40,000-plus siblings, and built on more than 7 million genetic markers, according to the company.
Here's what that headline number actually means once you read past it. Vitruvian doesn't edit anyone's genes. It scores embryos that already exist, side by side, and tells parents which one statistically carries the higher predicted intelligence score. The DNA is fixed before the model ever sees it. What changes is which embryo gets implanted.
And the 14-point figure is the outer edge of the claim, not the typical outcome. According to a detailed read of the underlying whitepaper published by onhealthcare.tech, the expected spread between the highest and lowest scoring embryo in a batch of five is roughly 10.8 points, with the gain against an average embryo in that batch closer to half of that, call it 5 points, not 14. Older peer-reviewed modeling of the same statistical problem, cited in that same analysis, put the realistic expected gain at about 2.5 points when picking the best embryo out of five. Vitruvian's number is a best-case scenario dressed up as the headline.
Nucleus isn't a random biotech afterthought chasing a viral tweet. The company has raised $32 million, including a $14 million Series A in January 2025, backed by Founders Fund, Asylum Ventures, Seven Seven Six, Samsung Next and One Eight Capital. Founders Fund's involvement tracks: it's the firm that has bankrolled contrarian, category-defining bets before, and genetic optimization for embryos fits that mold precisely. In June 2025, Nucleus launched Nucleus Embryo, a software platform used during IVF that screens embryo samples for more than 2,000 traits and conditions, everything from eye color and hair color to predisposition for depression, autism and bipolar disorder. It reached that market through a partnership with Genomic Prediction, a longtime provider of polygenic embryo testing, which Nucleus used to undercut rivals on price.
Nucleus isn't alone in this space, and it isn't even the most cautious player in it. Herasight publicly markets embryo screening that includes intelligence, while Heliospect Genomics has been reported to have pitched IQ screening and now directs embryo-screening inquiries to Herasight. Orchid and Genomic Prediction operate in adjacent polygenic embryo screening, with products centered more publicly on disease risk and broader genomic analysis. Sadeghi's willingness to put a specific number in a tweet, unhedged, is itself the product decision. It's marketing built for virality, and it's working: the post has been circulating across X and product-adjacent newsletters since it went up.
The professional response has been blunt. On Sept. 11, 2026, the National Advertising Division recommended that Nucleus modify advertising that conveyed it could accurately identify which embryo would grow up to have the highest IQ; Nucleus said it would appeal. The American College of Medical Genetics and Genomics has said the practice of using polygenic risk scores for embryo selection has moved too fast with too little evidence behind it. Many IVF clinics still decline to offer these tests at all, worried they set expectations no embryo can actually deliver on. That's not a fringe objection. It's the position of the doctors who would have to sit across from a parent whose "optimized" child doesn't test the way the model predicted.
Frankly, the eugenics comparison isn't a stretch invented by critics looking for a fight. Recent scoping reviews on polygenic embryo screening found that disapproval among the US public tracks most strongly with the fear that this kind of trait selection is eugenics by another name, not disease prevention. Screening for a BRCA mutation and screening for predicted IQ are not the same moral category, even if the software presents them on the same dashboard, and healthcare professionals surveyed in that research argued screening should stay limited to medical conditions, not traits like intelligence or height.
The regulatory map is close to blank, and that gap is doing real work for companies like Nucleus. The United Kingdom bars selecting embryos for certain non-medical characteristics outright, and India has restricted some embryo-testing uses too. Nucleus announced joint ventures in India and the Middle East, including Jordan, in March 2026 anyway, threading whatever regulatory needle each market leaves open. In the US, there's no federal rule stopping any of this. The FDA doesn't approve polygenic embryo scores the way it approves a drug, and the fertility industry has historically regulated itself.
None of this makes Vitruvian fake. The underlying statistical approach, ranking embryos by polygenic score and picking from the pool you already have, is a real and increasingly common practice in IVF clinics that never generated a single tweet. What's new here is the packaging: a single, near-standard-deviation number, pushed out by a well-funded founder with Founders Fund money behind him, at a moment when Silicon Valley's pronatalist wing is already primed to cheer for exactly this kind of claim. The number that will actually show up in a nursery, per Nucleus's own math, is closer to five points than fourteen. That's still not nothing. It's just not the headline.
Also read: DHH admits AI agents struggled on Basecamp even as they built all of Omarchy • How a Secondary Sale Actually Works for Startup Employees Before an IPO • Corgi's Viral Dating Post Turned Into a Referendum on Its Startup Culture
This article is posted in Startup News, check it out for more related stories.