Lilly and Novo say GLP-1 drugs slow biological aging. The ruler they used has never been validated, and the independent test has not started enrolling.
In my line of work there is a name for a measurement that the vendor chose, ran and reported by itself. We call it a claim. We do not call it a finding, and we do not rebuild our architecture around it. So when I read that the two biggest GLP-1 drugmakers say their products slow biological aging, my first question was not whether it was true. My first question was who built the ruler.
Eli Lilly and Novo Nordisk both presented at an aging research conference in Boston. Novo reported blood draws from 10,052 people, half on semaglutide and half on placebo, and read them with proteomic clocks, which estimate age and mortality risk from blood proteins. The overall difference was roughly two to three years against placebo, and as much as four years on a clock built from heart proteins. Lilly, using epigenetic clocks that count accumulated DNA changes, said molecular time for the most part moved more slowly on tirzepatide, and called the effect a significant reduction rather than a reversal. [1] MIT Technology Review’s account does not say either set of results has been published in a peer-reviewed journal, and it does not link to either company’s data. [1]
What a clock actually measures
An aging clock is a model. Someone trains it on blood proteins or DNA methylation to predict age or risk, and the output is a number that feels like a measurement. The Biomarkers of Aging Consortium, writing in Cell in 2023, is the field’s most visible attempt at a shared validation framework, and it is blunt about where things stand. Its authors state that there are not yet any formally validated or likely surrogate endpoint biomarkers of aging. [3] Showing that a clock responds to an intervention is only the first step. The reduction in biological age matters only if it is associated with a real clinical improvement, such as less frailty, less disease or lower risk of death. [3]
The consortium lists other open problems that read like a threat model for any metric. A near-perfect age predictor can lose its link to mortality, which the authors call the paradox of biomarkers. Many clocks are trained to predict chronological age, which may be the wrong target. Blood-based markers may mostly reflect the blood system and miss slower tissues. Nobody agrees on accuracy metrics or on clinical endpoints for aging trials. [3] There is also the odometer and speedometer problem. Biological age says how much aging has accumulated, while pace of aging says how fast it is happening now, and the field has not settled which a drug claim should rest on. Individual components such as methylation probes can be noisy, and normal fluctuations in these markers are poorly understood, which makes real change hard to separate from signal. [3] The authors also describe a loop in which geroprotectors are used to benchmark biomarkers while biomarkers are used to test candidate geroprotectors. [3] They present that as normal practice borrowed from animal research. I read it as a reason to want someone outside the loop.
Most of the human evidence for clocks moving at all comes from small or exploratory studies, many of them post hoc analyses. [3] Novo’s 10,052-person dataset is far larger than those, which is exactly why I want to see it. Size without independent inspection is still a vendor dataset.
The fair case for the drugmakers
I am not arguing that these drugs do nothing. That would be as lazy as the press release. Steve Horvath, the UCLA scientist credited with inventing aging clocks, called two years a pretty strong effect and said the data could support classifying GLP-1 drugs as geroprotectors, though he is not yet convinced they help people who are not obese. [1] Vadim Gladyshev, the Harvard biologist who helped Novo with the molecular measurements, thinks the drugs act as anti-aging drugs in unhealthy people but says no one knows about healthy ones. [1] The same reporting notes real-world benefits for kidney function, blood pressure and overall mortality. [1] A drug can be good medicine for a metabolically sick population and still have no proven anti-aging effect in a healthy fifty-year-old.
That distinction is the whole argument, and it is the one the marketing will flatten. Novo’s own researcher declined to say whether a reasonable person should take semaglutide, noting he is not a physician. [1] Meanwhile Alex Zhavoronkov of Insilico Medicine told MIT Technology Review that he takes both drugs, including microdosing despite not being overweight, and framed the question as whether the first mass-market anti-aging remedy already exists. [1] People are acting on the number before anyone has shown it means what they think it means.
The side of the ledger nobody advertises
Every security assessment has a cost column, and so does this one. MIT Technology Review’s separate piece on GLP-1 side effects lists nausea, vomiting, diarrhea and constipation as the most common, and links the drugs to pancreatitis and gallstone disease. Dozens of people are suing both companies over sudden vision loss, which the companies dispute and which the research treats as mixed. Hair loss has been associated with use in a large health-records study, though it is not thought to be common. [2] Horvath himself flagged muscle loss. [1] The article is explicit that how common each effect is remains unknown, and that data on children is thin. [2] It also points out that the biological-age findings come from the drugmakers, which is a caveat. [2]
Uptake is not waiting for answers. A Gallup poll cited there found 11 percent of Americans taking a GLP-1 for weight loss, up from 3 percent in 2024. [2] When a number that is not yet validated meets a population that large, the number starts doing work in the world whether the science is ready or not.
What a real test looks like
The most telling detail is what the one serious independent trial chose to measure. ARPA-H is funding a study, led by the Barshop Institute at UT Health San Antonio, called VITAL-H. It will enroll 726 adults in their 60s and compare rapamycin, dapagliflozin and semaglutide against placebo across four arms, with participants taking one pill a day and wearing a ring that tracks activity, heart health and sleep. [4] The primary goal is not a clock. It is to validate intrinsic capacity, a composite of physical and mental function tested across five domains, as a meaningful, regulatory-grade endpoint the FDA could accept. [4] Enrollment is expected to begin in 2027. [4] Stanford leads the intrinsic capacity validation work, and the consensus framework behind it sets out a stepwise validation plan toward using intrinsic capacity as an endpoint in healthspan trials. [4][5] That is what slow, deliberate validation looks like, and it is the opposite of a conference slide.
Read that as a signal. The people designing the trial that is meant to define a regulatory pathway for anti-aging drugs went to function, not to a blood-based age estimate. ARPA-H’s PROSPR program, which sits behind the study, aims to assess age-associated health outcomes in three years instead of decades, and has an expert consensus document on intrinsic capacity intended to give the FDA recommendations. [5] There is a small mismatch in the public record worth noting. MIT Technology Review describes the funded study as a semaglutide test in healthy people over 60 that will look at cognition, mobility and sensory acuity, and says answers could take one to two years. [1] The university describes three drugs and says enrollment starts next year. [4] Neither account is wrong about its own piece, but the timeline is not as short as the headline energy suggests.
Who holds the ruler
I have spent my career watching what happens when the party that benefits from a metric also controls how it is measured. Detection rates, mean time to respond, coverage percentages. The number improves, the dashboard goes green and the underlying reality is anyone’s guess. Biology is not immune to that dynamic just because the instruments are expensive.
So here is what I would want before anyone builds a business, a clinic menu or a personal protocol on a GLP-1 aging claim. The clock should be named and fixed in advance, so nobody can choose the flattering one afterward. The raw data should be deposited where independent groups can reanalyze it. Adverse events should be reported in the same breath as the biological-age result. The finding should replicate in people who are not already sick. The good news is that one such test exists, because semaglutide is one of the three drugs in the independent trial. Someone with no stake in the answer will eventually check this claim against function, not against a model. And the endpoint should be one regulators can accept, a status the consortium says no aging biomarker has yet earned. [3]
Until then, a drugmaker saying its drug makes you two to three years younger is a vendor claim, however large the sample and however respected the scientists who helped. It may well turn out to be right. The drugs may be the first real geroprotectors. If so, the independent evidence will arrive, and it will survive scrutiny because it was gathered by people with no stake in the answer. That is the bar I would set for any product that asks for my trust, and it should hold for a pill that asks for my decades.
About the Author
Len Noe is a Solutions Architect at BeyondTrust, Transhuman, Podcaster, International Cyber Security Speaker, Author, Technical Evangelist, and Biohacker with 11 implanted microchips. A former blackhat and ex-1% motorcycle club member with more than 30 years in technology, he has presented in over 70 countries and is featured in the documentary I Am Machine, which premiered at DEF CON 2025.
References
- Regalado, Antonio. “Weight-loss drugs show signs of slowing biological aging, say drugmakers.” MIT Technology Review, 6 October 2026. technologyreview.com
- MIT Technology Review. “We’re still figuring out the side effects of GLP-1 weight-loss drugs.” MIT Technology Review, 9 October 2026 (byline not captured). technologyreview.com
- Moqri, Mahdi; Gladyshev, Vadim N.; and the Biomarkers of Aging Consortium. “Biomarkers of aging for the identification and evaluation of longevity interventions.” Cell 186(18):3758-3775, 31 August 2023 (online). pmc.ncbi.nlm.nih.gov
- UT Health San Antonio. “UT San Antonio to lead $38 million national trial testing drugs to extend healthspan.” UT Health San Antonio news release, 4 March 2026. news.uthscsa.edu
- ARPA-H. “PROSPR: Proactive Solutions for Prolonging Resilience.” Advanced Research Projects Agency for Health, program page (accessed 21 October 2026). arpa-h.gov