Pauling.AI’s Javier Tordable on why aging biology outpaces the business model and how AI can reset the cost curve.
We keep talking about longevity like it’s a moonshot science problem. It isn’t. The science is moving fast; what’s lagging is the economic machinery needed to turn that science into approved interventions at scale. Until we fix that, the most exciting biology in the world will stay trapped in papers, not prescriptions.
The fundamental challenge of longevity research is not scientific but economic. The traditional pharmaceutical model requires clear disease endpoints, regulatory pathways, and billion-dollar market opportunities to justify investment. Longevity research has almost none of that working in its favor. Longevity biotechs target aging itself rather than a specific disease, which creates regulatory uncertainty and makes the blockbuster economics required for the pharma business nearly impossible. That’s why despite enormous public interest and obvious impact to society, institutional capital largely stays away.
Over the past few years, longevity has had a strange kind of success: it’s become intellectually mainstream without becoming economically feasible. The geroscience case for targeting the mechanisms of aging is stronger than ever, biomarkers are improving, and early clinical signals are starting to emerge. But from the outside, the field still looks like a patchwork of workarounds and companies forced to pick narrow disease indications not because that’s the most coherent scientific path, but because it’s the only path the current system reliably funds.
The core problem is that “aging” isn’t a billable, reimbursable label. Regulators want hard endpoints, payers want clear populations, and investors want timelines that fit venture math. That naturally pushes longevity companies toward diseases of aging – Alzheimer’s, fibrosis, frailty, metabolic dysfunction – because those have established trial designs and regulatory precedent. But the moment you translate aging biology into single-disease products, you inherit the slowest part of the whole pipeline: long, expensive trials and binary outcomes.
That’s why the economics are so brutal. It’s not that longevity is unimportant, it’s that the system is structurally optimized for “one drug, one disease,” not “one intervention, many downstream benefits.” Longevity promises the latter, and the current model punishes it. The current system also disincentivizes long term therapies that may have substantial effect only when applied for many years or decades.
This is exactly why tooling matters more in longevity than in almost any other therapeutic area. If the discovery phase is expensive and slow, only the most conventional, highest-certainty bets survive. But if discovery becomes cheap and fast, we can test more hypotheses early, explore more chemistry, and fail quickly, which is the only rational way to navigate biology as complex as aging.
What we’re doing at Pauling.AI, shrinking early discovery from months to weeks and cutting costs by orders of magnitude, benefits all drug development, but it’s particularly critical for longevity. By making the science radically more efficient, screening millions of molecules in days instead of months, for thousands of dollars instead of hundreds of thousands, we change the fundamental economics. Suddenly biotechs can afford to pursue mechanisms and targets that don’t have massive markets behind them.
In practice, lack of an economic model means a huge amount of potentially valuable work never gets funded. Mechanisms that look promising but uncertain – anything involving multi-pathway effects, combination strategies, or interventions that might show broad benefit only over time – become financially radioactive. You either over-promise a blockbuster indication you don’t truly believe in, or you die in a valley of translational cost before you ever reach a definitive result. The field doesn’t fail because the biology is wrong. It fails because the early-stage economics are misaligned with what the biology requires.
Here’s the gap we’re filling: computational chemistry today still involves enormous amounts of manual work and human intuition. A researcher needs to know which tools to use, how to set up simulations, how to interpret results, and how to iterate based on what they find. It’s expert-driven, time-consuming, and expensive. What we’ve done is leverage LLMs and agentic AI to automate that entire workflow.
To be clear, this doesn’t eliminate the need for wet-lab biology, smart clinical strategy, or human judgment. It changes what humans spend their time on. Instead of researchers burning weeks on tool setup, parameter sweeps, failed runs, and repetitive iteration, they get to spend that time asking better questions: which targets are actually worth pursuing, which hypotheses deserve validation, which models predict the real world. The point isn’t “AI replaces scientists.” The point is that AI can finally remove a huge amount of friction from a process that has been artificially expensive for decades.
Our agent doesn’t just run calculations, it makes decisions about what to do next based on results, corrects its course when something doesn’t work, and orchestrates dozens of specialized tools without human intervention. We’re talking about tasks that used to take a computational chemist weeks of hands-on work, now running autonomously in hours with broader exploration and fewer bottlenecks because the system can explore far more possibilities than any human could manually.
For the first time at scale, you don’t need a team of PhDs in computational chemistry to do world-class drug discovery. A biologist with a target can get a ranked list of optimized drug candidates without touching a single piece of chemistry software. That democratization, that radical reduction in both expertise required and resources consumed, is what makes previously impossible research suddenly viable. It’s not an incremental improvement, it’s a complete reset of what’s economically and practically feasible in early drug discovery.
And as biomarkers mature, this becomes even more important. The faster we can generate and refine candidates upstream, the more shots we can take on downstream strategies that rely on surrogate endpoints – epigenetic clocks, proteomic aging signatures, functional measures – instead of waiting years for hard outcomes. The field’s future depends on making early discovery cheap enough that we can run more experiments, not fewer, while regulators and payers catch up.
We believe that’s the way to unlock orders of magnitude more longevity interventions: make discovery so cheap and fast that it no longer requires pharma-scale returns to justify the work. Lowering discovery costs doesn’t just help one company, it changes what the entire ecosystem can afford to attempt. When the price of exploration drops, the field becomes less dependent on blockbuster logic and more able to pursue interventions that produce real healthspan gains even if they don’t map neatly onto a single disease label.
The longevity field won’t win by arguing harder that aging matters. It will win by making it economically irrational not to pursue aging biology, because the cost of testing new ideas becomes low enough that the upside is always worth the attempt. The first domino is discovery.
About Javier Tordable
Javier is the Founder and CEO of Pauling.AI, an automated agentic system for drug discovery. Previously, during 16 years at Google, Javier led technical work and managed teams in Search, Ads, Supply Chain and Cloud.
