Longevity has never had a science problem so much as a business model problem. No regulator anywhere approves a drug for being old. Zhavoronkov’s keynote was an argument about how you build a company that survives that fact.
There is a particular kind of company that gives the longevity field a bad name. It has a magnificent mission, a charismatic founder, a deck full of hallmarks of ageing, and roughly eighteen months of runway. It does not have a product, because the thing it wants to treat is not an approvable indication, and it does not have revenue, because nobody buys a mission.
Alex Zhavoronkov, Founder, Chief Executive Officer and Chief Business Officer of Insilico Medicine, has spent a decade building the counterexample. His keynote at BIO Asia-Taiwan 2026, How to Build a Sustainable Longevity Company, was not a talk about ageing biology. It was a talk about engineering, in the corporate sense: how you construct an organisation that can pursue an unfundable goal for long enough to reach it.
The framing he brought is deceptively plain. A bold mission is not a company. A company needs an engine, something that generates money, data and credibility faster than the mission consumes them. Most longevity ventures never build one. His argument, drawn from what Insilico has actually done rather than from theory, rests on three pillars.
A mission is not a business model. The longevity field has spent a decade confusing the two, and the graveyard is well populated.
THE PREMISE · WHY MOST LONGEVITY COMPANIES FAIL
PILLAR ONE Productivity you can actually measure
The first pillar is scientific productivity measured against rigorous benchmarks. This sounds like a platitude until you notice how rarely it is done. The AI drug discovery sector has commoditised at the model layer. Everyone has neural networks, everyone has open data, everyone has an architecture slide. What separates the companies is not the claim but the count: how many programmes run, how fast a target becomes a development candidate, and how often the resulting molecule survives contact with biology.
Insilico’s numbers are the argument. The company has nominated 27 preclinical candidates, of which 11 have reached clinical stage, across fibrosis, oncology, immunology, pain and metabolic disease. The Pharma.AI platform that produces them is built in three layers: Biology42 for target discovery, Chemistry42 for molecular design and Medicine42 for clinical insight. Compression of the earliest and most failure-prone stretch of drug development, from target identification to candidate nomination, is the measurable output.
THE DISCIPLINE UNDERNEATH
Zhavoronkov has been consistent that the hard constraint is not chemistry benchmarks, where progress is comparatively easy to measure, but biological validation in animal and human systems. A benchmark culture only works if the benchmarks are the honest ones.
PILLAR TWO A portfolio that funds itself
The second pillar is where the talk earns its title. Insilico runs a portfolio strategy that balances internal progression against licensing, and the licensing is not a consolation prize. It is the engine.
The company has licensed several AI-generated assets over the past three years, including a collaboration with Eli Lilly announced in March valued at up to 2.75 billion dollars, and an earlier deal worth up to 1 billion dollars with 80 million upfront. The proceeds are reinvested into the broader ageing and longevity pipeline. That loop is the whole thesis. The commercially legible work, drugs for indications regulators recognise and partners will pay for, funds the work that no regulator will approve and no generalist investor will underwrite. Zhavoronkov has described this cycle as what makes the ageing research fully sustainable, and the word is doing precise work: not profitable, not visionary, but self-funding.
The mechanism underneath is subtler still. Insilico’s approach has been to wrap ageing-derived targets inside conventional clinical programmes. A target found by asking questions about ageing biology gets developed against a specific disease with a specific endpoint. The longevity work does not have to announce itself to get done. It simply has to be right.
Pharma does not buy an effect on an ageing clock. It buys a primary outcome measure. Build the company that can sell one and fund the other.
PILLAR THREE A platform that compounds
The third pillar is the one with the longest tail: an AI platform that improves with every programme it runs. This is the difference between a tool and an asset. A tool performs the same on its thousandth job as its first. A platform that learns from each programme gets cheaper and better as it goes, which means the marginal cost of the next discovery falls while the value of the accumulated data rises.
It also explains a strategic move that puzzles people who file Insilico under pharma. The same Pharma.AI capability has been pointed at advanced materials, agriculture and veterinary medicine, and through a collaboration with Aramco at carbon capture, hydrogen storage and clean synthetic fuels. Read as diversification, that looks unfocused. Read as platform economics, it is the opposite: every additional domain is another set of programmes teaching the same engine, funded by somebody else.
WHY THIS MATTERS FOR THE ROOM
Compounding is the only honest answer to a mission with a thirty-year horizon. A company that gets better at discovery every year can afford to be patient. A company that does not is simply spending down a fixed endowment of investor optimism.
THE PROOF What the structure has produced
|
INSILICO |
The structure, in numbers |
|
HKEX:3696 |
Listed on the Main Board of the Hong Kong Stock Exchange on 30 December 2025, a capital structure few longevity-focused biotechs have reached |
|
27 / 11 |
Preclinical candidates nominated, of which 11 have reached clinical stage |
|
$2.75bn |
Value of the Eli Lilly licensing collaboration announced in March 2026, with proceeds reinvested into the ageing pipeline |
|
3 layers |
Pharma.AI: Biology42 for targets, Chemistry42 for molecules, Medicine42 for clinical insight |
|
April 2026 |
Insilico convened what it described as the industry’s first longevity board, chaired by Eli Lilly molecular discovery executive Andrew Adams, with Nobel laureate Michael Levitt among its members |
The Hong Kong listing is the detail worth pausing on. A public market imposes quarterly candour on a company whose central project will not report for decades. That Insilico chose to accept that discipline, rather than remain in the forgiving privacy of venture capital, is itself part of the sustainability argument. It also happened in Asia, which is not incidental to why he is speaking in Taipei.
THE VENUE Why this argument lands here
An Asian audience is the right audience for this talk, and not for sentimental reasons. Three things line up.
The first is demographic. No region has a more immediate stake in age-related disease than Asia, and the BIO Asia-Taiwan programme reflects it, with dementia given a dedicated Special Forum rather than a corner of a poster hall. The second is capital. Insilico is listed in Hong Kong, its Greater China operations are based there, and the Investment Summit downstairs is built around cross-border financing, IPO and M&A. The third is the point Zhavoronkov keeps making about consequence: when discovery gets faster and cheaper, the bottleneck relocates downstream into manufacturing, capacity and execution, which is precisely the ground this region has spent forty years mastering.
THE CAVEAT What the talk does not resolve
A feature that ended here would be an advertisement. Two things deserve stating plainly.
The first is definitional. Longevity remains a word that covers everything from serious drug development against age-related disease to peptide influencers and cosmetic aesthetics. A field that cannot police its own vocabulary will keep paying a credibility tax, and the sustainability argument is partly a response to that tax rather than a solution to it.
The second is scientific, and Zhavoronkov has been unusually willing to say it himself. The constraint on this entire enterprise is not computational. It is biological validation. Chemistry benchmarks improve in ways that are straightforward to demonstrate. Living systems do not cooperate on that schedule. A platform that compounds is a genuine advantage, but it compounds against a problem that has humbled every generation of tools brought to it so far.
The model is elegant and the numbers are real. The biology remains undefeated, and everyone serious in this field says so out loud.
THE TAKEAWAY
The most useful thing about this keynote is that it was not inspirational. It was structural. Zhavoronkov did not ask the room to believe that ageing is treatable. He assumed it, and then spent his time on the far more practical question of how an organisation stays alive long enough to find out: measure productivity honestly, license the legible work to fund the illegible work, and build an engine that gets better every time you use it.
For a region full of founders with long-horizon science and short-horizon capital, that is the more valuable talk. The mission was never the hard part. The engine is.
*** Alex Zhavoronkov delivered How to Build a Sustainable Longevity Company at BIO Asia-Taiwan 2026, Taipei Nangang Exhibition Center, on 15 July. Company figures and the keynote framework per Insilico Medicine and the published session abstract.
A BioSpectrum Asia feature.
