AI Drug Discovery Is Reshaping Longevity Medicine. Is Your Practice Ready?

Last week a patient asked me about a senolytic compound she heard about on a podcast. It had been discovered by an AI platform, was already in Phase II trials, and she wanted to know if she should try it.

I had not heard of it. That is happening more often.

On March 29, Eli Lilly committed $2.75 billion to Insilico Medicine for AI-discovered drug candidates. Insilico has 28 drugs in its pipeline. Nearly half are already in clinical trials. Their first compound went from target identification to Phase I in under 30 months. Traditional pharma averages 4 to 6 years for the same milestone.

For those of us practicing longevity and preventive medicine, this deal is not just a pharma headline. It is a preview of what our clinical lives are about to look like.

What Lilly is actually buying

This is not Lilly’s first bet on Insilico. The two companies have worked together since 2023 and signed a $100 million partnership in November 2025. This latest deal is a tenfold escalation. Lilly paid $115 million upfront, with the remaining $2.63 billion tied to milestones and royalties, in exchange for global rights to develop and commercialize novel oral therapeutics from Insilico’s platform.

What makes Insilico different from most AI drug companies is where it started. Its founder, Alex Zhavoronkov, built the platform around aging biology. PandaOmics, their target discovery engine, has been used to identify dual-purpose targets linked to the hallmarks of aging. Chemistry42, their molecular design engine, generates novel compounds from scratch in days.

Weeks before this deal, researchers from both companies published a paper in ACS Central Science called “From Prompt to Drug: Toward Pharmaceutical Superintelligence”. The title is ambitious. But the framework they describe, fully autonomous AI-orchestrated drug discovery from target to clinical planning, is not theoretical anymore. They are building it.

Lilly is also the company behind Mounjaro and Zepbound, and is reportedly planning a longevity-focused trial evaluating tirzepatide’s effects on biological aging. They are placing parallel bets on AI discovery and geroscience. I find that combination genuinely significant.

The pipeline is bigger than one deal

The Lilly-Insilico collaboration is the largest, but it is not isolated.

As of early 2026, more than 173 AI-discovered programs are in clinical development. Roughly 94 in Phase I, 56 in Phase II, and 15 in Phase III. Between 15 and 20 programs are expected to enter pivotal trials this year.

The early performance data gives me cautious optimism. AI-discovered compounds show 80 to 90 percent Phase I success rates compared to historically around 40-65 percent. Phase I performance is clearly better with AI; Phase II shows no demonstrated superiority in the current evidence base.

Phase III is where I hold my breath. That is where most drugs fail, and the AI-discovered pipeline is only now reaching that stage. We will know a lot more by the end of this year.

Other deals worth watching: Gero signed a research agreement with Chugai Pharmaceutical for AI-discovered therapies targeting age-related diseases.Biophytis presented its AI longevity platform at NVIDIA GTC 2026. Researchers at Scripps Research and the biotechnology company Gero used AI to identify novel anti-aging candidates that extended lifespan in animal models, with over 70 percent of the identified compounds showing significant results.

The field is moving. Fast. Here is what I keep thinking about as a practicing clinician.

Right now, I might be managing a patient on a GLP-1 agonist, monitoring their NAD+ levels, and considering off-label rapamycin. That is already complex. Each intervention generates data across labs, wearables, imaging, and genetics that I need to synthesize into a coherent picture of how this specific patient is responding.

Within the next few years, AI-discovered compounds targeting senescence, fibrosis, mitochondrial dysfunction, and inflammatory pathways will start entering clinical availability. The number of potential interventions per patient is going to increase. The data per intervention is going to increase. The need to track interactions between interventions is going to increase.

And the time I have per patient is not going to increase.

This is the tension I think most of us feel but rarely name. We want to offer the best available care. We want to be early adopters when the evidence supports it. But the operational reality of synthesizing multi-source data for every patient, manually, is already at capacity. The bottleneck is real.

When that patient asked me about the AI-discovered senolytic, what I actually needed in that moment was not more clinical knowledge. I needed a system that could pull up her full longitudinal profile, cross-reference the compound’s mechanism with her existing protocol, and flag potential interactions. In under five minutes, not five hours.

The gap between discovery and delivery

AI is compressing the discovery timeline. Insilico can scan billions of compounds in weeks. What used to take a decade of bench work now takes months of computation.

But nothing is compressing the delivery timeline. The way most longevity practices operate today, manually consolidating data from multiple lab portals, wearable platforms, imaging centers, and genetic testing companies, is the same as it was five years ago. A single comprehensive patient analysis can still take 5 to 10 hours.  

I am not making a technology argument. I am making a math argument. More therapeutic options times more data per option times the same number of clinical hours equals something that breaks.

The practices that will thrive when the next wave of longevity therapeutics arrives are the ones that have already solved this. Unified patient data. Longitudinal tracking that measures each intervention against the patient’s own baseline. Clinical AI that operates with traceability and physician oversight, not as a black box.

This is the problem we are building against. Not replacing what physicians do but making the complexity manageable so we can actually use these new tools well.

What I would tell a colleague

If you are practicing longevity or preventive medicine, here is what I think matters right now:

Pay attention to the pipeline. 173 AI-discovered programs in clinical development is not a number you can ignore. Some of these will reach your patients within a few years, whether you prescribe them or not.

Prepare for the conversation. Patients are already hearing about AI-discovered drugs on podcasts and social media. They will ask. Having a framework for evaluating these compounds, what evidence you need before incorporating them, is more valuable than memorizing each one.

Invest in your data infrastructure. This is not glamorous advice. But the clinics that can synthesize a patient’s full profile across multiple data sources quickly will have a real advantage when therapeutic complexity increases.

Stay skeptical in the right places. Phase I and II success rates are encouraging. But we have been here before with promising early data. Phase III is the test. Watch the results coming later this year before drawing conclusions.

Do not wait for the drugs to arrive to prepare. The infrastructure problem is already here. It just gets worse with each new therapeutic option.

Photo: metamorworks, Getty Images

This post appears through the MedCity Influencers program. Anyone can publish their perspective on business and innovation in healthcare on MedCity News through MedCity Influencers. Click here to find out how.