GLP‑1, SGLT2 and AI in Longevity: Where the Real Value Lies

Background: Biology Meets Commercial Reality

Santulli et al. (2024) present a compelling scientific hypothesis: that GLP‑1 receptor agonists and SGLT2 inhibitors, originally developed for type 2 diabetes, may exert systemic effects relevant to human longevity. These include reductions in cardiovascular and renal risk, anti-inflammatory properties, and potential neuroprotective effects. Concurrently, artificial intelligence is transforming how we quantify biological age, stratify populations, and model intervention impact. Tools such as epigenetic clocks and digital biomarkers (Horvath and Raj, 2018; Zhavoronkov et al., 2019) are increasingly used in early-phase development and observational studies. The convergence of drug repurposing and computational biomarkers appears to represent a promising direction for translational geroscience. However, for investors and pharmaceutical decision-makers, the pertinent question is not whether these ideas are interesting, but whether they are investable.

Context: What Is Viable, Reimbursable, and Strategically Defensible

Aging is not a regulatory indication

Despite scientific progress, neither the FDA nor EMA currently recognise “ageing” or “healthspan extension” as formal clinical indications. Ageing remains a risk factor rather than a disease. Trials such as TAME (metformin) have helped build momentum, but without a validated surrogate endpoint or reimbursed clinical condition, the commercial pathway remains undefined.

SGLT2 inhibitors are approaching loss of exclusivity. There is no viable business case for longevity

SGLT2 inhibitors, such as dapagliflozin and empagliflozin, are nearing generic entry in key markets. In the US, exclusivity expires by 2025 or 2026, and many patents in Europe are already weak or expired. Pursuing a longevity indication on these compounds would entail considerable clinical development cost, without the benefit of exclusivity, premium pricing, or payer traction. From an investment perspective, this offers poor return on capital.

GLP‑1 receptor agonists are more promising, but require precision

By contrast, GLP‑1 analogues, such as semaglutide and tirzepatide, retain patent protection well into the next decade. These assets are actively expanding into obesity, cardiovascular prevention, NASH, and potentially age-related muscle or cognitive decline. In this context, AI biomarkers could help refine population targeting or dose optimisation. However, the focus must remain on reimbursable endpoints, not speculative “longevity” claims.

AI biomarkers have potential, but should not be overstated

AI-generated ageing clocks and biological signatures are a useful innovation in the toolkit of R&D efficiency. They can support responder identification, adaptive trial design, and long-term risk modelling. However:

• No AI biomarker is currently accepted as a surrogate endpoint by regulators.

• Most are trained on publicly available data, limiting intellectual property defensibility.

• Their biological mechanisms are often unclear, which reduces their regulatory credibility.

In short, AI biomarkers may streamline development. However, they are unlikely to unlock reimbursement or regulatory approval on their own.

Implications: What Should Actually Guide Investment and Development

A credible longevity strategy must be built on three foundations: exclusivity, clinical relevance, and regulatory feasibility. Five principles apply.

1. Focus on reimbursable, age-related indications: Investors and developers should prioritise conditions where biological ageing plays a known role, such as frailty, sarcopenia, cognitive impairment in metabolic disease, or CKD-related muscle loss. These offer viable trial endpoints, payer engagement pathways, and regulatory precedents.

2. Do not pursue ageing claims on compounds nearing patent expiry: SGLT2 inhibitors, while mechanistically interesting, offer no remaining exclusivity. There is no justification for launching a clinical programme in this area unless paired with a differentiated formulation, delivery device, or diagnostic layer that restores commercial defensibility.

3. GLP‑1s provide a viable commercial bridge, but require positioning discipline: The GLP‑1 class offers scope for lifecycle extension and indication stacking. However, any expansion into age-adjacent use cases must be underpinned by reimbursable endpoints and robust clinical design. “Healthspan” as a label is unworkable. Function, independence, or cardiovascular protection may be more defensible framings.

4. Treat AI biomarkers as operational tools, not endpoints: AI biomarkers should be embedded in development workflows to enhance trial efficiency. They are most valuable when supported by proprietary datasets, feedback loops, and early-stage regulatory dialogue. Mere model accuracy is insufficient. Regulators require clarity, reproducibility, and proven clinical relevance.

5. Prioritise companies that operate at the interface of biology, regulation, and data: The most investable platforms will combine:

• A validated biological mechanism

• Exclusive access to trial-grade datasets

• Biomarkers that inform development or payer strategy

• Ongoing engagement with regulators

These firms are more likely to secure co-development partnerships, command premium valuations, and deliver strong, risk-adjusted returns.

Final Thought

Santulli et al. (2024) describe a scientifically rich and emerging area of exploration. Yet in the commercial world, strategy must go beyond enthusiasm. The next wave of value in longevity will be captured not by those who make the boldest claims, but by those who align biology, regulatory reality, and market structure with discipline and precision.

In short, investors should look not for “anti-ageing” stories, but for commercially coherent, age-relevant strategies that deliver exclusivity, reimbursement, and a measurable return on investment.

Sources:

• Fuellen, G., Liesenfeld, D., Kowald, A., et al., 2019. Health and aging: unifying concepts, scores, biomarkers and pathways. Aging Disease, 10(4), pp.883–900.

• Horvath, S. and Raj, K., 2018. DNA methylation-based biomarkers and the epigenetic clock theory of ageing. Nature Reviews Genetics, 19(6), pp.371–384.

• Santulli, G., Mone, P., Varzideh, F., et al., 2024. GLP-1 receptor agonists and SGLT2 inhibitors: new anti-aging tools? Future Cardiology, 21(1), pp.5–8.

• Zhavoronkov, A., Mamoshina, P., Vanhaelen, Q., et al., 2019. Artificial intelligence for aging and longevity research: recent advances and perspectives. Ageing Research Reviews, 49, pp.49–66.

Dillon Shokar, (Biostatistics, King's College London; Data Science, Harvard), CFA

Biopharma Venture Partner | Strategy Consultant