By Sergey Jakimov and Artem Trotsyuk, managing and operating partners, LongeVC
Artificial intelligence is rapidly rewriting medicine – from clinics to labs – and scientists, startups, regulators, and investors are shaping that shift together.
The global healthcare AI market is projected to soar from just over $1 billion in 2016 to more than $28 billion by 2025.
The first AI-designed molecules are already in human trials, compressing timelines that once stretched a decade into just a few years.
This marks a structural change in how therapies are designed and how we define patient care.
A data-driven revolution is underway, and its upside is enormous.
Healthcare has felt AI’s presence in imaging, triage, and early diagnostics, but the deeper transformation is happening in labs and discovery pipelines.
AI systems now predict protein folding (Demis Hassabis and John Jumper received the Nobel Prize in Chemistry 2024 for their AlphaFold2 model), map new drug targets, and propose molecular structures at a pace that traditional methods cannot match.
Where traditional discovery might take five years and hundreds of millions before a first clinical trial, AI is cutting that in half.
Insilico Medicine, for example, went from target identification to a Phase I trial in about 30 months with its idiopathic pulmonary fibrosis candidate, giving proof that generative chemistry can yield clinical assets.
AOA Dx offers another glimpse of what this transformation looks like in practice.
Sergey Jakimov
The company is using AI to interpret complex biomarker data from blood, aiming to detect ovarian cancer at its earliest stages: when survival rates are above 90 per cent, compared with less than 20 per cent for late diagnoses.
In its recent trial with over 500 women, AOA’s algorithm-driven approach achieved specificity and sensitivity levels that exceeded FDA thresholds, positioning it as the first to demonstrate reliable blood-based early detection for this cancer.
It’s a case study in how AI can open entirely new diagnostic frontiers, not just streamline existing processes.
For years, “personalised medicine” was more slogan than reality. Early efforts often relied on broad biomarkers or risk categories, treating patients as members of subpopulations rather than as unique individuals.
Truly individualising therapy requires integrating vast multi-omic and real-time data streams. And this challenge was simply unfeasible with older tools.
Now, AI is providing the infrastructure to make it practical.
Companies are integrating genomics, microbiome data, and digital biomarkers into algorithms that guide individual-level decisions.
A European SaaS startup, Haut.AI, is doing this in dermatology by analysing skin phenotypes to recommend targeted interventions.
The company reports its model achieves roughly 98 per cent diagnostic accuracy across diverse skin types by analyzing more than 20 clinically relevant skin metrics, enabling finely tuned interventions rather than one-size-fits-all skincare.
Ani Biome is applying generative AI to microbiome analysis, producing supplements that aim to optimize gut health and reduce inflammation.
These are early steps toward a future where therapies and preventions are tuned not to populations, but to individuals. We see it as a signal of a true shift in preventive care.
By continuously integrating genomic, microbiome, and biomarker data, future algorithms could update each person’s health plan in real time to slow ageing.
That this kind of n-of-1 precision is exactly what will underpin breakthroughs in age-related disease and healthspan extension.
The capital markets are noticing the shift as well. Investment in AI-driven healthcare ventures has surged, even though the investing market stays pretty rough (make this sentence better.
Pharma companies are partnering with AI labs not as experiments but as core R&D bets.
For example, Merck recently inked a $349M collaboration with Variational AI to co-design novel small molecules.
As reported by Silicon Valley Bank, venture investment in healthcare AI reached US$5.7B across 320+ deals in the first half of 2025, making it one of the most active subsectors.
Biopharma accounted for the majority, reflecting growing confidence that AI is not just an efficiency tool but a true source of pipeline innovation.
The capital markets are also noticing the shift.
Venture dollars are tilting toward AI-native plays, even amid a cautious funding environment.
Artem Trotsyuk
As reported by Silicon Valley Bank, U.S. healthtech has seen the biggest share of the AI healthcare boom: trailing 12-month AI deal activity in healthtech has roughly doubled since 2022 and accounted for nearly a third of all healthcare investment in the first half of 2025.
Healthcare-specific AI models, pure software drug-discovery platforms, and a wave of new administrative tools have made the sector one of the most attractive categories for investors.
Pharma companies are also signing AI-first partnerships: Merck recently inked a US$349M collaboration with Variational AI to co-design novel small molecules, in September 2024 Novartis signed a deal with Generate:Biomedicines (up to US$1B+) for AI-designed proteins.
Startups like Ani Biome, Glyphic Bio, Melio, and Valink Therapeutics are applying AI to everything from microbiome analysis to signal deconvolution and drug design, showing how machine learning is now embedded across the discovery value chain.
Better prediction reduces failed trials, saving billions while de-risking pipelines.
But the deeper opportunity is strategic: AI allows small, early-stage teams to generate assets that once required the budget of Big Pharma, changing who gets to compete in therapeutics.
The implication for longevity biotech is profound: a tiny team can now explore aging-related pathways without a billion-dollar budget.
By democratising drug discovery, AI could spark a wave of startups tackling neglected geroscience targets.
This lower barrier to entry may rapidly accelerate the development of geroprotective therapies that were previously too complex or niche for larger companies.
Many commercial “anti-aging” or “AI health” products overpromise, and clinicians rightly point to bias and explainability gaps in black-box models.
Regulators are starting to respond: the FDA and EMA have both issued early guidance on adaptive algorithms in healthcare.
If trust is to grow, transparency and clinical validation will need to be non-negotiable.
In fact, the FDA’s AI Action Plan explicitly emphasises bias mitigation and robust validation of medical AI.
Any AI-derived therapy will need the same rigorous trials and peer-reviewed evidence as a traditional drug.
In practice, this means the field must police hype aggressively – promising results backed by data, not snake oil.
Without clear, evidence-based validation, even groundbreaking algorithms will be viewed skeptically.
The thesis is simple: drug discovery is being redefined by algorithms, and medicine will not look the same in a decade.
AI has already delivered its first clinical candidates. The question is no longer “if,” but “how fast”, and whether the industry can scale these tools responsibly.
What’s clear is that the winners in biotech and longevity will be those who learn to work with algorithms not as black boxes, but as partners in the next era of medical innovation.
