The medicine of tomorrow: where AI can really take us

A frequently caricatured promise

"AI will cure cancer." "We'll live to 150." "Death will be optional." These phrases have been circulating for five years, pushed by Californian entrepreneurs with ambitious PR. They have helped discredit any serious discussion of what AI is actually doing to medicine, because nobody wants to be associated with the marketing promise of the billionaire who wants to "defeat death".

Yet behind these sometimes ridiculous statements, something real is happening. In 2026, several concrete advances in AI-assisted medicine are moving from the lab to the clinic. Not the end of disease. Not eternal life. But structural transformations that will probably change how we fall ill, how we recover, and how many years we spend in good health. This article gives an honest assessment of what is credible over the next 30 years, following on from our article on prospective AI scenarios.

What has already been achieved by 2026

Let's start with what's already in place. Several things have already been accomplished today, even if they don't make the front pages.

AlphaFold has mapped 200 million proteins. DeepMind (Alphabet's AI subsidiary, Alphabet being Google's parent company, based in London and led by Demis Hassabis) published in 2021 and then 2024 (with AlphaFold 3) the structure of virtually every known protein on Earth. As a reminder, proteins are the molecules that make every living thing work: enzymes, antibodies, hormones. Knowing their 3D structure makes it possible to understand how they act and to design targeted drugs. It is a century of biological work condensed into a few years. Demis Hassabis (co-founder and CEO of DeepMind, a British researcher trained in neuroscience) and John Jumper (lead researcher on AlphaFold) received the 2024 Nobel Prize in Chemistry.

The drug development cycle is getting shorter. Insilico Medicine (a company based in Hong Kong and founded in 2014, valued at several billion and listed on the Hong Kong stock exchange) uses AI to identify therapeutic targets and design molecules. Its Pharma.AI platform has produced around ten drug candidates now in clinical trials. The average time to go from target to clinical candidate has fallen from 4-5 years to 12-18 months. Several other companies (Recursion, Exscientia, BenevolentAI) are following the same trajectory.

AI diagnostics match or exceed humans. On the detection of skin cancers, certain diabetic retinopathies (damage to the retina caused by diabetes), and several radiological cancers, the best AI systems now match or exceed experienced human radiologists. These tools are already deployed in hospitals in Asia, Europe and North America, under strict medical validation protocols.

The figure that changes everything 🧬
In 2026, around 15% of oncology clinical trials are designed with the help of AI algorithms to select the patients most likely to respond to treatment. This personalisation speeds up trials (fewer patients needed) and increases success rates. That figure stood at 2% five years ago. It could exceed 50% by 2030.

The cellular regeneration gamble: what is really happening at Altos Labs

The most widely covered topic remains "cellular reprogramming", popularised by Altos Labs (a Californian laboratory founded in 2022 with a record initial investment of 3 billion dollars, backed by Jeff Bezos and other billionaires, dedicated to research on cellular rejuvenation). Let us break down what it is and what it can really do.

In 2006, Shinya Yamanaka (a Japanese researcher and professor at Kyoto University) discovered that four proteins, now known as "Yamanaka factors", can turn an adult cell into a pluripotent stem cell (a cell capable of becoming any type of cell in the body). This discovery, which earned him the Nobel Prize in Medicine in 2012, opened up an immense therapeutic field, but with one major problem: applied bluntly, reprogramming turns cells into embryonic cells, which causes cancers and malfunctions.

The brilliant idea behind current research is partial reprogramming. Instead of resetting the cell to zero, just enough Yamanaka factors are injected to erase the markers of ageing, without touching its identity (liver, heart, neuron). This is what Jerry McLaughlin, CEO of Life Biosciences (an American company specialising in longevity, founded in 2017), calls "taking that corrupted software and restoring it to factory settings".

The preliminary results in mice are impressive. Several studies (including a major Altos publication in 2024) show that partial reprogramming extends lifespan and healthspan in mice. Altos is also working "ex vivo" (outside the organism, on harvested organs): taking an old kidney, rejuvenating it outside the body, then transplanting it.

The critical step in 2026: human clinical trials begin. Life Biosciences is entering phase 1 clinical trials this year (the stage where a new treatment is tested on a small number of patients to check its safety before phases 2 and 3, which test efficacy in larger populations) on patients with optic neuropathy (disease of the optic nerve that can cause blindness). Altos Labs, which had so far remained in the preclinical phase (before human testing), recruited Joan Mannick (an expert in longevity clinical trials) in 2025, signalling the forthcoming move to human studies. The first indications targeted: neurodegenerative diseases, idiopathic fibrosis (a disease in which lung tissue gradually becomes stiff and scarred), age-related immune disorders.

A note on scientific honesty ⚖️
A recent study from UC San Diego (University of California San Diego, one of the leading American research universities) argues that somatic mutations (changes to DNA that arise during life in non-reproductive cells) rather than epigenetic drift (changes in gene expression without altering their DNA sequence) could be the true cause of ageing. If that is correct, Altos's therapies would be addressing symptoms rather than the root of the problem. The scientific debate is far from settled.

The market: $420 billion by 2030

The economic scale of the longevity bet deserves closer attention. According to several converging analyses, the longevity medicine market will reach around $420 billion by 2030, with annual growth of 8 to 9%. Venture capital investment (high-potential, high-risk investment in young companies) in the sector tripled between 2022 and 2025.

The main players and their funding:

Altos Labs: launched in 2022 with a historic record of $3 billion. Backed by Jeff Bezos (founder of Amazon) and several other billionaires. Based in San Francisco, San Diego and Cambridge in the United Kingdom.
Calico: a Google subsidiary created in 2013, with more than $2.5 billion invested. Focused on the biology of ageing.
Retro Biosciences: funded personally by Sam Altman (CEO of OpenAI) to the tune of $180 million. Focused on cellular reprogramming and autophagy (the natural mechanism by which cells recycle their own damaged components).
Life Biosciences: entering clinical trials in 2026. Approach based on "partial epigenetic reprogramming".
Insilico Medicine: focused on AI plus chemistry. Listed in Hong Kong in 2024.
NewLimit: founded by Brian Armstrong (CEO and co-founder of Coinbase, the largest American cryptocurrency exchange platform). Focused on transcription factors (proteins that regulate gene activation) and cellular rejuvenation.

This accumulation of private capital in longevity is unprecedented. No biomedical field has ever attracted so much investment in so little time, with the possible exception of cancer immunotherapy (treatment that uses the patient's immune system to fight tumours) in the years 2015-2020.

The other frontiers AI is pushing back

Beyond longevity, several other fields are advancing rapidly thanks to AI.

Digital patient twins. The concept of a "digital twin" (a complete virtual replica of an object or living being that simulates its real behaviour) is being applied to patients by several university hospitals (Stanford, MIT, Karolinska in Stockholm). They are developing personalised models of each patient: their genome, their microbiome (the set of micro-organisms living in their body, particularly the gut), their medical history, their biomarkers (measurable biological indicators). These digital twins make it possible to simulate, before treatment, how a patient will respond to different options. First clinical applications expected around 2028-2030.

Proteomics and personalised genomic medicine. Proteomics is the study of all the proteins in an organism. Advances in AlphaFold now make it possible to understand how each mutated protein in a patient affects their metabolism. Eventually, a doctor will be able to prescribe a drug specifically tailored to a patient's genetic variations, rather than a standard treatment. This approach already exists in part in oncology, and will become widespread in other specialities by 2030.

Mass early detection. Combining blood tests (in particular "liquid biopsy" tests, which detect fragments of tumour DNA in the blood), medical imaging and behavioural data (smartwatches), AI can detect several cancers and cardiovascular diseases several years before symptoms appear. Galleri (a test developed by the company Grail) already offers a blood test that detects 50 types of cancer, with steadily improving accuracy.

Designing new antibiotics. Faced with growing bacterial resistance, laboratories are using AI to discover new classes of antibiotics. In 2023, MIT published the discovery of a next-generation antibiotic (Halicin) identified entirely by AI. Several others are in clinical development.

Organ regeneration through bioprinting. Bioprinting is 3D printing applied to biological tissues. Combined with AI, it is progressing rapidly. First functional kidneys printed and tested on animals. Human application expected in the 2030-2040 decade, mainly for skin, cartilage and certain simple structures.

What this concretely changes in 10 years

If we cautiously extrapolate current trends, here is what will probably be true in 2036.

Area	2026	2036 (likely)
Cancers with 5-year+ survival	~70%	~90%
Early detection of serious diseases	Limited to certain cancers	Widespread annual preventive diagnosis
Neurodegenerative diseases (Alzheimer's, Parkinson's)	Modest slowing	First genuinely effective treatments
Healthy life expectancy (wealthy countries)	70 years	85-95 years
Drug development cycle	10-15 years	3-5 years

Important note: these projections concern wealthy countries with access to cutting-edge care. The question of the equitable distribution of these advances is probably the biggest political question in twenty-first-century medicine. A cell reprogramming therapy costing 500,000 dollars per cycle would benefit only a tiny minority, creating a "biological inequality" unprecedented in human history.

The limits you should honestly be aware of

To avoid falling into the naive enthusiasm that often characterises this subject, four limits must be named.

Limit 1: human biological complexity. Many approaches that work in mice fail when they move to humans. Human biology is more complex, takes longer to study, and side effects can take years to appear. Many of the promises of 2020-2025 have already disappointed in the clinic.

Limitation 2: the costs. Personalised gene therapy often costs several hundred thousand dollars per patient. If these costs do not fall drastically, these treatments will remain out of reach for most of humanity.

Limitation 3: slow regulation. The FDA (Food and Drug Administration, the US medicines agency that authorises the marketing of treatments in the United States) and the EMA (European Medicines Agency, its European equivalent) apply rigorous approval protocols that typically take 7-10 years. This caution is necessary (poorly validated medicines kill people), but it also slows the arrival of innovations that could save lives.

Limitation 4: cultural expectations. Society is not prepared for advances that would massively extend life. Pension systems, the labour market, family structures, urban planning: all of this assumes stable lifespans. A biological breakthrough could precede the necessary social adaptations by several decades.

Why this subject deserves your attention right now

Three reasons to actively follow these developments, even if you are not ill today.

Firstly, today's political choices will determine who has access to these advances. The question of reimbursement by the French social security system for longevity therapies will be one of the great debates of the 2030s. Positions are being formed now.

Secondly, personalised medicine is arriving faster than collective awareness. In 10 years, your doctor will probably be able to prescribe a treatment tailored to your genome rather than a generic one. This transition requires medical education, infrastructure, and a rethinking of the carer-patient relationship.

Thirdly, healthy life expectancy radically influences individual choices. If living 90 years in good health becomes likely for the next generation, that changes the calculation around long studies, career strategy, savings, late parenthood, social commitment. As we explored in our article on prospective scenarios, these social adjustments are as important as the technology itself.

The medicine of the future will not be the eternal life promised by transhumanists. But nor will it be today's medicine simply improved. It will be structurally different, more personalised, earlier, more effective for many pathologies that are fatal today. To go further, you can consult our article on scientific discoveries by AI, which explores how AlphaFold and its successors are transforming all medical research, and our analysis of the impossibility of regulating AI, which places these medical issues within a broader legal debate.