How Abu Dhabi's M42 Is Building AI Into Gulf Clinical Care

Most conversations about artificial intelligence in the Gulf focus on data centres and sovereign funds. A quieter shift is happening inside hospitals. M42, the Abu Dhabi health company formed in 2023 from the merger of G42 Healthcare and Mubadala Health, has spent the past two years assembling the pieces of a clinical AI system that reaches from its own language model down to the patient record. With more than 480 facilities across 26 countries, according to M42, it is large enough that decisions made in Abu Dhabi ripple across a sizable care network.

One company, a full clinical stack

The foundation is Med42, an open access clinical large language model with 70 billion parameters that M42 released so researchers and health systems could test medical question answering without a closed commercial black box, as detailed by G42. Around that model sit genomics capabilities and biobanks, and the everyday systems that clinicians actually use. The strategy is less about a single product than about owning each layer, from the model to the record, so that AI can be applied where care decisions are made rather than bolted on afterwards.

Genomics at the point of care

The clearest example is a collaboration with Oracle Health announced in late 2025, which sets out to integrate data from the Emirati Genome Programme directly into the electronic health record, as Oracle described. The aim is pharmacogenomic guidance at the point of care, meaning that when a doctor prescribes a drug, the record can flag how a particular patient is likely to respond based on their genetics. Done well, that cuts the trial and error of prescribing and the waste of treatments that were never going to work for a given person. It also depends on unified, lifelong records that follow a patient across the system, which is the harder part of the project.

From research models to daily care

The stack is starting to show up in specific conditions. In May 2026, M42 partnered with the dialysis provider Diaverum to launch an AI driven kidney health platform aimed at raising awareness and catching chronic kidney disease earlier, as Healthcare Middle East and Africa reported. Earlier in the year, an M42 data unit agreed to work with TELUS Health on personalised wellbeing services that combine precision medicine with AI. None of these is a science fiction robot doctor. They are narrow tools aimed at specific points in the care journey, which is generally where clinical AI earns its keep.

The region offers a useful contrast in ambition. Saudi Arabia drew global attention when an AI doctor clinic, a collaboration between the medical technology firm Synyi AI and Almoosa Health, began letting an AI system generate diagnoses for review by a licensed doctor, while the Kingdom's Seha Virtual Hospital has connected hundreds of facilities into one of the world's largest virtual care platforms, as Gulf News has documented. M42's approach is more incremental, wiring intelligence into existing records and pathways rather than standing up a new front door.

The governance question

The harder issues are not technical. A model that suggests a diagnosis or a prescription has to be explainable, auditable and clearly subordinate to a human clinician, especially once genetic data enters the record. The UAE has been moving to regulate AI in healthcare, and the credibility of the whole effort rests on getting consent, data protection and accountability right. A pharmacogenomic alert is only useful if clinicians trust it, and that trust is built on transparency about how the system reaches its conclusions and who is responsible when it is wrong.

M42's bet is that the winners in clinical AI will be the institutions that own the unglamorous layers, the record, the genome and the workflow, not just the headline model. That is probably right. The polished demonstrations get attention, but the durable value comes from a prescribing alert that quietly prevents a bad reaction, or a record complete enough that a patient does not have to repeat their history at every visit. The risk is that ambition outruns governance, and that a system sold on prevention and longevity ends up under scrutiny for how it handled consent or a wrong call. For now, the Gulf has a credible case that it is building clinical AI from the patient outward, which is a more promising place to start than from the press release in.