As global Big Tech companies wage an all-encompassing battle for the medical artificial intelligence (AI) market, the scale of revenue Microsoft generates through OpenAI has been disclosed for the first time. The medical AI competition, which began with diagnostic assistance tools, is now spreading across hospital workflow systems and drug discovery platforms, escalating into a “platform war” to seize control of the industry’s foundational infrastructure.
According to industry sources on the 6th, Microsoft revealed in recently disclosed regulatory filings that it generated approximately $24.1 billion (roughly 34.2 trillion won) in revenue from OpenAI during its 2026 fiscal year (July 2025 – June 2026). When compared to the annual AI revenue target of $37 billion (roughly 52.5 trillion won) that Microsoft had disclosed through the March quarter, and the actual AI revenue of $34 billion (roughly 48.2 trillion won) estimated by Bloomberg Intelligence, this figure suggests that OpenAI could account for more than 70% of Microsoft’s total AI business revenue.
This revenue concentration demonstrates that Microsoft’s AI business remains heavily dependent on OpenAI for the foreseeable future. OpenAI is the largest customer training and running inference on large language models via Microsoft’s Azure cloud service, while simultaneously serving as the core intelligence behind the Copilot services Microsoft sells to enterprise customers. Microsoft’s structure generates profit by integrating OpenAI’s technology into its productivity tools, such as Microsoft 365 and GitHub.
Competition Moves Beyond Diagnosis to Dominating Hospital Workflow Systems
Microsoft’s influence is particularly pronounced in the medical field. Dragon Copilot, Microsoft’s healthcare-specific AI assistant, listens to conversations between doctors and patients in real time and automatically drafts electronic medical records (EMRs). It is a tool designed to help clinicians move away from simple documentation tasks and focus on patient care itself. This year, Microsoft is expanding the service from a simple voice recording tool into a clinical AI platform that integrates overall hospital workflows. The goal is to enable medical information searches during consultations, patient data lookups, and integration with external medical AI agents, all processed on a single screen.
Nvidia is expanding its market presence by simultaneously supplying the hardware and software stack that serves as the “brain” for medical AI. Nvidia’s healthcare and life sciences platform, BioNeMo, is a tool that analyzes protein, DNA, and RNA structures to design and virtually validate new drug candidates. In June, the company unveiled a research AI agent toolkit, evolving its capabilities to the point where AI can collect and analyze papers and experimental data to suggest the next research direction. Nvidia is already collaborating with companies like Eli Lilly and Thermo Fisher Scientific to apply its computing infrastructure to drug discovery and automated laboratory infrastructure.
Google is pursuing a strategy that combines a healthcare-specific generative AI model with its cloud data platform. MedGemma, developed by Google DeepMind, is an open AI model capable of analyzing not only medical documents but also medical images such as X-rays and pathology slides. Google is fostering an open ecosystem so that hospitals and medical AI startups can train the model on their own proprietary data to develop specialized services. In parallel, the company is also engaged in an infrastructure business that converts scattered hospital medical data into a format usable by AI through the Google Cloud Healthcare API.
Rather than developing medical AI models directly, Amazon Web Services (AWS) is adopting a strategy of providing cloud infrastructure and distributing a variety of partner models so that hospitals, pharmaceutical companies, and medical AI firms can build generative AI services. Pfizer is also collaborating with AWS on experiments to integrate generative AI into the drug discovery process. AWS is moving to dominate the distribution network of the ecosystem by serving as a marketplace where externally developed healthcare and life sciences AI models can be bought and sold through the AWS Marketplace.
AI Supremacy Shifts from “Most Accurate Diagnosis” to “Platform Control”
The changing competitive landscape in the medical AI market is interpreted as a fight to preempt industry standards and the operating system (OS), moving beyond a simple technology race. In the past, the core focus was on how accurately an AI could read an X-ray or CT scan. Now, who controls the entire workflow of medical staff has become more critical.
An industry insider analyzed, “The ultimate winner of the medical AI territory war is likely to be not the company that developed the most accurate diagnostic algorithm, but the one that occupies the platform through which all medical AI must pass.” The outlook suggests that as hospitals, pharmaceutical companies, and medical AI startups grow, an industrial structure could solidify where they pay Big Tech for graphics processing units (GPUs), cloud services, and software usage fees.
This trend aligns with movements in the investment market. On Wall Street, AI investment sentiment is rapidly shifting from “who is buying more AI chips” to “who is converting AI into actual revenue.” In Microsoft’s case, a key strength is its ability to cross-sell AI services to existing enterprise customers without separate customer acquisition costs by integrating AI into its existing Office software and cloud services through its partnership with OpenAI. In fact, paid users of Microsoft 365 Copilot surged from 20 million in the previous quarter to over 30 million, and Azure cloud revenue surpassed an annualized $100 billion (roughly 141.8 trillion won) for the first time.
However, high dependency also represents a risk. To reduce its reliance on OpenAI, Microsoft is attempting portfolio diversification by investing in Anthropic and developing its own AI models. The market views the proportion of cloud usage fees and profit-sharing payments from OpenAI within Microsoft’s total AI revenue as a key yardstick for evaluating the future value of Microsoft’s AI business.