During the peak of his career, Ma Yun led Alibaba in several ventures that appeared, on the surface, to be less than "cost-effective," including the establishment of the Hupan University, the DAMO Academy, and the Luohan Academy. He repeatedly emphasized at the time that if a company dedicated all its resources solely to profit-making, it would be "a company with no promise." Within that narrative, these institutions were imbued with a significance transcending business cycles, envisioned as long-term assets Alibaba would bequeath to the world.
Amid Alibaba Group's recent organizational transformations, the DAMO Academy inevitably underwent strategic refinement and structural adjustments; compared to teams like Alibaba Cloud, it seemed to be receding from the center stage. However, at the beginning of 2026, The New York Times dedicated a full-page report that thrust DAMO Academy back into the global spotlight. The report detailed DAMO's AI model for early pancreatic cancer screening—DAMO PANDA—which has assisted doctors in detecting lethal tumors that might otherwise have been missed, saving the lives of several patients.
Such a story is undoubtedly inspiring globally, not only showcasing the formidable potential of AI in healthcare to a broader public but also serving as another successful case study of Chinese AI technology being applied in real-world scenarios. Furthermore, this achievement is poised to enhance the global influence of Alibaba's AI technology.
According to insights, the DAMO Academy's medical AI team chose its first adversary in cancer screening research to be the "king of cancers"—pancreatic cancer. After years of foundational research, the team unveiled the DAMO PANDA model in November 2023. DAMO PANDA constructed the largest-ever training dataset of pancreatic tumor CT scans and, through a retrospective validation involving 20,530 real-world patients, identified 31 clinically missed lesions; two patients with early-stage pancreatic cancer detected this way have already undergone successful curative surgery, with the related paper published in the top-tier medical journal Nature Medicine.
Diverging from the traditional cancer screening approach that often "relies on contrast-enhanced CT or invasive examinations," DAMO PANDA is built upon a technical pathway systematically pioneered by the DAMO Academy—"non-contrast CT + AI." Non-contrast CT is one of the most widely available imaging techniques within China's healthcare system, characterized by low cost and broad accessibility, even at grassroots hospitals and county-level medical institutions. However, due to limited image contrast, its use has long been primarily confined to diagnosing conditions like pulmonary nodules, fractures, and stones.
With the accumulation of large-scale data and advances in deep learning technologies, the DAMO research team discovered that AI can identify minute density variations in non-contrast CT images—variations极易 overlooked by the human eye—thereby uncovering valuable clues for disease detection. The significance of this pathway lies in its approach: instead of introducing an expensive, complex, and hard-to-disseminate new examination method into the healthcare system, it integrates into existing workflows, extracting additional health insights from scans "that patients were already scheduled to undergo."
According to DAMO Academy's estimates, screening for the seven most common cancers (lung, colorectal, liver, stomach, breast, esophageal) using traditional methods like ultrasound, CT, gastroscopy, and colonoscopy would cumulatively cost at least 3,000 RMB. In contrast, the "one-scan-multi-screening" approach based on non-contrast CT+AI costs less than 200 RMB, making large-scale screening of asymptomatic populations a realistic possibility.
Simultaneously, DAMO PANDA does not exist in isolation. Over the past two years, the DAMO Academy has clearly designated medical AI as one of its core research priorities. Beyond pancreatic cancer, the team has achieved a series of breakthroughs in other high-incidence cancers like esophageal, gastric, colorectal, and liver cancer, as well as in chronic conditions like osteoporosis and fatty liver disease, and emergencies like acute aortic syndrome, with multiple results published in Nature Medicine.
In June 2025, DAMO Academy, in collaboration with Zhejiang Cancer Hospital, also released the world's first AI model for gastric cancer imaging screening, DAMO GRAPE. In Alibaba's first decade, Ma Yun forbade any internal discussion about establishing a research institute or R&D department. During those initial ten years, survival was the daily concern for Alibaba; it "needed good products, good services, and profitability to move forward." After Alibaba reached its 18th anniversary, boasting over 20,000 technical talents and more than 500 million users, Ma Yun began contemplating longer-term strategies, with the first major initiative being the founding of the DAMO Academy in October 2017.
The DAMO Academy was dedicated to tackling scientific and technological challenges that enhance productivity, with Ant Group pledging at the time to invest 100 billion RMB into DAMO over three years. However, against the backdrop of Alibaba Group's significant organizational reforms over the past three years, the DAMO Academy has also undergone successive adjustments and reshufflings. Its previously diverse "4+X" research fields have been streamlined, retaining only "Intelligence + Computing." The Intelligence direction encompasses Medical AI, Decision Intelligence, Video Technology, Embodied AI, and Genomic AI, while the Computing direction includes Computing Technology and RISC-V.
Evidently, post-adjustment, projects that were vague in direction, excessively long-cycle, or difficult to commercialize have been spun off, scaled back, or terminated. What truly remains are not necessarily the most cutting-edge concepts with the grandest imaginations, but rather those achievements that have been validated in the real world, can be integrated into existing systems, and generate stable positive feedback.
In this context, medical AI has become one of the most stable and continuous focus areas for the DAMO Academy in recent years. Unlike frontier fields such as autonomous driving and quantum computing, which require massive investments and face highly uncertain commercialization paths, medical AI was embedded into specific application scenarios from the outset: hospitals, medical examination centers, and primary care institutions.
According to available information, beyond the DAMO AI team, Alibaba's current AI-related organizational structure also includes Alibaba Cloud, Alibaba's Intelligent Information Business Group, and AI teams within various business units. Among these, Alibaba Cloud serves as the core platform for Alibaba's AI technology deployment and commercialization, focusing on AI infrastructure development, large language models, and B2B commercialization.
The positioning of the DAMO Academy has also become clearer—acting as the group's frontier research hub, undertaking long-term technological exploration tasks with significant potential social value. While Alibaba emphasizes commercial efficiency and cash flow more strongly, and while "poetry and distant horizons" are continually compressed by现实 pressures, the DAMO Academy appears to persist in pursuing work that can genuinely benefit humanity.
Comments