How Massive Data Shapes the AGI Era: Challenges and Opportunities

At the third OceanBase Developer Conference, Ant Group CTO He Zhengyu delivered a keynote titled “AGI Era, Qualitative Changes Brought by Massive Data,” exploring how the AI era reshapes data usage and the opportunities and challenges generative AI presents for data infrastructure.

Four Major Data Challenges

Rising acquisition cost: Public internet data is becoming scarce, turning cheap, abundant data into a depleted resource and making high‑quality, costly data the new competitive advantage.

Industry data scarcity and flow difficulty: Highly regulated sectors such as law and healthcare generate valuable but hard‑to‑access data, creating structural gaps that hinder generative AI applications.

Multimodal data processing: Future AI will need to handle not only text but also visual, tactile, and sensor data, dramatically increasing the volume and complexity of data to be processed.

Quality evaluation: Assessing model performance requires large, high‑quality evaluation datasets, and the lack of such data makes model validation akin to alchemy.

He emphasized that data quality and comprehensive data systems are the root solutions to large‑model hallucinations, and that the ability to generate high‑quality data will become the decisive factor for digital enterprises.

The keynote also highlighted the strategic role of infrastructure providers: building scalable service architectures, reducing compute costs, and pushing performance limits are essential to capture the upcoming wave of exponential long‑tail AI applications.

He announced Ant Group’s commitment to support OceanBase in key AI scenarios across finance, healthcare, and lifestyle, promoting the “Data × AI” concept, open‑source collaboration, and the vision that all data companies will eventually become AI companies.