How Ant Group’s OceanBase Database Emerged from Double‑11 Pressure: Lessons in Autonomous R&D

The ZTE incident turned the topic of independent R&D into a social hot‑spot and revived Alibaba’s "de‑IOE" movement launched five years earlier; the massive transaction volume driven by Double‑11 forced a technical choice that became a classic case of Chinese internet companies embarking on self‑development.

On May 11, at an FT Chinese forum titled “What Kind of Independent R&D Does China Need?”, Ant Financial Vice President Hu Xi shared this story and his reflections on autonomous development.

Hu Xi, an 80‑born Alibaba technologist and the youngest partner in the company’s history, joined Alibaba in 2007, led the three‑generation architecture of Alipay, and witnessed the full cycle from business‑driven self‑development to self‑development driving business innovation.

Self‑R&D Forced by Double‑11

Alibaba and Ant Financial’s self‑built database, OceanBase (literally “ocean‑like database”), set several records: it handled a peak of 42 million transactions per second and reduced hardware costs to one‑tenth of a comparable centralized database while maintaining performance, positioning China among the world’s leading database powers.

According to Hu Xi, the decision was a matter of life and death: Double‑11’s payment peak grew 1 280‑fold in just eight years. After a near‑catastrophic outage in the final minute of the 2010 Double‑11, the team realized that only a home‑grown database could meet their needs.

Before the technology matured, the database team endured five to six years of isolation and uncertainty. Their leader had previously spent three years on a database project at another internet company that never materialized, leading to the team’s dissolution. Alibaba later recruited the leader and several core members, who persisted for four years before seeing clear progress.

“If you don’t have a core of technical idealism, you can’t achieve autonomous R&D,” Hu Xi said.

In recent years, the centralized IOE architecture has become increasingly unsuitable for China’s mobile‑internet landscape, and distributed architectures have become mainstream among Chinese internet firms.

Alibaba and Ant are among the few companies that insist on self‑development, yet Hu Xi stresses that self‑development is not inherently superior to open‑source or commercial solutions; the key is solving real business problems and advancing technology where needed. Ant’s choice to self‑develop stems from a desire to avoid being constrained by external frameworks.

Navigating the ‘No‑Man’s Land’: Self‑R&D as the Only Way

The database forced out of the market was only the beginning of Ant’s self‑development journey. As a benchmark fintech company, Ant has delivered several milestone innovations such as guarantee transactions, quick payments, reverse QR codes, facial recognition payments, Yu‑e‑Bao, and the “3‑minute application, 1‑second disbursement, zero‑human‑intervention” 310 loan model.

The Economist highlighted last year that Chinese fintech has outpaced the United States.

Hu Xi noted that behind these innovations is Ant’s shift from “forced self‑development” to proactive, comprehensive autonomous development of five foundational technologies: blockchain, artificial intelligence, security risk control, Internet of Things, and computing.

▲These foundational technologies are all self‑developed by Ant

Such heavy investment is necessary because foreign technology stacks do not support the development of Chinese fintech; self‑development is the only viable path. The broader and deeper the self‑developed foundation, the greater the strategic control, allowing Ant to feed innovation back into the entire industry chain.

Self‑R&D Is Not Meant to Be Hoarded, but to Be Open

Openness marks Ant’s R&D 3.0 stage. In the past two years, Ant has opened AI, risk‑control, and IoT technologies, and in April 2018 released the distributed middleware SOFA as open‑source. Ant also shares core risk‑control capabilities with nine Belt‑and‑Road countries and regions.

Hu Xi explained that while self‑development has limits, openness is the way to break those limits, ultimately allowing ordinary people to enjoy the benefits of technology.

He concluded with three reflections on autonomous R&D:

1. Commercial value is essential. If Alipay had tried to build a database at its inception, it would have failed due to lack of sustainable value and feedback loops.

2. Technical idealism is required. Without a deep‑seated belief in technology, it would be impossible to spend five, ten, or more years on cloud computing and databases, and to attract top talent from abroad to relocate to Hangzhou.

3. A sense of national responsibility is crucial. The fourth industrial revolution (digital economy) is arriving faster than previous revolutions; seizing this opportunity demands both corporate and patriotic commitment. Autonomous controllability is ultimately about national, not just corporate, sovereignty.