Ant Group spent a record ~$2.9B on research in 2023, as the Alibaba affiliate develops its BaiLing LLM; Ant and affiliate MYbank served 87M businesses by 2023
Context & Ripple Effects
Ant had already formed a dedicated unit for ChatGPT-style technology and committed about $2.8B to R&D in 2022 through its earlier Zhen Yi LLM effort. The 2023 research total shows that work remained a strategic investment rather than a one-off experiment.
The company is pairing model development with a large existing business footprint: Ant and MYbank served 87M businesses by 2023. That makes BaiLing relevant not only as a general AI project, but as research that could eventually be applied across Ant's business-facing products.
First-order effects
- Ant commits a record ~$2.9B to research while continuing development of its BaiLing LLM, increasing the resources available to its in-house AI program.
- Ant and MYbank's 87M-business reach gives the group a large potential base for testing or deploying business-oriented AI capabilities, subject to product and regulatory decisions not described here.
Second-order effects
- The spending raises the competitive bar for financial-platform AI efforts: rivals without comparable research budgets or business distribution may need to partner, specialize, or spend more to keep pace.
- MYbank and Ant's business relationships become more strategically valuable because they connect AI research to a broad customer base rather than leaving it solely as a standalone model-development effort.
Third-order effects
- If sustained, this points to AI development becoming more concentrated among large platforms that can fund research and connect models to established financial and business-service ecosystems.
- The key structural question is whether proprietary models create differentiated services for existing platforms, or whether model capabilities become broadly available and competition shifts back to distribution and customer trust.
The trend: Financial platforms are increasingly treating proprietary AI research as core infrastructure, combining capital-intensive model development with existing business networks.