Ant Group’s Wang Xu on Generative AI, the Emerging LAMP Paradigm, and Infrastructure Evolution

At the MEET2025 Intelligent Future Conference, Wang Xu, Vice‑Chair of Ant Group’s Open‑Source Technology Committee, presented a data‑driven view of large‑model architectures, infrastructure, and innovative applications.

He asserted that models with generative data capabilities are beginning to replace databases as the core of traditional applications, heralding a new LAMP stack where models are more portable than databases and new data‑transfer and security requirements are emerging.

Wang highlighted the rapid explosion of AI agent frameworks following the release of open‑source models such as Llama 2, noting that while many projects have surged, none has yet dominated the landscape.

He described how Ant Group leverages open‑source community insights to guide its own architectural evolution, using large‑model‑driven analysis to accelerate tasks like financial report review and building AI‑powered applications with frameworks like agentUniverse .

The talk also covered trends in developer tools, emphasizing the shift from simple data‑entry utilities to AI‑augmented collaborative partners, and cited the rise of low‑code, 30‑minute AI app development.

Regarding infrastructure, Wang explained that while hardware (GPU‑centric compute clusters) has dramatically changed, the underlying software stack remains similar, with new trade‑offs driven by generative AI workloads, model serving, and security considerations.

He concluded that the AI era is reshaping both native AI applications and traditional software, urging infrastructure teams to focus on supporting model‑centric workloads, data migration, and secure AI integration.