On April 15, 2026, the STAC Summit Hong Kong 2026, a premier industry forum in the global fintech sector, was held at the HKEX Connect Hall. DolphinDB was invited to attend, and its Founder and CEO, Dr. Davis Zhou, delivered a keynote speech at the AI & Analytics track, sharing DolphinDB's latest technical practices and cutting-edge explorations in high-performance quantitative data infrastructure.
The summit brought together technical leaders and decision-makers from world-class fintech institutions including HKEX, Jefferies, Optiver, AWS, and AMD. The event focused on financial technology trends in the Asia-Pacific market, facilitating in-depth discussions on AI-era trading system architecture, high-frequency computing, and market structure optimization.
Keynot: Simplifying the Quantitative Foundation in the AI Era
At the AI & Analytics track, Dr. Davis Zhou delivered an insightful keynote titled "AI-Ready High Performance Quant Stack Without the Complexity." He pointed out that the rapid adoption of emerging paradigms — AI Agents and MCP among them — is introducing unprecedented complexity into quantitative infrastructure. DolphinDB's mission is to tame this complexity—making real-time data systems simpler and faster, and better serving both humans and intelligent agents.
Highlights from the Speech:
The presentation focused on four core technical directions:
1. Full-Stack Integrated Infrastructure
DolphinDB natively integrates storage, analytics, and stream computing capabilities on a single platform, offering a multi-model storage engine that supports TSDB, OLAP, vector, text, and other data models. On the programming front, it is compatible with both SQL and a Python-like scripting language, with over 2,000 built-in professional functions and more than 20 ready-to-use stream computing engines, easily supporting millisecond-level queries on trillions of rows of data.
2. Unified Framework for Research and Production
Quant teams have long faced the challenge of maintaining dual codebases—Python for research and C++ for production. DolphinDB completely solves this problem through a unified scripting framework: the same factor logic can be deployed directly from backtesting environment to real-time trading without rewriting, ensuring 100% logical consistency. Leveraging DolphinDB's unified stream-batch capability, teams can shorten factor go-live cycles from one month to just three days.
3. Microsecond-Level Low-Latency Stream Computing
DolphinDB has introduced a low-latency stream computing engine optimized for high-performance scenarios, achieved through a combination of low-level optimizations: incremental computation, a row-based execution engine, CPU core affinity, and in-process embedding. Benchmarked on an Intel i9-14900KS platform, end-to-end average latency for processing 10 factors — triggered by individual ticks across three cascaded stream computing engines — stands at just 2.7 microseconds, scaling to approximately 30 microseconds for 100 factors, with virtually no degradation under multithreaded concurrency.
4. Quantitative Intelligence Foundation for the AI Era
New paradigms such as MCP, Skills, and Agents are currently reshaping workflows in the financial industry. In his speech, Dr. Davis Zhou noted that DolphinDB is building a comprehensive enablement layer around AI workflows—launching an MCP Server and a professional Skills toolkit tailored for quantitative finance scenarios. These support AI Agents in directly calling core capabilities such as data queries, factor calculations, and real-time market data subscriptions, while the built-in vector storage engine enables AI-enhanced analytics scenarios including semantic search and similar market pattern recognition.
On the Exhibition Floor
At the exhibition area, the DolphinDB team engaged in in-depth conversations with technical heads and quantitative engineers from top global financial institutions. Through video demos, they showcased core practical scenarios including financial time-series data processing, high-frequency factor calculation, real-time risk control, and the integration of AI Agents with quantitative foundations.
Looking Ahead
STAC Summit Hong Kong 2026 underscored DolphinDB's growing presence on the world stage. At its core, Dr. Davis Zhou's keynote carried one clear message: in the AI era, infrastructure should be an accelerator — not a bottleneck.
Backed by unified stream-batch processing, seamless research-to-production deployment, and microsecond-level latency, DolphinDB is purpose-built to serve as the high-performance, AI-ready foundation that quantitative financial institutions worldwide demand.