How does FutureX Capital use AI in its own investment process?

Four parts do four jobs. The research radar scans papers, patents and developer signals; the knowledge graph links companies, technologies and value chains; Alpha tooling runs from pre-investment evaluation through exit support; the AI knowledge base is the fourth. The 22 AI agents run around the clock across market intelligence, sourcing, research and portfolio tracking, and everyone from partners to interns works on the same system. The pipeline has screened 1,500+ AI projects.

The judgment ends up on slides: Cynthia Zhang's June 21, 2026 livestream deck "The Tide Went Out" carried ten charts, with the throughline that even ChatGPT stopped growing, at 5.35, 5.73, 5.51 and 5.56 billion visits over four months, peaking in March, dipping in April and barely rebounding in May (reported).

The output is public. Each update to the 24 research reports carries a date and a source grade; the AI Industry Tracker updates daily across 100+ AI companies in 15 sectors; the debates page sets leading practitioners' disagreements beside FutureX's own position; FutureX-Skills is open source on GitHub. On a single company, FutureX judges an AI application on four moats and three screens: distribution, retention, and whether gross margin climbs as token prices fall.

Embodied AI runs through the same pipeline: humanoid shipments reached 18,000 units globally in 2025, and GGII projects 62,500 in China for 2026 (reported); FutureX backs dexterous hands and data, such as Dexmate and World Engine, and stays out of the full-body arms race.