Since the last technical update in June, Gravitio has seen its user base climb from 4,000 to 5,500, with internal records indicating that 3,960 of these accounts are actively engaged. The platform’s activity has scaled alongside this growth, with AI-agent predictions rising 77% and the number of cloned agents reaching 9,077. This expansion coincides with the rollout of the GRV-2.1 inference architecture, which introduces a modular design to separate data acquisition from AI reasoning and outcome verification.

The new infrastructure aims to improve efficiency by reducing duplicate AI workloads, allowing the system to reference centralized intelligence while maintaining individual performance histories. By integrating real-time data from sources like Polymarket alongside traditional market indicators and news sentiment, the platform seeks to refine its predictive outputs through an automated feedback loop. CEO Bobby Moshar noted that the current technical foundation is designed to support a larger user base without requiring a linear increase in computational costs.

While Belgravia Hartford has explored the possibility of spinning out Gravitio as a separate entity for shareholders, management clarified that no formal decision has been reached. Future development will focus on expanding data sources and scaling user acquisition, as the company navigates a micro-cap market that increasingly prioritizes technical maturity and verifiable execution.