Fabarta Announces a Major Upgrade to Its AI for Science Platform: RSI Engine Connects Materials Discovery, Digital Pilot and Volume Production

BEIJING, Nov. 3, 2026 /PRNewswire/ -- Fabarta (Beijing Farbarta Technology Co., Ltd.), a leading provider of enterprise-grade AI agents powered by its knowledge engine, today announced a major strategic upgrade to its AI for Science (AI4S) platform, introducing Fabarta Science Intelligent — a core engine driven by Recursive Self-Improvement (RSI) — and elevating the platform to AI for Science & Engineering (AI4S&E), which covers the entire chain from research to volume production.

The upgrade deeply integrates the company's self-developed Context Graph (enterprise knowledge graph) with physics-informed AI hybrid modeling, closing an end-to-end loop that spans literature research, molecular and formulation design, Wet & Dry Lab validation, digital pilot and scale-up, and volume production. Through three nested loops — the P0 Validation Loop (precise hypothesis validation with real-time model evolution), the P1 Domain Loop (cross-disciplinary deep exploration and multiscale transfer) and the P2 Industrial Loop (from laboratory to production line, a digital-twin closed loop) — the platform addresses two long-standing pain points in new materials R&D, namely balancing multiple target properties and bridging the data gap between R&D and production, while eliminating LLM hallucinations and making every computation traceable and verifiable.

Fabarta's AI4S&E solutions are already deployed at a number of leading central state-owned enterprises (SOEs), national-level chemical industrial parks and leading new materials companies. Going forward, Fabarta will build on its dual-engine model of knowledge engine and large language models (LLMs) to accelerate expansion into the Greater Bay Area and overseas markets such as the Middle East, delivering highly reliable, enterprise-grade AI infrastructure to chemical and energy, new materials and advanced manufacturing companies worldwide.