Path to the Scientific Intelligence 2.0 Era: Xinghe Qizhi Science and Intelligence Open Platform Launching Soon}

In the era of "Science and Intelligence 1.0," a few scientists with advanced AI engineering skills achieved breakthroughs. Now, in "Science and Intelligence 2.0," there is an urgent need for more powerful, user-friendly infrastructure and collaboration platforms, enabling a large community of scientists to focus on core scientific questions and truly become explorers.

Against this backdrop, the Xinghe Qizhi Science and Intelligence Open Platform (NovaInspire: Scientist-Centered AI Open Platform) was born. Developed jointly by the Shanghai Institute of Scientific and Intelligent Innovation, Infinity Light Year, and Fudan University, it is a full-stack, scientist-centered scientific intelligence platform designed to provide high-value scientific data, open-source models, efficient computing, closed-loop experiments, multi-agent reasoning and planning, and interdisciplinary collaboration, aiming to create the "strongest brain" for the Science and Intelligence 2.0 era.

The platform will be officially launched at the WAIC 2025 (World Artificial Intelligence Conference 2025) in Shanghai on July 26.

Six Core Capabilities: Covering the Entire Research Process, Empowering Scientific AI

The platform builds a scientific intelligence ecosystem around six core capabilities, driving paradigm shifts in research:

1. Autonomous Scientific Exploration Engine: Through multi-agent collaboration of data, models, and experiments, it enables automated scientific reasoning, dynamic decision-making, and experimental closed-loop, improving discovery efficiency from data to insights. For example, AI Einstein handles theoretical thinking, AI Fermi manages experimental sensitivity, and AI Gaussian handles mathematical expression, coordinating across agents.
2. Universal Repository of Scientific Models: An open platform integrating multimodal and multidisciplinary models, with over 200 models from 12 fields and 40+ institutions, ready to use, lowering barriers for researchers to build AI-driven complex research chains.
3. Accelerated Scientific Computing Platform: A high-stability, high-efficiency computing platform with fault tolerance, dynamic deployment, and GPU/CPU fusion, reducing task time from hours to minutes (99.7% of tasks), doubling the number of models deployed with the same resources, and supporting uninterrupted long-term training.
4. Closed-Loop Wet and Dry Experiments: Integrates AI simulation (dry experiments) with physical experiments (wet experiments), enabling validation of models in real-world scenarios. It includes the world’s first multi-scale deep phenotyping platform, autonomous intelligent lab agents, and robotic experimenters, supporting full research workflows.
5. High-Value Trusted Scientific Data: A blockchain-based data ecosystem with over 40,000 datasets, totaling 12PB, with daily collection of 150TB and processing of 50TB, ensuring trustworthy, high-quality data for scientific AI.
6. Interdisciplinary Collaborative Community: Facilitates major scientific questions through matchmaking and incentive systems, bringing together scientists, AI developers, engineers, data engineers, algorithm engineers, and experimentalists to co-create and share innovations globally.

Three Major Features: Scientist-Centered, User-Friendly, Open Collaboration

With a solid technical foundation and collaborative network, the platform’s core focus is on scientists. Its features include:

• Scientist-Centered: AI serves scientists, not the other way around. Researchers focus on asking the right questions, while AI helps find answers.
• Easy to Use: No complex deployment required. Intuitive, ready-to-use platform that makes scientific research accessible to everyone.
• Open Collaboration: Promotes open-source, cross-institutional cooperation, building a global network of scientific AI innovation.

These features are rooted in six core capabilities, enabling scientists to naturally harness powerful intelligent agents, truly realizing "Science^AI"—the AI of science—effectively unleashing AI productivity, accelerating discovery, and creating the "strongest brain" for the Science and Intelligence 2.0 era. On this stage, the star remains the scientist.

Scientist-centered is not just a philosophical stance but a core marker of the new era. In Science and Intelligence 1.0, AI was like a high branch shear, used unidirectionally by researchers to harvest mature "scientific fruits." In the 2.0 era, AI becomes a full-process research partner, accompanying scientists from inspiration to validation, helping them directly address the profound questions in the universe.

As we move into the new epoch of scientific intelligence, Xinghe Qizhi invites global scientists, AI engineers, and developers to participate and explore together. The detailed platform features and initial projects will be unveiled at the conference on July 26. Stay tuned for updates from the official WeChat account.

Original link: https://mp.weixin.qq.com/s/yG1plNDSm9KmvN7xCie_vg