Mapping the Cosmos of Innovation: AI Model Charts the Age and Trajectory of Over 23,000 Technologies
University of Technology Sydney researchers unveil Cosmos 1.0, a groundbreaking AI-powered framework that maps and analyzes emerging technologies, providing insights into their maturity, impact, and scientific roots.
Sydney, Australia — December 2, 2025
A team led by the University of Technology Sydney (UTS) has developed one of the most comprehensive and open-source maps of the technology landscape to date. This AI-driven framework, called Cosmos 1.0, applies machine learning to analyze millions of Wikipedia pages, books, and patents, organizing over 23,000 distinct technologies into an intricate, multidimensional map. By doing so, Cosmos 1.0 enables governments, businesses, investors, and researchers worldwide to visualize the vast cosmos of innovation and better understand the evolution and significance of individual technologies.
Published recently in the journal Scientific Data, the Cosmos 1.0 framework introduces a novel approach to tracking and categorizing technology fields. It combines large-scale data analysis with machine learning to provide a detailed picture of how technologies relate, cluster, and develop over time.
A Multilevel Map of Innovation
Cosmos 1.0 allows users to start from broad fields such as artificial intelligence (AI) or quantum computing and then drill down into subfields and specific technologies. For example, within AI, the map distinguishes between different strands such as deep learning, transfer learning, computer vision, and reinforcement learning, and shows their relative maturity and public visibility. Similarly, quantum technologies are broken down into post-quantum cryptography, quantum sensing, and various qubit designs.
“The depth and granularity of this tool enable a nuanced understanding of technological ecosystems," explains lead author Xian (Elaine) Gong, a doctoral candidate at UTS. “You can explore each technology’s trajectory and assess how deeply it is embedded in scientific research.”
Clusters of Technological Innovation
The research team found that technologies naturally group into seven major clusters:
- Autonomous Systems
- Biotechnology
- Data & Analytics
- Health & Medical
- Nanotechnology
- Networking & Connectivity
- Converging Technologies
Among these, the cluster of “Converging Technologies” sits at the center, where materials science, engineering, and digital systems intersect, fostering hybrid innovations. Notably, many renewable energy and climate-related technologies are located within this central cluster, underscoring their hybrid and interdisciplinary nature.
Claire McFarland, another UTS Ph.D. candidate and co-author, noted, “We observe a dynamic interplay where six peripheral clusters orbit a central, convergent core. This highlights the increasingly interdisciplinary nature of cutting-edge technologies.”
New Indexes for Technology Assessment
Cosmos 1.0 introduces four key indexes designed to quantify critical dimensions of each technology’s profile:
- Age of Tech: Estimates when a technology actually integrates into common public use.
- Awareness: Measures public attention and visibility across platforms.
- Generality: Indicates whether a technology is specialized or widely applicable across multiple fields.
- Deeptech Intensity: Assesses how strongly a technology is anchored in scientific research, distinguishing truly science-driven technologies from those propelled more by marketing or hype.
According to co-author Colin Griffith, “The Deeptech Intensity index is especially valuable as it highlights sophisticated technologies grounded in specialized scientific knowledge — technologies that typically represent strategic national advantages in sectors like clean energy, advanced manufacturing, and healthcare.”
By cross-analyzing these indexes, stakeholders can separate hype from substance, identifying technologies that are both gaining traction and rooted in robust science.
Practical Applications and Open Access
Led by Associate Professor Marian-Andrei Rizoiu, head of UTS’s Behavioral Data Science lab, the research team demonstrated Cosmos 1.0’s versatility. The framework can reconstruct technology rollouts over time in various industrial sectors such as automotive and mining, benchmark national innovation strengths, and recommend adjacent technologies for strategic development.
“We envision Cosmos 1.0 as a foundational dataset for policymakers, investors, and academics," Rizoiu said. “Because it aggregates validated data from patents, venture investments, scientific publications, and public awareness, it enables rigorous, multidimensional analyses.”
The entire Cosmos 1.0 dataset is openly accessible to researchers and analysts to integrate into their own decision-support models, dashboards, and innovation strategies.
Looking Ahead
As technology advances at a rapid pace, tools like Cosmos 1.0 are vital to keep pace with the shifting terrain of innovation. By providing a transparent, data-driven map of the technology universe, Cosmos 1.0 empowers diverse stakeholders to make more informed decisions about research investment, policy directions, and commercial opportunities.
For further information and access to the dataset:
Xian Gong et al., Cosmos 1.0: a multidimensional map of the emerging technology frontier, Scientific Data (2025). DOI: 10.1038/s41597-025-06125-y
University of Technology Sydney Behavioral Data Science Lab: UTS Behavioral Data Science
This article is based on a peer-reviewed publication and has been fact-checked and edited for clarity and accuracy.
