Terminus Group Spatial Intelligence Solution Lands in iFlytek AI Headquarters

In the advancement of spatial intelligence, understanding and aligning multi-source heterogeneous data within spaces remains a common industry challenge. To address this, Terminus Group has introduced its Meta-Space Element Management Platform, equipped with multi-layered structuring capabilities to overcome the difficulties of managing unstructured spatial elements. The Terminus Group Meta-Space platform and its full-stack solutions have now been deployed at the iFLYTEK AI Town project, creating a scalable paradigm for spatial intelligence implementation.

Meta-Space: Conquering Unstructured Spatial Governance with Triple Structuring Capabilities

Unstructured data refers to information that lacks a predefined model or fixed format. This type of data is diverse in content and form, typically existing in its natural state. It requires technologies such as NLP, CV, and OCR to be transformed into structured formats before its value can be fully extracted.

From geospatial data in the physical world (such as satellite imagery and GIS coordinates) to environmental and scientific datasets (such as remote sensing images or geological surveys), as well as everyday text (emails, documents, comments), multimedia (images, video, audio), and IoT data—all are typically unstructured. According to IDC, global data volume will rise from 16ZB in 2016 to 163ZB in 2025, with 80% being unstructured or semi-structured. In the era of intelligence, rapidly converting unstructured data into structured formats through innovative solutions is pivotal to accelerating spatial intelligence and enabling universal computation. The Meta-Space solution was developed in this context.

As the smallest management unit of the physical world, Meta-Space is designed to connect complex business scenarios in physical environments, orchestrate IoT devices, enable real-time spatial strategy generation, and rapidly build digital twins—enhancing the bidirectional interaction between the physical and digital worlds.

Meta-Space delivers three core structuring capabilities:

· IoT Data Structuring (Device Control Spatialization): Aggregates discrete device data (lighting, HVAC, access control, etc.) into Meta-Space units for unified parsing.

· Rule Data Structuring (Business Strategy Knowledge Encoding): Converts expert experience (e.g., energy-saving protocols, security processes) into an AI strategy library of 127 iterative strategies, enabling data-driven decision-making and automated execution.

· Environmental Data Structuring (Physical Space Digitalization): Converts physical entities such as buildings and equipment into computable digital twin models with standardized spatial coordinates.

This enables the transformation of physical space from unstructured complexity to computable systems.

Space-as-a-Service: Terminus Group Spatial Intelligence Solutions Landed in iFLYTEK AI Headquarters

Powered by Meta-Space, physical space becomes a service entity, creating a brand new business model—Space-as-a-Service (SaaS). The Meta-Space platform-based solution has now been deployed in iFLYTEK’s AI Headquarters.

iFLYTEK AI Headquarters accommodates over 15,000 employees, offering more than 5,000 parking spaces. Located in Hefei, China, the project aims to create a hub for technological innovation and emerging industries, delivering safe, green, and efficient work and business experiences for its occupants.

Large-scale integrated industrial parks like iFLYTEK AI Headquarters often face challenges such as fragmented security management, rough energy consumption control, inefficient space utilization, and delayed operation response. Terminus Group designed a Meta-Space platform-based solution to address these pain points across various spaces such as office, dining, reception, meeting, wellness, operational, etc..

· Security Coordination: Through 3D digital twin models, fragmented spatial data becomes interoperable. The strategy library generates coordinated logic by spatial unit, solving the challenge of isolated security systems.

· Precision Energy Management: Physical space is broken down into the smallest controllable structured units. IoT hardware integration enables energy management to evolve from whole-building estimates to square-meter-level precision control.

· Optimized Space Utilization: Discrete spatial elements (meeting rooms, equipment, occupancy schedules, etc.) are structured and integrated into automated space workflows, replacing manual scheduling and improving efficiency.

· Proactive Operations: By binding spatial and device twins, each physical device (e.g., elevators, water pipes, meters) has a structured digital mirror for precise fault mapping, reducing maintenance delays.

Additionally, the Meta-Space platform integrates IoT device control, spatial strategy execution, and digital twin scene creation, enabling clients to achieve low-cost, customized, full-scale smart space management.

According to data from January 2025, iFLYTEK AI Headquarters achieved a 5.7% reduction in total electricity usage and significantly improved security management efficiency through Meta-Space implementation.

Unlike humans, who intuitively understand spatial relationships (e.g. reconstructing 3D space instantly from a 2D video), machines require this capability as an acquired skill. Training spatial models remains nascent globally, hindered by the unstructured nature of spatial data.

Terminus Group experts view Meta-Space as an attempt to conquer the structured representation of unstructured data and a pathway for implementing their "Model + System" approach. Moving forward, Terminus Group will further explore the integrated training and inference strategies for unstructured data, accelerating the integration of direct and indirect intelligent computing to enable the scaled application of spatial intelligence across diverse scenarios.