AI has driven the demand for Spatial Intelligence among various machines. As of November 2025, SpatiX's monthly spatial intelligence service calls have exceeded 1,000,000,000,000 for two consecutive months, making it the world's first spatial intelligence service platform to reach this milestone.
What does 1 trillion calls mean? If we average these 1 trillion calls per second, it means we have to handle 380,000 spatial service requests every second. In other words, in the blink of an eye, tens of thousands of smart devices such as intelligent cars, mobile phones, drones, and robots complete spatial-temporal perception through SpatiX's services.
Benefiting from the rapid development of large AI models, these smart devices have acquired thinking and decision-making capabilities, with significantly enhanced interactivity with the real world, leading to a substantial increase in the use of spatial intelligence services.
"The AI that works for us needs to understand the world, perceive the world, and see the world clearly. The integration of AI and spatial intelligence gives rise to intelligent agents that can truly work in the physical world," said Chen Jinpei, CEO of SpatiX. He believes that spatial intelligence is the infrastructure that enables AI to truly "land in the physical world," providing AI with high-precision spatial-temporal perception, planning, decision-making, and control capabilities regarding "where, when, and how to move."
Physical AI has become the next wave in global AI development, and spatial intelligence serves as the spatial infrastructure for "Physical AI." It provides the necessary "nervous system + perception map + physical common sense" for AI to act in the real world, allowing various AI agents to not only know their own positions but also perceive the positions of other agents. Currently, scientists both domestically and internationally are researching world models to enable AI to gradually learn and cognize the common sense of the physical world through observation and experience, just like humans.
Spatial-temporal service calls surge 10 times in three years
From 100 billion calls in March 2022 to over 1 trillion calls now, SpatiX's monthly spatial-temporal service calls have increased 10 times in three and a half years!
The spatial intelligence provided by SpatiX is based on a unified, high-precision spatial-temporal (time + space) benchmark, integrating multi-source perception data and using AI algorithms for modeling, deduction, prediction, and decision support. This transcends the limitations of traditional 2D vision, endowing machines with in-depth spatial cognition, enabling them to perceive, plan, and make decisions in complex 3D worlds like humans.
In the field of intelligent driving, the demand for spatial-temporal services in high-level intelligent driving continues to rise. Qianxun's spatial intelligence serves over 3.5 million intelligent connected vehicles, providing spatial-temporal services with an integrity risk as low as 10⁻⁷ per hour (once in a millennium), helping them achieve high-level intelligent driving functions such as high-speed NOA, covering mainstream brands like XPeng, AION, Li Auto, IM Motors, AITO, Leapmotor, Geely, SAIC, and FAW Hongqi. Currently, SpatiX serves over 10 billion kilometers of intelligent driving mileage monthly. With the large-scale implementation of L2-level assisted driving and the accelerated approval of L3-level, BeiDou spatial-temporal services have become an indispensable part of intelligent driving systems, showing broad market prospects.
In the low-altitude economy sector, SpatiX's spatial-temporal infrastructure accurately maps real-world low-altitude scenarios in virtual space and supports air route planning, ensuring precise low-altitude flights every time. Currently, SpatiX has helped over 200,000 low-altitude aircraft such as drones complete professional tasks including power grid inspection, water area management, agricultural plant protection, and night sky performances. The cumulative low-altitude service mileage is nearly 4 billion kilometers, equivalent to escorting these aircraft around the Earth nearly 100,000 times.
In the consumer sector, SpatiX and its partners have launched innovative applications such as "lane-level navigation" and "traffic light countdown," upgrading mobile phone positioning accuracy from "road-level" to "lane-level." Currently, the shipment of mobile phones supporting SpatiX's spatial-temporal capabilities has exceeded 50 million units, covering mainstream brands like Huawei, Honor, OPPO, and VIVO, bringing revolutionary changes to users' navigation experience.
Chen Jinpei pointed out that today is just the prelude to the era of universal intelligence. As the industry moves from "Internet of Things" to "Intelligent Connection of Things," the number of intelligent cars, robots, and low-altitude aircraft will grow exponentially, and their demand for high-precision spatial-temporal services will shift from "optional" to "essential." Spatial intelligence is transitioning from technology supply to demand-driven, becoming a fundamental element driving industrial innovation.
Spatial intelligence enables AI to truly "land in the physical world"
As AI technology continues to penetrate autonomous machine systems such as robots and self-driving vehicles, the demand for physical interaction capabilities in the real world has become increasingly prominent. Overnight, from Shanghai to Silicon Valley, "Physical AI" has become the latest hotspot in AI development.
"Nature Machine Intelligence" once defined "Physical AI" as: Physical AI refers to physical systems capable of performing tasks usually associated with intelligent organisms, enabling the coordinated evolution of the body, control, form, action execution, and perception.
International Data Corporation (IDC) believes that the core value of "Physical AI" lies in endowing autonomous machines with the ability to achieve a "perception-understanding-execution" closed loop in the real physical world, serving as a key bridge for artificial intelligence to evolve from virtual intelligence to embodied intelligence.
In the eyes of NVIDIA founder Jensen Huang, robots, cars, etc., all belong to the category of Physical AI, and the next stage of AI is Physical AI. "Digital twin factories, warehouses, operating rooms, etc. — AI is learning in real-time how to understand and manipulate the real world."
Recently, World Labs, founded by Stanford University professor Fei-Fei Li, launched its first commercial product, Marble, a generative AI-driven 3D world generation system, allowing users to create spatially coherent, high-fidelity, and sustainable 3D worlds with just a picture, a video clip, or a text prompt. "The current direction of AI focusing on parameters and algorithms may be off track. If models do not understand the world, true intelligence cannot be achieved," Fei-Fei Li said in a long article.
Chen Jinpei, CEO of SpatiX, believes that the core challenge for AI in understanding the physical world, reasoning, and executing tasks is converting digitally driven instructions into precise operations in the physical world. He proposed "spatial intelligence" in 2019 — it is an indispensable bridge connecting the physical and digital worlds, helping various devices in different scenarios and with different needs to conduct spatial-temporal planning, decision-making, and control.
At the perception layer, SpatiX integrates multiple advanced technologies such as BeiDou's spatial-temporal intelligence basic capabilities, vision systems, and inertial navigation systems, providing machines with an action hub integrating spatial-temporal perception and edge computing power. It endows machines with accurate, reliable, and secure spatial-temporal information and multi-model collaborative reasoning capabilities, enabling model reasoning at the edge, giving robots a human-like sense of direction, and allowing them to complete diverse tasks in complex environments.
At the planning layer, SpatiX provides machines with high-precision environmental perception and map prior information, helping machines reasonably plan paths and act autonomously, avoid obstacles, and reach target positions smoothly.
Since its establishment in 2015, SpatiX has been committed to building precise spatial-temporal capabilities into a large-scale public service, making spatial intelligence services as accessible as water, electricity, coal, the Internet, and cloud computing. Currently, SpatiX utilizes over 10,000 GNSS space-based/ground-based augmentation stations worldwide, independently developed intelligent algorithms, chips, and large-scale Internet service platforms to provide machines with centimeter-level positioning, millimeter-level perception, and nanosecond-level timing spatial intelligence services.
The cornerstone supporting this transformation is SpatiX's genes from cloud-native architecture to AI-driven intelligence, enabling it to rapidly expand in the era of the Internet of Everything and large models, meeting the explosive growth in demand for spatial-temporal services from various AI and digital applications. Currently, SpatiX has the capability to handle millions of concurrent requests per second, coping with the pressure of high concurrency and supporting the simultaneous access of massive smart devices.
