Saido Technology launched the automotive brand AIVA, backed by state-owned capital, Seres, CATL, and Volcano Engine. What kind of AI car did they create together?
In 2026, as the automotive industry collectively pivots to AI, the birth of a new brand is often accompanied by dense parameter releases and technological visions. On June 9, Saido Technology unveiled the AI pioneer ecosystem mobility brand AIVA in Beijing. The brand's approach to car manufacturing is "AI defines the car, first there is AI, then there is the car." AIVA also announced a deep cooperation with Volcano Engine.
According to reporters, Saido Technology was jointly established with state-owned capital, Seres, CATL, and other industrial investors. Alongside the brand launch, the first concept car, the Origin Concept, was also unveiled. The production model AIVA ME7 is planned to debut within the year, targeting the mainstream market above 200,000 yuan. Volcano Engine's role is defined as "jointly defining, jointly designing, and co-creating the AI car experience," providing intelligent cockpit and other technical services for AIVA.
Putting AI at the Starting Point of Product Definition
AIVA stands for Artificial Intelligence Voyage Ahead, meaning AI accompanies users forward. Li Bo, President and Product Manager of AIVA, told China Times that traditional car product definition relies on market research, scenario deduction, and human experience judgment. The key change brought by AI is that needs can be proactively and batch-sorted. Li Bo used an analogy: in the past, people mined ore; now AI mines ore, and people shift to panning for gold. The product team's role thus transforms from demand discoverer to demand screener, judging which directions are truly worth pursuing and feasible among the many possibilities explored by AI.
To support this product definition logic, AIVA built an internal organizational system. Li Bo stated that the perception layer gathers user feedback, market changes, supply chain dynamics, and other information in real time, trying to solve the information loss in traditional hierarchical transmission. The decision-making layer uses AI sandbox simulation, shifting decision-making from "I think" to "data tells me," but final decisions remain with humans. The execution layer consists of AI assistants and intelligent agents forming an execution network, with each execution result and user feedback automatically flowing back to the perception layer, forming a continuous learning loop. Li Bo described this mechanism as a "flywheel that gets smarter with every action," with a clear goal: to let the organization building the car and the car itself evolve synchronously.
Interestingly, AIVA's brand logo is also AI-related. The central ring symbolizes the AI large model as the "brain," with an internal form of a person embracing the world under the guardianship of an AI lifeform. Officially, the three layers are interpreted as AI large model brain, embodied intelligence, and full-ecosystem lifeform.
At the brand launch, the concept car AIVA Origin Concept was also showcased. The body features G4-level surface treatment with no hard edges, and the headlight area is designed as an interactive "Light Eye" system that responds to user proximity and gestures. AIVA stated that this design aims to bring AI from digital space to the physical world, turning it into a perceptive, responsive embodied AI lifeform. The production car AIVA ME7 is expected to be unveiled within 2026, with the entire lineup entering the market above 200,000 yuan.
In terms of user experience, AIVA advocates "Live Alive, Love Sets Free." Li Bo explained with specific scenarios: setting the air conditioner to 22°C, the system should distinguish between summer T-shirts, winter sweaters after taking off a down jacket, and post-exercise sweating, providing matching temperature sensations. When traveling, the system should detect the user's urgency to recover time when leaving 5 minutes late, automatically open the surround-view camera when passing underground garage blind spots, tighten following distance to prevent queue-cutting in traffic, then automatically widen on highways. In interaction, it can chat during solo commutes, switch to restrained professional mode when carrying elders or leaders, and interact with children during family trips. The underlying logic is unified: transform the car from a passive tool awaiting commands to a partner that understands states and actively contributes.
Yang Liwei, Vice President of Volcano Engine, outlined the technical evolution background of this cooperation. He divided AI capability development into three stages: the chatbot that entered public view in late 2022 was "turn-based," requiring step-by-step user prompts and stalling on complex tasks; agentic AI was proposed in 2024 and rapidly expanded in 2025, with enhanced model reasoning and tool integration, gaining the ability to autonomously follow up tasks until completion; 2026 is seen as the first year of physical AI, where AI is embedded in machines with sensors and actuators, capable of perceiving, reasoning, planning, acting, and reflecting in the real world. Yang noted that the car is currently one of the world's largest, most commercially mature, and most complete industrial chain applications for physical AI.
Based on this judgment, Yang Liwei broke down the changes in physical AI era cars into three directions. Interaction shifts from "mechanical and rigid" to "universal and vibrant," where users no longer need to adapt to the machine, allowing natural communication across ages and language backgrounds. Intelligence moves from "function stacking" to "capability emergence," replacing fixed function lists with flexible task-completing abilities, where the system understands context and uses tools, no longer bound by preset boundaries. Experience evolves from "monotonous and tedious" to "relaxed and joyful," as AI handles repetitive tasks, freeing users' attention and time. He emphasized that the cooperation between Volcano Engine and AIVA is not just stuffing a voice assistant into the car, but deep co-creation around "AI defines the car," making AI the "core power that understands, serves, and accompanies you."
New Logic in the Physical AI Era
The launch of the AIVA brand is actually an experiment in the organizational approach of the automotive industry by Saido Technology and its shareholders.
From the current market perspective, China's new energy vehicle market is no longer a blue ocean. Data from the China Association of Automobile Manufacturers shows that in 2025, the domestic new energy vehicle penetration rate exceeded 50%, with the market shifting from incremental expansion to stock competition. The rankings of the first half of electrification are gradually being determined, and intelligentization has become the core variable of the second half. But intelligentization itself is facing a leap, as industry focus moves from "software-defined vehicles" to "AI-defined vehicles."
Automotive industry observer Zhao Yongqi told China Times that the logic of "software-defined vehicles" uses software to call vehicle hardware capabilities and achieve function iteration via OTA. This paradigm has been widely accepted by the industry in recent years. "AI-defined vehicles" goes further, requiring AI to be the architectural origin from the product definition phase, rather than superimposing AI function modules on an already defined car.
Li Bo further explained that traditional approaches are limited by human cognitive boundaries, often producing what the team thinks is best, not necessarily what users truly need. The logic of AI-defined vehicles is to let the car "grow out of AI's understanding of people." Yang Liwei added from a technical perspective that if a car is defined around AI from day one, its interaction mode, intelligence ceiling, and user experience will fundamentally change. Moreover, with AI iterating rapidly, the car will no longer be a "once defined, long unchanged" product, but a continuously evolving intelligent partner.
So, can the traditional chain structure of the automotive industry adapt to this logic? OEMs sit in the middle, with Tier 1 and Tier 2 suppliers providing components step by step. Product definition is based on existing technical capabilities and market demand, with development cycles of three to five years. But large model iteration cycles are measured in months, and the chain structure's response rhythm begins to strain.
To this end, the shareholder composition and cooperation model of Saido Technology offer another organizational template. The company, jointly funded by state-owned capital, Seres, CATL, and multiple industrial investors, operates as an independent entity responsible for brand, product, and business advancement.
Seres contributes vehicle manufacturing, supply chain, quality systems, and engineering experience. CATL provides battery technology, battery safety, and charging/swapping systems. Volcano Engine offers intelligent cockpit technical capabilities. State-owned platforms provide industrial resource support. Each party contributes its longest strengths, collaborating around the same goal, no longer the traditional one-way relationship of OEM ordering and supplier receiving orders.
The emergence of this organizational model is inseparable from industry chain maturity. By 2026, core technologies such as batteries, motors, and electronic controls in domestic new energy vehicles have moved toward standardization and platformization. The differentiation space in vehicle manufacturing is narrowing, and the true differentiation focus increasingly leans toward AI experience. When manufacturing capability shifts from a unique barrier to a deployable industrial resource, the core proposition of industrial organization naturally evolves from "who can build a car" to "who can more effectively integrate AI and the car."
In Zhao Yongqi's view, the launch of the AIVA brand is an important probe by the industry before the full arrival of the physical AI era. Its model of "AI front-loading, independent entity, longboard synergy" reconfigures vehicle manufacturing, energy systems, supply chain capabilities, and AI technical capabilities under the same goal. Whether this approach works ultimately depends on the actual performance of the production model AIVA ME7 and whether the multi-party coordination mechanism can operate efficiently and sustainably.
In the current market above 200,000 yuan for new energy vehicles, where strong competitors surround and the top echelon is initially forming, Saido Technology and AIVA's step provides a sample worth tracking for the industry to observe the development path of AI cars.
(This article is sourced from China Times)
