A live demonstration of future technologies: The third stop of the Shanghai Artificial Intelligence Industry Special Session of the 4th "Double Carbon Star Species · Carbon Exploration Plan" concluded.

In the fifth year since the proposal of the "dual carbon" goals, discussions about the industrial ecosystem increasingly call for voices from the front line.

On December 23rd, the third stop of the Fourth "Dual Carbon Star Species · Carbon Exploration Plan", jointly initiated by 36Kr, Orient Securities, and its subsidiaries Orient Securities Innovation and Orient Securities Capital, was launched in Shanghai. This is also the final stop of this year's Carbon Exploration Plan.

From the "Future Transportation" in Zhengzhou Airport Economic Zone, to the "Future Energy" in Beijing Daxing, and then to the focus on "Future Technology" in Shanghai, the Carbon Exploration Plan has been advancing deeper into the industry along the practical implementation path of the dual carbon goals.

That's why this event chose to visit the "Model Speed Space", an important carrier of the Shanghai Artificial Intelligence Industry, to explore on - site the real - world samples of AI's transition from technological innovation to systematic implementation.

As the initiators, Orient Securities and 36Kr organized investment institutions and enterprises in the upstream and downstream of the industrial chain to visit the industrial site and conduct systematic discussions on the innovation path of the industry driven by AI and the practical possibilities of low - carbon intelligent manufacturing.

On the day of the event, guests such as Xie Zuoqiang, the vice - president of 36Kr, Zheng Wangyu, the vice - president of 36Kr Fund, Chen Gang, the vice - president of Orient Securities, and Zhang Yun, the deputy general manager of Model Speed Space, attended the event. Meanwhile, roadshows of AI - related projects were also held on site.

"If in the previous two stops, we focused on 'seeing' industrial practices and 'connecting' multiple resources, then today we hope to 'precipitate' practical insights and 'elevate' the understanding of development with all of you," said Chen Gang, the vice - president of Orient Securities. As a participant and promoter in the capital market, Orient Securities has always deeply embedded the concept of green development in the core of its business and actively fulfilled its mission of green finance. In recent years, it has continuously promoted green equity and bond investment and financing services, guided all parties in the market to jointly build a green finance ecosystem, and delved into carbon finance business to help enterprises revitalize carbon assets.

This is also one of the reasons behind Orient Securities' continuous participation in the Dual Carbon Star Species · Carbon Exploration Plan with 36Kr for the fourth consecutive year and the upgrade of its form this year.

A on - site deduction in the real industrial hinterland has begun. Perspectives from different identities converge here to explore how AI will participate in the next stage of low - carbon transformation.

01

For AI to enter the low - carbon narrative, an innovation site at the industrial front line is needed

"The green competition for carbon peak by 2030, which concerns the future, has entered the substantial stress - testing stage," said Xie Zuoqiang, the vice - president of 36Kr on site. "Innovation should not only exist in the PPTs of roadshows but also at the industrial front line." This judgment directly points out the reason why this year's Dual Carbon Star Species · Carbon Exploration Plan keeps moving towards the industrial hinterland.

When the dual carbon goals enter the implementation stage, a more practical problem arises: How can green transformation be truly implemented in a highly mature and closely - structured industrial system?

Manufacturing, the foundation of industrial society, is precisely the most complex and difficult part to be rapidly rewritten. There is no lack of consensus on carbon reduction in the manufacturing industry. The problem is that it is a highly engineered system: production lines, processes, quality, yield, delivery cycle, and cost structure, each variable is interrelated. Any single - point optimization may create new waste elsewhere.

"The achievement of the national dual carbon goals has never been an isolated technological breakthrough, but rather the close symbiosis and collaborative evolution of the industrial chain, innovation chain, and capital chain in real - world scenarios," said Chen Gang, the vice - president of Orient Securities, giving a clearer response from the perspective of capital - industry collaboration.

Against this background, AI has begun to be re - evaluated. On the premise of cost reduction and efficiency improvement, it is gradually being given a more practical expectation: Can technological innovation disassemble, quantify, and optimize complex systems, turning carbon reduction from a one - time project into a continuously operating ability?

The final stop of this Carbon Exploration Plan is at the Model Speed Space in Shanghai. This is an attempt to bring together multiple forces to further explore the integration path of "AI + Green" and jointly discuss how technology can be transformed into the core driving force for green development.

02

Shanghai Model Speed Space: The industrial foundation for low - carbon intelligent manufacturing

As the first large - model innovation ecological community in China, Model Speed Space is different from traditional industrial parks.

In terms of function, it is more like a high - density reaction field of "technology, industry, and capital". With strong policy support and a full - factor platform, Model Speed Space helps large - model enterprises accelerate their development and gradually form an independent, controllable, and internationally competitive industrial cluster.

The five functional platforms of computing power, corpus, finance, talent, and scenarios built by Model Speed Space form a high - frequency collaboration mechanism, providing full - life - cycle support for enterprises from their inception to growth, and promoting the implementation of AI in specific scenarios such as finance, culture and tourism, healthcare, manufacturing, and energy management. This ecological system significantly accelerates technological innovation and result transformation, helps incubate a number of lightweight and easily replicable AI application products, and continuously promotes their iterative upgrading.

Since its establishment, Model Speed Space has attracted more than 200 enterprises to settle in, covering the core links of the large - model industrial chain. The investment rate of incubated enterprises has exceeded 35%, and the overall valuation has increased by more than 10 billion yuan. More importantly, this sustainable industrial ecosystem has been gradually improved, forming a closed - loop industrial ecosystem covering the upstream element layer, the mid - stream model layer, and the downstream vertical application layer, and gradually creating a "flywheel effect" where enterprises drive the ecosystem. "The upstairs and downstairs are the upstream and downstream" and "neighbors are partners" have become the norm in "Model Speed Space".

This change is particularly crucial for the dual carbon agenda. AI technology is transitioning from the "capability demonstration stage" to the "systematic implementation stage". At this stage, whether the technology can be embedded in the existing industrial system and operate stably in the long term has become a new dividing line. This determines whether technological innovation can be replicated and promoted across different factories and production lines.

From this perspective, visiting Model Speed Space is also a judgment of the current progress: How much of the ever - flowing AI technology under the emergence of intelligence has the conditions to become the infrastructure of the green economy?

This has also become the main theme of this on - site roadshow. The roadshow enterprises cover multiple fields such as industrial AI, spatial intelligence, embodied intelligence, intelligent hardware, and safety supervision. Although their technological forms are significantly different, they all have one thing in common: they are ready to enter the real industrial system and take on specific responsibilities in the green transformation.

• As an industrial AI solution provider, Xiwei Technology creates a new - generation industrial Agentic AI platform through its self - developed vertical industrial process data, industrial analysis algorithms, and industrial intelligent agents, helping to upgrade the industry and enabling enterprises to continuously improve in core scenarios such as quality management, process optimization, and data analysis capabilities.
• Founded in 2021 and headquartered in Hangzhou, MizarVision is committed to transforming the traditional way of intelligence application and using innovative geospatial intelligence services to illuminate various industries. MizarVision aims to be the "Bloomberg" in the industry intelligence field, promoting the commercialization of remote sensing in China and the implementation of the open - source intelligence service framework. Its multi - modal sensor fusion model and intelligence mining services have taken root in the top - tier customer system of Chinese intelligence, and MizarVision has also emerged on the international stage, becoming a benchmark brand for Chinese open - source intelligence and remote sensing applications.
• Screaming Intelligence focuses on the research and development of artificial intelligence robot exoskeleton assistance systems, and its products have been applied in scenarios such as firefighting, healthcare, and industry. Although exoskeleton technology does not directly aim at energy substitution, it addresses the long - standing problems of high - intensity and high - risk labor in manufacturing and public services.
• Xinyu Technology focuses on the fields of artificial intelligence and wireless communication, serving digital economy and low - altitude economy application scenarios in the energy and transportation industries around AI synthetic data and full - scenario data application intelligent agents. In the green transformation, the integrity, continuity, and high quality of data often determine the quality of decision - making. Scenarios such as energy dispatch, urban operation, and low - altitude supervision highly depend on the ability to control the "invisible systems". The advantage of Xinyu Technology lies in the modeling and perception of complex environmental data and the construction of intelligent agents for relevant application scenarios.
• Zhishi Shenggong showcases a non - consensus technological route: through the full - stack architecture of "self - developed chips + self - developed models + self - developed intelligent agent systems", it provides an "end - to - end intelligent agent software factory", achieving an end - to - end closed - loop from requirement analysis to software testing and deployment, and directly delivering runnable software systems to enterprises or organizations in need of customized software. This product breaks the high dependence on large models, has low energy consumption, and high efficiency. It covers all professional positions in the software development process, can handle enterprise - level complex engineering software with a code volume of over one million lines in context, and supports isolated deployment. While bringing a new AI - native software development experience to enterprise developers, it provides full - life - cycle encryption protection for code data security.
• Light Cloud focuses on the field of cloud - native graphics technology. With nearly two decades of research accumulation in computer graphics by its founding team, it has created an internationally leading cloud - native real - time rendering engine. Currently, its engine has been successfully applied in many high - tech fields such as cloud gaming, AR/VR, industrial digital twin, and online exhibitions, and will also become a powerful technological foundation for the metaverse. Products and solutions: The fully self - developed cloud - native real - time rendering engine RAYSENGINE is the world's first cross - domain "cloud - edge - end" architecture, realizing a cloud - native commercial engine for real - time rendering. RAYSENGINE uses an innovative solution to effectively overcome the physical barrier problem of current pure cloud streaming rendering, and has received support from the key research and development project of the Ministry of Science and Technology and more than ten national invention patents.
• INFIFORCE, founded in 2023, is an innovative high - tech enterprise driven by AI and led by technology. INFIFORCE has gathered executives from global top 500 enterprises, top scientists, and an elite team of master's and doctoral students. Relying on the global R & D network of "Hangzhou headquarters + Shenzhen Research Institute + Beijing Laboratory", it is always committed to the R & D and commercialization of leading embodied intelligence technology. The core strategy of INFIFORCE is "one brain, multiple bodies, and multiple scenarios". By constructing a leading "embodied brain" - based on the full - stack self - developed Hyper - VLA (Vision - Language - Action) end - to - end multi - modal large model + causal world model, it forms a complete intelligent closed - loop of "perception - cognition - decision - execution", breaking the boundary between the virtual and the real, enabling intelligent agents to truly understand the world, continuously learn, and act autonomously, thus creating a "super brain + the strongest body". Through the growth flywheel of "data - model - body - scenario", it constructs a globally leading embodied intelligence ecosystem.
• Xili Technology takes "IP + AI" as the core model, developing children's AI toys with voice interaction as the main feature and a screen - less design. Through emotional and warm - hearted character companionship, it stimulates children's language ability, imagination, and learning interest. At the same time, it introduces environmentally friendly materials and the concept of sustainable manufacturing, exploring the path of low - carbon design and carbon neutrality, and constructing the next - generation children's product system with educational value, emotional value, and social responsibility.
• Shanghai Zhipaike Robot is a startup focusing on embodied intelligent robots. Zhipaike Robot adheres to the corporate vision of "making wisdom deliverable and life more beautiful", and is committed to creating cost - effective, highly safe, and user - friendly intelligent robot products. It is also making efforts in the development of the large - model - native embodied intelligent robot operating system, developer hardware and software tools, and the ecological platform. Through its unique technological advantages, differentiated product positioning, and innovative business model, it aims to promote the transformation of the embodied intelligent robot industry and the industrial chain, and realize the dream of creating highly personalized and personified robot assistants and companions for everyone.
• Shanghai Yuben Intelligent Technology Co., Ltd., with the mission of "making factories more intelligent", has self - developed two product series, the "YobiAI Industrial Safety Eye" and the "YobiAI Industrial Inspection Eye", including multiple industrial - grade AI standard products such as the intelligent safety management system for personnel in robot, stamping machine, three - dimensional warehouse, and bending machine operation areas, and the intelligent tablet inspection system. These products are applied in the scenarios of intelligent safety management and intelligent defect detection in manufacturing factories, and belong to the development direction of intelligent manufacturing and domestic substitution. The core standard component of the product, the YobiAI EdgeX intelligent super - brain controller, is an industrial - grade AI vision controller, which has obtained the PL d - level safety performance level calculation result issued by SGS and technical certifications from China Academy of Information and Communications Technology, Huawei, etc. The products have been sold to factories such as Huawei's Southern Factory, Emerson, GlaxoSmithKline, Baoti, Goertek, Orion, Pujie Non - woven Fabric, and Nanhai Nansin.

In industrial scenarios, the accuracy, stability, interpretability, and long - term operating cost of AI have become more important indicators than model parameters. Many enterprises repeatedly emphasized in the roadshow that the core of their technological innovation lies in whether it can operate stably.

This change itself is a clear industrial signal.

It means that AI is transitioning from the "capability demonstration stage" to the "responsibility - taking stage". Once the technology starts to take on responsibilities, it is no longer just an efficiency tool but will directly affect energy consumption, yield, and loss, and will truly enter the core of the dual carbon narrative.

03

Re - calibrate the positions of industry and technology in the dual carbon landscape

Looking back at the three - stop route of the Fourth Dual Carbon Star Species · Carbon Exploration Plan, its value does not lie in the number of industrial scenarios covered, but in gradually clarifying a key question: After the dual carbon constraints become a long - term premise, where do different industries stand and what roles do they play respectively?

This is exactly the perspective that the organizers, Orient Securities and 36Kr, as the capital side and the media side, need to supplement. In the overall environment, there is no shortage of narratives about the dual carbon goals, but there is a lack of judgment coordinates.

Visiting the artificial intelligence industry in Shanghai brings this judgment process to the industrial site. From the industrial density of Model Speed Space to the technological maturity shown in the project roadshows, this stop does not attempt to give a unified answer but is constantly narrowing the judgment range. Discussions on issues such as whether the technology can be deployed, whether it has long - term operational stability, and whether it can achieve a balance between cost and efficiency... clarify the real position of "future technology" in the dual carbon narrative, making the dual carbon goals no longer just policy targets or technological fantasies, but choices that can be disassembled, compared, and allocated.

In this sense, the role of this Dual Carbon Star Species · Carbon Exploration Plan has become clearer. It is an intermediate field where multiple forces such as capital, industry, policy, and technology calibrate with each other. Through repeated visits to real - world scenarios, it continuously approaches the answer to the same question: In an era where the dual carbon goals are a long - term premise, which innovations deserve long - term attention?

This is particularly crucial for the dual carbon field, which has a long cycle and high uncertainty.