GuShiio.com This Spring Festival, DeepSeek can be said to be a big hit. Its open source in many versions such as V3 and R1 makes advanced AI models no longer a black technology accessible to overseas giants and a few people. DeepSeek has taken off the high school of advanced AI models and become AI technology that is universally beneficial to all.
Industry intelligence based on DeepSeek is inseparable from the convenient supply of massive computing power, while advanced computing power service platforms such as the National Supercomputing Internet allow the formation of a combination of “advanced models + strong computing power” and promote more AI application innovations have emerged.
Recently, the national supercomputing Internet platform quickly launched the DeepSeek family bucket model, including the full blood version of R1, and provides rapid deployment of API calls and AI web application services, helping enterprises and developers flexibly build exclusive AI applications.
Advanced algorithmic models and Pratt & Whitney’s ultra-intelligent integration capabilities are turning AI into infrastructure like water, electricity and gas. A new technology movement to accelerate the implementation of inclusive AI applications is kicking off across the country.
DeepSeek’s AI inclusive era requires us to understand, explore, and embrace deeply. The first step in understanding Inclusive AI is to understand why the integration of powerful models and super-intelligence is cost-effective and was once the flower of high school in the eyes of ordinary people.
Without the intelligence of thousands of businesses, there would be no intelligence of the entire China. The so-called AI inclusiveness means that all walks of life can have abundant and low-cost access to advanced AI algorithms and integrated computing power.
This requires the sinking of two basic AI elements:
The first is the sinking of advanced models.
Previously, reasoning models such as OpenAI o1/o3 were the best flowers in the field of algorithms. The subscription fees were high, making it difficult for ordinary people to use them, and they did not open services to the country. Since the beginning of 2025, DeepSeek has significantly reduced model training costs through technological breakthroughs such as the Hybrid Expert Model (MoE) and the Multiple Potential Attention Mechanism (MLA), as well as innovations in enhanced learning and training methods. DeepSeek R1 is completely open source for algorithm architecture, weights, etc., breaking overseas monopoly on advanced reasoning models, and supporting small and medium-sized enterprises to access high-performance AI at low cost.
The other is the sinking of super-intelligent and cost-effective resources.
DeepSeek has high performance and reduces computing power requirements for single model training. However, the introduction of DeepSeek upgrade and intelligence by Qianhang Baiye will lead to an increase in the total number of AI models and applications. As the Jevons Paradox puts it, when the efficiency of using a resource increases, the total consumption of that resource may increase rather than decrease.
The widespread use of productivity tools like DeepSeek, which improve efficiency through technological innovation, will drive an increase in computing power demand on a large scale. Looking at the current domestic computing power landscape, GPU-based AI dedicated computing power is expensive and lacks resources, which cannot support the intelligent computing power needs of thousands of industries.
At this time, the integrated computing power cluster dominated by supercomputers can provide surging computing power guarantee to meet the computing power needs of large model training and reasoning and large-scale applications in the industry.
Therefore, it is imperative to reduce the computing power shortage and cost burden of introducing AI in all walks of life, and to deepen and benefit the cost-effectiveness of super-intelligent integration. At this time, new challenges emerged.
For a long time, supercomputing has been regarded as the flower of the top in the computing field. It is not easy for ordinary companies and developers to use it. The challenges mainly come from several aspects:
The first is abilityMaking good use of multi-heterogeneous fusion computing power and adapting it to AI tasks to fully release performance requires certain R & D and engineering capabilities, which exceeds the capabilities of many companies and developers.
The second is efficiencyThe requirements for large-scale, high concurrency, stability and reliability of AI tasks are highly dependent on multi-cluster and multi-center collaboration. This involves comprehensive capabilities such as cross-center resource coordination, inter-computing interconnection, and inter-computing interconnection, which comprehensively tests the efficient operation and management capabilities of the computing power platform.
third, the serviceAI development and operation to commercial monetization have many links and long processes. Supporting the development of Agents, vertical models, applications, etc. requires a complete tool chain and ecological support. This requires the platform to transform from a shovel seller to a service provider, which is a lot of traditional data center operators have not yet transformed.
For the prosperity of industry intelligence, advanced algorithms and supercomputing power are indispensable. Relying on the massive resources of supercomputing and smart computing centers across the country behind the supercomputing Internet has become a must-have option. DeepSeek allows enterprises and developers to achieve model benefits. Who can take advantage of supercomputing for the industry? The national supercomputing Internet platform has taken action.
At the critical moment when DeepSeek exploded and servers were frequently down, the national supercomputing Internet platform stepped in and released a series of models and Chatbot services such as DeepSeek-R1, V3, and Coder on February 3. On the first day of launch, the number of clicks exceeded one million.
What does the national supercomputing Internet rely on to support millions of visits?
It turned out that before the outbreak of DeepSeek, the national supercomputing Internet had cultivated a fertile ground for supercomputing that could be used by the industry through intensive cultivation.
On the supply side, from the perspective of basic computing power.
Supercomputing Internet is abundant in resources and has connected more than 20 supercomputing and smart computing centers in 14 provinces in China. This national-level comprehensive computing power service platform is also a fertile ground for computing power.
Supported by the abundant, safe and efficient supercomputing Internet computing power base, users can now quickly access and have smooth conversations with Deepseek on the platform through the web and mobile phones.
On the application side, from the perspective of the use threshold of enterprises.
Supercomputing Internet has accumulated rich experience in AI ecology and deployment. There are more than 6600 products of various types on the platform, including nearly 240 AI model service products. There are domestic open source models such as DeepSeek and Qwen, as well as foreign AI open source models such as Llama, Stable Diffusion, and Gemma. The rich model matrix and lightweight tools make the Supercomputing Internet a more efficient AI platform. The platform also opens up the AI community for developers to conduct development debugging and knowledge sharing.
With the multiple empowerment of models, tools, and communities, ordinary people can also quickly experience the capabilities of various AI models on the platform, and easily get started with the development of model deployment, optimization, and invocation, advancing from AI users to AI developers., embracing the infinite opportunities of the intelligent era.
At the ecological level, from the perspective of a series of subsequent service guarantees.
The professional service team and ecological support plan of the Supercomputing Internet platform further promote public participation in AI technology innovation and help AI startups like DeepSeek grow and grow.
On the one hand, the professional team of Supercomputing Internet can quickly respond to enterprises ‘personalized Kubernetes Engine and other needs. Enterprise customers can focus more on their own business development without having to spend energy on trivial matters such as building an AI training environment.
Also,Through the “AI Ecosystem Partner Acceleration Plan”, enterprise users can also get incentives such as free use of the DeepSeek API interface for 3 months and a tens of millions of card hours computing power resource poolEmpowering more companies to achieve AI innovation is expected to further accelerate the popularization of AI in more industry scenarios.
It can be said that the Supercomputing Internet has completed the comprehensive integration from computing power supply to enterprise applications and ecological support, allowing AI to enter thousands of industries, protecting it in all directions and carrying out inclusive AI to the end.
DeepSeek is in flames, but without the continuous access and supply of inclusive computing power, resource constraints and busy access will quickly consume the enthusiasm of the public and enterprises to use and explore AI. Without the AI innovation of thousands of industries, breakthroughs in domestic AI can only stop at the developer community and cannot be exchanged for business opportunities. Domestic AI technology cannot enter a virtuous cycle from R & D to commercial use.
From this perspective,The supply of computing power is a necessary condition for China to seize DeepSeek’s innovation opportunities.
The National Supercomputing Internet and DeepSeek work hand in hand to promote the sinking of multiple and massive computing power and make up for the crucial part of inclusive AI’s computing power.
The flowers of inclusive AI on the base of algorithms and supercomputing are blooming this spring, which will kick off the acceleration of intelligence in the industry.
When inclusive AI flows into thousands of industries, Smart China is full of vitality.
