As inclusive AI flows into various industries, an intelligent China is thriving.
This Spring Festival, DeepSeek has been exceptionally popular, with its V3 and R1 versions being open-sourced, making advanced AI models no longer the black technology accessible only to overseas giants and a select few. The once rare advanced AI models have been made available to the public through DeepSeek, becoming a technology for everyone.
The industry intelligence based on DeepSeek relies on the convenient supply of massive computing power, and advanced computing service platforms like the National Supercomputing Internet enable the combination of "advanced models + strong computing power," fostering more AI application innovations.
Recently, the National Supercomputing Internet platform quickly launched the full suite of DeepSeek models, including the R1 full version, and provided rapid deployment API calls and AI web application services, helping enterprises and developers flexibly build their own AI applications.
The integration of advanced algorithm models and inclusive super-intelligent computing power is turning AI into a basic infrastructure, akin to water, electricity, and gas. A new technological movement aimed at accelerating the implementation of inclusive AI applications is unfolding across the country.
The era of inclusive AI with DeepSeek requires us to understand, explore, and embrace it deeply. The first step in understanding inclusive AI is to grasp why powerful models and super-intelligent computing power were once seen as rare flowers in the eyes of ordinary people.
Without the intelligence of various industries, there can be no overall intelligence in China. Inclusive AI means that all industries can access advanced AI algorithms and integrated computing power abundantly and at low cost.
This requires the sinking of two basic elements of AI:
First is the sinking of advanced models.
Previously, inference models like OpenAI's o1/o3 were seen as rare flowers in the field of algorithms, with high subscription fees making them difficult for ordinary people to use, and they were not open to domestic services. However, since the beginning of 2025, DeepSeek has significantly reduced model training costs through breakthroughs in technologies such as mixed expert models (MoE) and multi-head latent attention mechanisms (MLA), as well as innovations in reinforcement learning training methods. DeepSeek R1 has fully open-sourced its algorithm architecture and weights, breaking the overseas monopoly on advanced inference models and supporting small and medium-sized enterprises in accessing high-performance AI at low costs.
The second is the sinking of super-intelligent integrated computing resources.
DeepSeek's high performance reduces the computing power requirements for training a single model. However, the introduction of DeepSeek to upgrade intelligence across various industries will lead to an increase in the total number of AI models and applications. As the Jevons Paradox states, when the efficiency of using a resource improves, the total consumption of that resource may actually increase rather than decrease.
Tools like DeepSeek, which enhance productivity through technological innovation, will significantly drive up the demand for computing power. However, looking at the current domestic computing power landscape, GPU-based AI-specific computing power is expensive and scarce, unable to support the computing power needs for the intelligence of various industries.
At this time, integrated computing clusters based on supercomputers can provide robust computing power support to meet the demands for large model training inference and large-scale industry applications.
Therefore, reducing the computing power shortage and cost burden for various industries to adopt AI is imperative. At this point, new challenges arise.
For a long time, supercomputing has been seen as a rare flower in the field of computing, making it difficult for ordinary enterprises and developers to use. The challenges mainly come from several aspects:
First is capability. To fully unleash performance using diverse heterogeneous integrated computing power that matches AI tasks requires a certain level of R&D and engineering capability, which exceeds the capability margins of many enterprises and developers.
Second is efficiency. The demands for large-scale, high-concurrency, and stable reliable AI tasks heavily rely on the collaboration of multiple clusters and centers. This involves comprehensive capabilities such as cross-center resource coordination, inter-computing connectivity, and network interconnectivity, which thoroughly tests the efficient operational management capabilities of the computing power platform.
Third is service. The process of AI development and commercial realization involves many steps and lengthy processes, requiring a complete toolchain and ecosystem support for the development of agents, vertical models, applications, etc. This necessitates a shift from being a mere provider of tools to becoming a service provider, which many traditional data center operators have yet to achieve.
The prosperity of industry intelligence relies on both advanced algorithms and supercomputing power. Relying on the massive resources of supercomputing and intelligent computing centers across the country behind the National Supercomputing Internet becomes a necessary option. DeepSeek enables enterprises and developers to achieve model inclusivity; who can pick the rare flower of supercomputing for the industry? The National Supercomputing Internet platform has taken action.
At the critical moment when DeepSeek was booming and servers frequently crashed, the National Supercomputing Internet platform stepped in, releasing the DeepSeek-R1, V3, Coder, and other series models and Chatbot services on February 3. On the first day of launch, the click volume exceeded one million.
The combination of the national-level computing power platform and DeepSeek has given the public hope for widespread use of the "full version" R1. Indeed, the National Supercomputing Internet platform has lived up to expectations and has now launched the 671B parameter scale DeepSeek model for enterprise development.
What exactly supports the National Supercomputing Internet to handle millions of visits?
Before the explosion of DeepSeek, the National Supercomputing Internet had already cultivated a fertile ground for industry-usable supercomputing through meticulous efforts.
On the supply side, from the basic computing power situation.
The supercomputing internet has abundant resources, connecting over 20 supercomputing and intelligent computing centers across 14 provinces in the country. This national-level comprehensive computing power service platform is also a fertile ground for computing power.
With the support of a rich, secure, and efficient supercomputing internet computing power foundation, users can now quickly access and smoothly interact with DeepSeek through web and mobile platforms.
On the application side, from the perspective of enterprise usage thresholds.
The supercomputing internet has previously accumulated rich AI ecosystem and deployment experience, with over 6,600 types of products on the platform, including nearly 240 AI model service products, featuring both domestic open-source models like DeepSeek and Qwen, as well as foreign AI open-source models like Llama, Stable Diffusion, and Gemma. The rich model matrix and lightweight tools make the supercomputing internet a more efficient AI platform. The platform has also opened an AI community for developers to conduct development debugging and knowledge sharing.
With the multiple empowerments of models, tools, and community, ordinary people can quickly experience the capabilities of various AI large models on the platform and easily get started with model deployment, optimization, and invocation, progressing from AI users to AI developers, embracing the infinite opportunities of the intelligent era.
On the ecological level, from the perspective 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 support the growth of AI startups like DeepSeek.
On one hand, the professional team of the supercomputing internet can quickly respond to enterprises' personalized container service needs, allowing enterprise clients to focus more on their business development without wasting energy on trivial matters like building AI training environments.
Additionally, through the "AI Ecosystem Partner Acceleration Program," enterprise users can also receive three months of free use of the DeepSeek API interface and access to a computing power resource pool of tens of millions of card hours, empowering more enterprises to achieve AI innovation and further accelerating the popularization of AI in more industry scenarios.
It can be said that the supercomputing internet has completed a comprehensive connection from computing power supply to enterprise application and ecological support, ensuring the full-chain and all-round support for AI to enter various industries, fully realizing inclusive AI.
DeepSeek is booming, but without the continuous access and supply of inclusive computing power, resource limitations and busy access will quickly deplete the enthusiasm of the public and enterprises to use and explore AI. Without AI innovation across various industries, breakthroughs in domestic AI can only remain within the developer community, unable to translate into commercial opportunities, and domestic AI technology cannot enter a virtuous cycle from R&D to commercialization.
From this perspective, the supply of computing power is a necessary condition for China to seize the innovative opportunities presented by DeepSeek.
The National Supercomputing Internet and DeepSeek are working hand in hand, advancing together to promote the sinking of diverse and massive computing power, filling the crucial gap in inclusive AI's computing power.
The flower of inclusive AI on the foundation of algorithms and supercomputing is blooming this spring, marking the beginning of accelerated industry intelligence.
As inclusive AI flows into various industries, an intelligent China is thriving.
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
