DataCanvas Debuts at MWC2026, Building a New Foundation for Sustainable AI Development with Inclusive Intelligent Computing

The Mobile World Congress 2026 (MWC 2026), hosted by the GSMA, officially opened in Barcelona. This year marks the 20th anniversary of MWC being held in Barcelona. The global technology event, themed “The IQ Era”, focuses on the deep integration of AI, network technologies, and various industries.

As a leading provider of AI infrastructure and intelligent computing cloud services, DataCanvas showcased its global expansion strategy and industry deployment achievements at the event, and shared its core insights on promoting the sustainable development of the AI industry through an inclusive computing power model.

At this year’s exhibition, Chinese companies became a core driving force, with the number of exhibitors increasing to 350. As 5G-A enters large-scale commercialization and AI agents rapidly expand, Chinese technology companies are accelerating globalization with an open approach, shifting from “going global with scale” to “going global with value.” As a representative enterprise, DataCanvas presented its key achievements at the exhibition, comprehensively showcasing its global deployment of AI infrastructure and practical applications across vertical industries.

Chairman Dr. FANG Lei shared on-site the company’s intelligent computing cloud platform designed for the second half of large-model training. Centered on reinforcement learning and oriented toward collective intelligence, the platform provides a new pathway for the intelligent upgrading of global industries. During the exhibition, delegations from international organizations, multiple global telecom operators, enterprises, and partners visited the booth and held in-depth discussions with the company’s leadership team—including Chairman Dr. FANG Lei, Co-founder and COO SHANG Mingdong, and Senior Vice President HUANG Beining—on topics such as global intelligent computing infrastructure development and the joint construction of a computing power ecosystem.

At present, the AI wave is deeply penetrating the capillaries of major industries, and a new era defined by “connected intelligence” is unfolding. According to a report released by GSMA, the construction of telecom private data centers continues to accelerate. As of December 2025, 50% of AI-related deployments had entered the commercial stage, doubling from 25% in September 2025, fully demonstrating the strong momentum of AI–industry integration. Meanwhile, the rapid adoption of large AI models and intelligent agents has created unprecedented demand for computing infrastructure, driving profound changes in AI training paradigms.

Dr. FANG Lei stated during an industry forum that artificial intelligence has entered a new stage centered on large models and intelligent agents, and AI training paradigms are undergoing profound transformation. Observations suggest that over the next five years, the proportion of reinforcement learning–driven post-training will increase significantly and become key to technology implementation and value realization.

He also pointed out that in the current development of the AI era, a noteworthy phenomenon of “imbalanced growth” has emerged. A small number of large-model enterprises have achieved significant value growth through the continuous learning of intelligent agents and the integration of industry knowledge. Meanwhile, some traditional industries—such as software, finance, and manufacturing—are facing pressures in adapting to this transformation, and structural adjustments are also appearing in the labor market. This reminds us that a healthy and sustainable AI ecosystem must move toward a collaborative and inclusive development model.

To address this challenge, DataCanvas has built an intelligent computing cloud platform as its core, connecting open-source foundation models with partners across industries to establish an open and collaborative collective intelligence agent ecosystem, enabling cross-domain knowledge co-creation. This model allows enterprises in various industries to embrace AI while maintaining technological autonomy, transforming them from passive value losers into active co-creators of value and promoting industrial upgrading toward an inclusive and sustainable development path.

At the technological ecosystem level, the DataCanvas intelligent computing cloud platform is centered on Agentic RL technology, precisely addressing the core needs of building professional expert agents across industries and promoting the large-scale deployment of collective intelligence. This technological framework differs fundamentally from traditional reinforcement learning approaches. Its core advantage lies in diversified and focused objective design, which efficiently transforms general intelligence into expert intelligence with domain specialization, and further amplifies the effects of collective intelligence through dynamic combinations, providing highly adaptable and efficient solutions for intelligent industrial upgrading.

This technological system has been fully validated in the field of automotive assembly automation. By enabling AI to take over the entire assembly process and achieve closed-loop self-evolution, it not only significantly improves production efficiency and assembly precision but also greatly reduces deployment costs, becoming a practical model for the intelligent upgrading of manufacturing.

As a pioneer of global inclusive computing power, DataCanvas has built a mature inclusive computing model based on several key innovations, including the world’s first standardized computing power measurement system and the first reinforcement learning cloud platform. The DataCanvas intelligent computing cloud platform adopts a Serverless architecture, enabling on-demand allocation of computing power with millisecond-level responsiveness—without the need to reserve GPU or manage underlying hardware—completely breaking through the limitations of traditional computing models.

The platform also provides a full-stack software-defined AIDC solution and adopts a pay-per-use model, significantly improving GPU utilization and reducing overall computing costs by more than 60%. This allows AI training to extend beyond leading enterprises to organizations of all sizes, enabling AI technologies to move from isolated industry applications to truly inclusive and widespread adoption.

Co-founder and COO SHANG Mingdong further explained the concept of DataCanvas’s “inclusive intelligent computing” model during the conference discussions: the key to successful AI industrialization lies in transforming cutting-edge technologies into stable, reliable, and economically accessible “productivity services.” The emerging trend is that future competitiveness in the AI industry will depend on the ability to integrate and apply continuously evolving global intelligent technologies in an agile, efficient, and cost-effective manner. DataCanvas hopes that by providing standardized, open, and efficient intelligent computing infrastructure, every valuable AI idea—regardless of where it originates—can receive equal computing support, ultimately fostering a more collaborative global AI innovation ecosystem rather than a zero-sum competition.

At the event, Senior Vice President HUANG Beining stated: “Feeling the pulse of the new era of intelligence is not limited to the major exhibitors in Barcelona; it is also reflected in the model training of startups in Southeast Asia and the practical exploration of model-calling tools by developers around the world. Regardless of company size, every AI innovation relies on strong computing power. We remain committed to advancing inclusive computing power practices, overcoming barriers of geography, economic disparity, and cultural differences, accelerating the evolution of the global cloud-network broadband industry, and enabling market participants worldwide to share the development dividends of the intelligent era.”

At this year’s MWC, “Tech4All” has become a highly prominent theme, reminding the entire industry that The IQ Era is not merely an era of “smarter” technologies, but an era of industries that are “more usable, more trustworthy, and more widely accessible.” Inclusiveness is the core principle of AI technology development, which closely aligns with the philosophy advocated by DataCanvas. Guided by technological leadership and driven by innovative models, DataCanvas continues to strengthen its full-stack intelligent computing capabilities, enabling computing power to flow as freely as water and electricity, and contributing to the inclusive and sustainable intelligent transformation of industries worldwide.