Que s'est-il passé
At HUAWEI CONNECT 2026 in Shanghai, Huawei unveiled the SCALE partner support system and a seven-step enterprise AI deployment framework. The company also introduced DIMAK, an engineering system designed to provide technical components for building and operating AI applications across data, infrastructure, models, agents, and knowledge.
Huawei's new SCALE partner support system focuses on five areas: scenario-based solutions, co-innovation, aligned marketing, local service, and efficient collaboration. The company reports that partners have already developed over 100 scenario-based solutions across eight industries using Huawei's infrastructure and tools.
The seven-step deployment framework guides enterprises from identifying business logic and core scenarios to full production deployment and value stream extension. This process includes a 12-question assessment tool for evaluating potential AI use cases.
The DIMAK engineering system provides components for Data, Infrastructure, Model, Agent, and Knowledge. It is designed to be compatible with major AI models and integrates with Huawei's open-source platform, openJiuwen, to manage agent responsibilities, permissions, and multi-agent collaboration.
Huawei has reorganized its internal structure to support these efforts, establishing 66 sub-business units and account departments across 10 industries, supported by specialized computing platform teams.
Détails de la source: crnasia.com ↗
Pourquoi c'est important
The initiative represents a strategic shift by Huawei to move enterprise AI from isolated pilot projects to integrated, large-scale production systems. By providing standardized reference architectures and a structured engineering methodology, Huawei aims to lower the barrier for partners and enterprises to deploy AI in complex industrial environments. This framework emphasizes organizational redesign, suggesting that successful AI adoption requires restructuring workflows around human-AI collaboration rather than simply layering technology onto existing processes.
Huawei Deputy Chairman of the Supervisory Board Tao Jingwen emphasized that enterprises must move beyond 'mounting an engine onto an old carriage' and instead redesign processes to facilitate collaboration between employees and AI. This reflects a broader industry trend toward agentic workflows.
The framework addresses the common enterprise challenge of 'pilot purgatory,' where AI projects fail to transition into routine operations. By providing a structured engineering path, Huawei aims to standardize the integration of AI into core IT infrastructure.
The inclusion of agent engineering within the DIMAK system highlights the increasing focus on multi-agent systems in enterprise environments, where governance and defined responsibilities are critical to preventing conflicting actions between autonomous agents.
Mécanisme interactif : comment cela fonctionne réellement
Explorez de manière interactive la technologie sous-jacente à ce développement.
crm_get_transaction(id='4092').An agent must create a draft calendar event for Tuesday at 2 p.m. Which evidence would establish the requested result?
Que regarder ensuite
The practical effectiveness of the DIMAK system and the seven-step framework in real-world industrial settings remains to be seen. Observers should monitor how effectively Huawei’s partners utilize the openJiuwen platform for multi-agent governance and whether the company’s 66 industry-specific sub-business units can successfully scale these deployments across the 10 industries targeted by the new support structure.
The actual adoption rate of the DIMAK system among Huawei's enterprise partners will be a key indicator of the framework's utility.
The performance of the openJiuwen platform in managing complex multi-agent environments in production, as opposed to controlled showcases, remains a significant variable.
Huawei has not disclosed specific pricing or access conditions for the SCALE program or the DIMAK engineering components, leaving the financial barrier to entry for enterprises unknown.