In the deep integration era of AI and industry by 2026, enterprise demand for "cloud" services has shifted from resource elasticity to intelligent-native capabilities, security, controllability, and practical implementation. Cloud products in the market are rapidly diverging: general-purpose public clouds focus on making computing power universally accessible, while intelligent cloud platforms targeting the government and enterprise market increasingly emphasize compatibility with domestic IT ecosystems, heterogeneous computing resource scheduling, embedded industry knowledge, and hybrid deployment support. Today, the launch of Phanthy Cloud by FOURTH PARADIGM (06682) is a quintessential example of this trend—leveraging a decade of AI engineering experience in vertical sectors to build a practical intelligent foundation that integrates models, computing power, and governance, marking the official entry of enterprise cloud services into a new "AI-as-a-Service" (AIaaS) phase.
FOURTH PARADIGM founder Dai Wenyuan attended the launch event and delivered a speech, while the company's co-founder and Chief Scientist, Chen Yuqiang, officially introduced and elaborated on the core architecture and positioning of Phanthy Cloud. Differentiating itself from simple model hosting services, Phanthy Cloud represents the comprehensive cloudification and productization of FOURTH PARADIGM's decade-long expertise in AI engineering, industry-specific modeling, and computing resource scheduling. It aims to provide enterprises with an "out-of-the-box, continuously evolving" AI infrastructure, directly addressing the market's urgent need for intelligent cloud platforms that enable rapid deployment, continuous iteration, and deep business integration.
Phanthy Cloud adopts a three-layer integrated design to offer enterprises a full-stack solution spanning from underlying resources to top-level intelligence: The Cloud Layer ensures seamless compatibility with public, private, and hybrid deployment models, meeting the stringent security and compliance requirements of government and enterprise clients while guaranteeing elastic scalability. The Computing Power Layer integrates a high-performance heterogeneous vGPU scheduling engine, supported by the globally active computing management community HAMi, enabling fine-grained partitioning, isolation, and unified scheduling of GPU resources. This design significantly enhances resource utilization and delivery efficiency in multi-team, multi-task, and fluctuating workload scenarios, making substantial investments in computing power more cost-controllable. The Capability Layer fully incorporates the core AI competencies accumulated through FOURTH PARADIGM's ten years of in-house R&D, specifically comprising three pillars.
1. Model Ecosystem: Tackling the "Last Mile" of Domestic Adoption. Addressing common pain points in domestic substitution initiatives—such as difficult adaptation, slow deployment, and fragmentation—Phanthy Cloud provides access to an industry-leading community for large domestic models. This community has achieved a remarkable leap from zero to over 10,000 certified models within just three months and is progressing towards 100,000+. Enterprises can thereby rapidly acquire validated model assets, significantly shortening the selection and integration cycle and allowing them to focus energy on business innovation.
2. Engineering Platform: Leveraging a Consistently Leading AI Development System. The longevity of scaled AI applications hinges on continuous iteration. Phanthy Cloud integrates the AI development platform—Prophet AIOS 5.2—which has ranked first in IDC market share for seven consecutive years. It supports enterprises in independently developing, deploying, evaluating, and governing models based on their own data and business changes, forming a closed-loop, evolvable AI production pipeline that ensures intelligent applications can create stable, long-term value.
3. Vertical World Models: Enabling AI to Truly "Understand Business and Be Implementable". In specialized fields like financial risk control, energy dispatch, and manufacturing scheduling, the "expressive power" of general-purpose large models is insufficient for solving complex decision-making problems. Phanthy Cloud innovatively platforms FOURTH PARADIGM's decade-long沉淀 (accumulated) vertical world model capabilities, transforming industry mechanisms and high-quality data into reusable model assets. This enables AI in critical scenarios not just to answer questions, but also to perform precise reasoning, evaluation, and optimization, achieving verifiable intelligent implementation.
The future competitiveness of AI lies not in parameter scale, but in the ability to construct "usable, trustworthy, and evolvable" enterprise-grade intelligent systems. The launch of Phanthy Cloud is a critical step for FOURTH PARADIGM in transforming intelligent capabilities into a new type of enterprise infrastructure. By integrating high-performance computing power, an open model ecosystem, vertical world knowledge, and a mature engineering system, it is committed to enabling every enterprise to build its own future-oriented AI capability stack with lower barriers to entry, higher efficiency, and a more robust pathway.
Comments