Shanghai has unveiled its ambitious “15th Five-Year” software and information service industry plan, targeting a total scale of 4 trillion yuan and an added value exceeding 1.1 trillion yuan by 2030, anchored by massive artificial intelligence computing infrastructure investments and cross-border data trade frameworks.
The Bottom Line
- The Target: Shanghai aims for a 4 trillion yuan software industry scale by 2030, driven by aggressive 100,000-card intelligence cluster deployments in Lingang, Songjiang, and Qingpu.
- Global Capital Inflows: NVIDIA established a strategic partnership with six Wall Street financial giants, including Blackstone and Goldman Sachs, aiming to mobilize over 500 billion dollars in third-party capital for AI infrastructure.
- Market Re-rating: Domestic equity and Hong Kong-listed tech names such as Qiniu Intelligent (HKG: 02567) and Zhipu (HKG: 02513) are capturing value as computational demand shifts from heavy training toward high-volume inference.
Infrastructure Upgrades and Municipal Blueprints
The Shanghai Municipal Commission of Economy and Informatification released its comprehensive development roadmap on August 11, setting explicit performance markers for the region’s digital economy. According to the planning document, the municipal government intends to position the local information technology sector as the primary economic growth engine and a global bridgehead for AI competitiveness.
Central to this blueprint is the execution of the “Hundred-Thousand-Ten Thousand” intelligent computing cluster project. Authorities plan to build ultra-scale 100,000-card clusters across suburban tech hubs while simultaneously transforming traditional data centers into hundred-card edge facilities. This architecture aims to reduce latency for industrial automation, smart medical platforms, and autonomous systems.
Here is the math: as multimodal models and autonomous AI agents penetrate enterprise workflows, computational requirements are transitioning rapidly. Industry analysts at Guosheng Securities note that cloud service provider capital expenditures show no signs of peaking. Instead, the demand slope is steepening as workloads pivot from foundational training to continuous inference execution.
Global Liquidity and Wall Street Financing Pools
While domestic policy frameworks solidify regional infrastructure, international capital markets are mobilizing at an unprecedented scale. NVIDIA (NASDAQ: NVDA) announced a strategic alliance with six premier Wall Street institutions—including Apollo Global Management, Blackstone, BlackRock, Brookfield Asset Management, Goldman Sachs, and KKR—to launch an independent financing platform dedicated to AI infrastructure.
This vehicle targets the deployment of more than 500 billion dollars in third-party capital. It addresses the growing capital expenditure hurdles facing data center developers worldwide. According to reports from the National Development and Reform Commission cited by CCTV News, China has simultaneously brought its first fully domestic 100,000-card artificial intelligence super-cluster online within the national supercomputing internet framework.
But the balance sheet tells a different story regarding hardware constraints. Channel prices for high-end domestic AI servers have climbed significantly, reflecting acute supply chain tightness. Guohai Securities points out that large-scale capital outlays by hyperscalers are successfully converting into top-line revenue, driving high-margin growth for upstream thermal management and hardware leasing firms.
Thermal Engineering and Supply Chain Valuations
The physical reality of running 100,000-card clusters has turned thermal management into a critical bottleneck. According to research from CICC, the global market for data center cooling devices is projected to experience a compound annual growth rate of 59.8% between 2025 and 2027, with total addressable market space expected to exceed 315.1 billion yuan by the end of that window.
| Company / Ticker | Core Business Focus | Key Financial / Operational Metric |
|---|---|---|
| Qiniu Intelligent (HKG: 02567) | MaaS Dispatch Hub / Cloud Services | 2025 AI revenue hit 437 million yuan; first-ever positive adjusted EBITDA of 900,000 yuan. |
| Huiju Technology (HKG: 01729) | MPO Optical Communications / Automotive Cabling | Projected 26/27/28 year total revenue for the company with net profit estimates. |
| Zhipu (HKG: 02513) | Coding Agents / Foundation Models | ZCode programming user base surpassed 1 million. |
| Tianshu Zhixin (HKG: 09903) | General-Purpose GPUs (GPGPU) | Initiated coverage by BofA Securities with a target price of 1013 HKD. |
Publicly traded equities tied to these supply chains are experiencing heightened investor attention. Huiju Technology (HKG: 01729), a key part of the Luxshare system, occupies a strategic position in MPO optical routing and high-speed server wiring. Analysts at China Merchants Securities project the firm’s net profits to scale.
Meanwhile, application layer innovators are reporting improving unit economics. Qiniu Intelligent (HKG: 02567) posted 437 million yuan in artificial intelligence-related revenues for 2025, pushing its adjusted EBITDA into positive territory for the first time as a publicly traded entity. The company’s registered developer count now exceeds 1.8 million, highlighting the commercial viability of localized model-as-a-service architectures.
The Cross-Border Software Trade Channel
Beyond domestic infrastructure, local authorities are opening channels for international expansion. The Shanghai Municipal Commission of Commerce, alongside six other government bodies, released an implementation plan for the National Service Trade Innovation Development Demonstration Zone. The policy explicitly encourages local developers to scale overseas deployment of foundational models, multi-modal generation engines, and autonomous agent frameworks.
This export push aligns with the operational momentum of domestic developers. Zhipu (HKG: 02513) recently upgraded its ZCode programming suite, pushing automated coding agents from basic chat-assistance to complex multi-file task execution. Internal benchmarking data indicates that integration with advanced model tiers has raised task completion rates above competing Western architectures, prompting increased enterprise subscription adoption.
At the hardware level, institutional backing is solidifying around domestic semiconductor designers. Bank of America initiated coverage on Tianshu Zhixin (HKG: 09903), issuing a 1013 HKD price target based on the firm’s dual-line GPGPU architecture and its capacity to capture market share amid ongoing import substitution mandates.
Market Outlook and Structural Risks
The intersection of municipal planning, Wall Street balance sheets, and domestic hardware deployment has created a distinct operational window for technology providers. Investors are prioritizing firms that demonstrate actual earnings conversion over speculative narrative generation.
However, execution risks remain visible. Persistent supply constraints for advanced components could compress margins for server assemblers if component cost inflation outpaces end-user pricing power. For institutional allocators, navigating this sector requires continuous tracking of capacity utilization rates across regional data centers and verified cash flow generation from enterprise AI deployments.
