Siemens Industrial AI Drives Green Efficiency Revolution in Infrastructure Cooling, Parks, and Data Centers
The State Council recently issued the 15th Five-Year Plan Carbon Peaking Action Plan, which calls for advancing green and low-carbon industrial transformation, accelerating construction of green computing facilities, and promoting energy-saving and carbon-reduction measures in buildings.
This policy sets the stage for a major efficiency overhaul in China's cooling plants, industrial parks, and data centers over the next five years.
Siemens is positioning itself as a key participant through its intelligent infrastructure portfolio that connects energy systems, buildings, and industrial processes.
The company presented its solutions at the 2026 World Artificial Intelligence Conference, demonstrating applications across infrastructure, life sciences, consumer goods, and automotive manufacturing.
A Lightweight AI Box for Cooling Optimization
Traditional cooling systems in hotels and commercial buildings rely heavily on manual experience, leading to inefficient matching of chillers, pumps, and cooling towers with real-time loads.
Siemens addresses these challenges with the AI BOX (智冷魔方), a non-invasive edge computing device that integrates with existing building automation systems without requiring equipment replacement or major piping changes.
The solution combines physics-based models of HVAC components with machine-learning algorithms that predict cooling demand 15 minutes ahead and dynamically optimize equipment operation every 15 minutes.
At the five-star Shanghai Yingyi Crowne Plaza Holiday Hotel operated by InterContinental Hotels Group, installation and commissioning were completed in just three days with no interruption to hotel operations.
The deployment delivered approximately 7% additional energy savings on top of the existing control system and shifted maintenance from manual experience to fully autonomous AI-driven operation.
From Single-Point Savings to System-Level Energy-Carbon Management
At the park level, Siemens Smart ECX platform provides unified oversight of distributed energy resources including photovoltaics and storage.
The platform supports metering of 17 energy categories, automatically generates consumption baselines, and predicts usage trends with over 92% deviation-warning accuracy.
Following deployment at Sichuan Chuanrun Chengdu factory, the site achieved a 30% reduction in energy costs, over 90% overall energy efficiency, and cumulative carbon-emission cuts of 64,000 tons.
In November 2025, Chuanrun signed a strategic cooperation agreement with Siemens to scale zero-carbon park solutions and explore AI-liquid cooling and energy digitalization.
Power Infrastructure as the Foundation for AI Compute
Rising rack densities in AI training clusters have elevated power reliability to a critical requirement for data centers.
Siemens supplies end-to-end distribution systems featuring NXAirS medium-voltage switchgear and SIVACON S8 low-voltage panels.
These products were deployed at the Zhongjin Data Ulanqab zero-carbon computing base, China's first source-grid-load-storage integrated project serving the Beijing-Tianjin-Hebei region.
In July 2026, Siemens announced a 300-million-euro investment in Germany to expand capacity for energy-transition and AI data-center technologies, planning to create up to 700 new jobs by 2030.
Industrial AI Grounded in Domain Knowledge
Siemens stresses that successful industrial AI must fuse long-standing domain physics models with modern data-driven techniques rather than applying generic algorithms.
The approach spans hardware, software, and services, enabling customers to obtain complete solutions from a single vendor and reducing integration risk.
Documented results across multiple sites demonstrate that standardized, rapidly deployable products can be replicated from one hotel or park to another, creating scalable industrial value.
Related articles
How Ideco Measures NGFW Performance Using TRex ASTF and RFC 9411 Methodology
Ideco engineers have published a detailed methodology for benchmarking the performance of their Ideco NGFW Novum product on the Ideco EX hardware platform. The team uses TRex in Advanced Stateful mode to generate realistic L4-L7 traffic and follows RFC 9411 guidelines to ensure reproducible results across different test environments. Key metrics include throughput on UDP 1518-byte frames reaching 200 Gbit/s, TCP/HTTP performance varying from 100 Gbit/s down to 62 Gbit/s depending on object size, and EMIX mixed traffic dropping from 92 Gbit/s with firewall only to 15.5 Gbit/s when all inspection engines are enabled. The methodology incorporates binary search for maximum load, a 1% session drop threshold, and 300-second hold times per test point to capture real-world behavior rather than short-term peaks. All tests were performed without TLS decryption, with separate plans for hardware-based Keysight BreakingPoint testing when decryption is required. The full article includes hardware specifications, traffic profiles, loss calculation formulas, and configuration details to allow anyone to reproduce the measurements.
RuBackup 2.9.0 Receives Official Compatibility Certificate with BAUM-Inform Storage Systems
RuBackup version 2.9.0 from the Astra Group has successfully passed compatibility testing with storage systems produced by BAUM-Inform. The verification confirmed that the backup solution can reliably perform data backup and recovery operations using BAUM-Inform hardware. Testing followed a predefined program and methodology, although detailed results and the full list of tested configurations were not disclosed. BAUM-Inform systems are designed for corporate data storage and processing, while RuBackup supports servers, workstations, virtualization environments, mail systems, databases, and domain infrastructure. The certification allows organizations to deploy both products together without additional integration validation. Both companies stated that the verified combination reduces deployment risks and supports the development of fully domestic data protection infrastructures.
VK WorkSpace Adds Guest Access Without Accounts, Password Protection, Polls, and File Uploads Up to 30 MB
VK Tech has updated the cloud version of its VK WorkSpace Board to allow external participants such as clients, contractors, and other guests to join collaborative boards without creating an account. Access can be protected by a password set by the owner, who can also assign specific permissions including view-only, commenting, or full editing rights while restricting guests from creating new boards or viewing version history. Domain administrators can enforce mandatory password protection across all boards that permit guest access. Additional features include built-in polls with automatic result counting, support for uploading files up to 30 MB that are scanned by antivirus software, a mini-map for navigating large projects, and export options in SVG, PNG, JPG, PDF, and CSV formats. The service also received an English-language interface, addressing key use cases involving external collaboration as highlighted by VK Tech executive Petr Shcheglov.
Solar SIEM 2026.2 Adds Full Solar JSOC Detection Library, TI Feeds Support, Enhanced AI Agent and Multi-Tenancy
GC Solar has released Solar SIEM 2026.2, which now includes the complete detection rule library developed by Solar JSOC over 14 years of monitoring approximately 300 customer infrastructures. The update eliminates the need for organizations to spend months building custom rule sets by providing ready-made detection scenarios that identify sophisticated attacks from the earliest implementation stages. In 2025, Solar JSOC recorded 1.16 million security events after false-positive filtering and confirmed more than 33,000 incidents, with malware accounting for 36 percent and unauthorized access attempts for 23 percent. The product also gained native support for TI Feeds, allowing automatic ingestion and correlation of indicators of compromise from Solar 4RAYS and customer-owned sources. The built-in AI agent has been significantly expanded so it can now independently query raw event data, perform deeper analysis, and recommend next steps, reducing manual workload for analysts. Multi-tenancy capabilities enable multiple organizations to share a single SIEM instance while keeping their event streams fully separated, targeting holdings, MSSP providers, and large enterprises with numerous divisions. More than 40 companies of various sizes participated in the pilot program.