무슨 일이 일어났나요?
Jacobs secured a three‑year agreement with NVIDIA to deploy its Data Center Digital Twin at a large‑scale AI research and development facility in the United States. The platform will ingest live sensor data, engineering models and operational technology into a single, real‑time virtual replica built on NVIDIA Omniverse libraries. Facility operators will be able to simulate workload changes, evaluate power and cooling impacts, and run applications such as dynamic load balancing, energy forecasting, leak detection, predictive maintenance and operator training. Jacobs also plans to embed AI agents that can automatically analyze data, flag abnormal conditions and suggest corrective actions, moving the twin beyond visualization toward autonomous management.
Jacobs, a global engineering and construction firm, announced a three‑year contract with NVIDIA to install its Data Center Digital Twin at NVIDIA’s U.S. AI research campus. The solution will combine engineering models, operational technology (OT) data and live sensor feeds into a unified, real‑time virtual environment.
Built on NVIDIA Omniverse libraries, the platform enables operators to model changes in AI workloads and instantly see the impact on power availability, cooling capacity and equipment health. The system supports use cases such as dynamic power load balancing, energy forecasting, liquid‑coolant leak detection, predictive maintenance and immersive operator training.
Jacobs plans to embed AI agents that continuously analyze facility data, identify abnormal conditions and generate recommendations for infrastructure management. The longer‑term goal is to shift the digital twin from a visualization tool to an automated operations assistant, reducing reliance on manual monitoring and physical testing.
소스 세부정보: constructionowners.com ↗
왜 중요한가요?
The deployment shows how AI‑intensive data centers are turning to digital twin technology to manage the growing complexity of power, cooling and equipment performance. By linking live telemetry with engineering simulations, operators can test scenarios without costly physical trials, potentially reducing downtime and energy waste. The inclusion of AI agents points to a broader industry trend of automating infrastructure decisions, which could set new standards for reliability and efficiency in high‑performance computing environments that power large language models and other ‑heavy AI workloads. For data‑center owners, the solution offers a way to anticipate bottlenecks before they affect AI training jobs, improving overall productivity.
AI research facilities consume massive amounts of electricity and generate significant heat, making precise infrastructure management critical for cost control and system reliability. A digital twin that can simulate workload‑induced changes in real time offers a proactive approach to avoid overloads, cooling failures, or unexpected outages.
Integrating AI agents into the twin could automate routine decisions—such as adjusting cooling set points or redistributing power loads—freeing human operators to focus on higher‑level tasks. This automation may become a for future AI‑centric data centers seeking to maximize uptime while minimizing energy waste.
The partnership expands Jacobs’ existing work with NVIDIA on digital‑twin applications, suggesting a growing ecosystem of engineering firms leveraging NVIDIA’s Omniverse platform for AI infrastructure. If successful, the model could be replicated across other hyperscale data centers, influencing industry standards for AI‑hardware operations.
대화형 메커니즘: 실제로 작동하는 방식
이 개발의 이면에 있는 기본 기술을 대화식으로 살펴보세요.
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?
다음에 무엇을 볼 것인가
Future updates on the pilot’s performance metrics, such as reductions in energy consumption or incident response times, will indicate whether the approach scales to other AI facilities. Watch for announcements on pricing, licensing terms for the Omniverse‑based twin, and any additional partners that Jacobs brings on board. The evolution of the AI‑agent component—whether it remains advisory or gains autonomous control—will also be a key indicator of how far digital twins will move toward fully automated data‑center operations.
Performance data from the pilot, including any measurable improvements in energy efficiency, incident response time, or maintenance cost reductions.
Details on commercial availability, pricing structures, and licensing for the Omniverse‑based digital twin, which will determine how widely the technology can be adopted beyond NVIDIA’s own facilities.
Progress on the AI‑agent component: whether it remains a decision‑support tool or evolves into an autonomous controller capable of executing infrastructure changes without human approval.
Potential expansion of the partnership to other NVIDIA sites or third‑party data‑center operators, indicating broader market acceptance.