GPUs behind all artificial intelligence | Counting Beans and Saving Green | Edmund Yeung

As sustainability becomes inseparable from business performance, infrastructure decisions are no longer purely operational. They shape long-term outcomes. With the rapid advancement of artificial intelligence, one critical enabler has moved to the forefront: the computing power that turns data into actionable intelligence.

What once took eight engineers 10 months to complete can now be accomplished overnight with a single graphics processing unit, a point frequently highlighted by NVIDIA’s chief scientist Bill Dally. Today, AI can analyze massive datasets and generate powerful insights in seconds. This speed and scale are made possible by GPUs, which provide the essential platform for training and running AI models effectively. However, achieving this performance requires more than just GPUs alone. It depends on specialized data centers equipped with advanced power and cooling systems.

At Towngas, we view GPU infrastructure not as experimental technology, but as a core strategic capability closely aligned with our ESG goals. This approach is supported by our self-built AI data center, operated by our wholly owned subsidiary, Towngas Telecommunications Company. When properly governed and deployed, it drives more efficient operations, enables earlier risk detection, and delivers greater transparency and accountability across our energy, telecommunications, and property businesses.

The quality of the data center environment is critical. Even the most advanced GPU infrastructure can only deliver its full value when housed in a world-class facility. From an environmental perspective, efficiency is paramount. AI demands significant computing power, and poorly optimized workloads can result in unnecessary energy consumption. By leveraging dual energy sources, our GPU platforms allow AI workloads to run faster and more efficiently, achieving the same outcomes with fewer resources.

GPUs also deliver tangible environmental benefits. In our energy operations, we combine drone inspections with GPU-accelerated vision models to detect gas leaks, corrosion, and equipment anomalies across vast and hard-to-reach areas. Early identification of issues reduces energy loss and minimizes the environmental impact of maintenance activities. Over time, these capabilities enable predictive maintenance, extend asset lifespan, and reduce unplanned outages, turning sustainability targets into measurable operational results. All of this is powered by a professionally managed, high-standard GPU infrastructure housed within our LEED Gold-rated AI data center.

Beyond environmental gains, GPUs significantly enhance safety across our operations. In plants, industrial facilities, commercial buildings, and within our own AI data centers, GPU-powered analytics process visual and sensor data in near real time. This capability also strengthens cybersecurity through our newly established Cyber Command Centre. Together, they shift teams from reactive responses to proactive intervention, enabling hazards to be identified and addressed earlier and with greater confidence.

As AI adoption accelerates, strong governance becomes equally essential. AI brings important responsibilities related to data privacy, cybersecurity, and regulatory compliance. Without proper controls, its benefits can quickly be undermined by risk. Our carrier-grade AI data center services and enterprise-grade GPU infrastructure help address these challenges by providing secure, controlled, and fully auditable environments for AI workloads. Centralized platforms also offer robust access management and comprehensive audit trails, enabling us to deploy AI with discipline, transparency, and scalability.

Ultimately, sustainability is about building long-term resilience – for our infrastructure, our business, and the communities we serve. When grounded in strong ESG principles and supported by robust data center management, GPU infrastructure empowers smarter decisions, safer operations, and stronger governance. In this sense, GPUs do more than power AI; they help power sustainable progress.

Edmund Yeung is the ED & CFO of Towngas

 

𝗗𝗼𝘄𝗻𝗹𝗼𝗮𝗱 𝗧𝗵𝗲 𝗦𝘁𝗮𝗻𝗱𝗮𝗿𝗱 𝗔𝗽𝗽 ↓