The disruptive power of Energy Efficient Computing to different industries

1. Industrial and Manufacturing

Even though some industries have been researched and optimized for hundreds of years, by using artificial intelligence, companies can still use energy-efficient computing to process data faster and suggest even better systems. It has been proven that smart sensors and AI-driven robotics could optimize power usage and adjust machinery types to different usages at any given moment.

Research published in 2021 compared DeepAR and Prophet to the traditional approach using decline curve models to analyze short-term oil production forecasts. They found “a forecast which outperforms the standard DCA methods” (Tadjer, 2021). Note that the standard decline curve analysis has been widely used since 1940 with many believing that it had been perfected.

With the change to smart factories, the power of energy-efficient computing could help the transformation be more effective and more efficient. By using sensors, companies could adjust production based on the machinery performance, the raw materials’ quality, and external factors such as weather.

2. Artificial Intelligence & High-Performance Computing (HPC)

As mentioned, the disruption of DeepSeek has changed how current generative artificial intelligence investing and customizing is done. Before, it was believed that to create a complex GEN-AI would take plentiful resources, which include capital investment, human resources, and expensive structures. However, it had been proven that this energy-intensive model training could be changed, just a few months before the announcement of DeepSeek, a research paper predicted that the training costs for a single frontier AI model could be billions of dollars by 2023 (MARTENS, 2024). Moreover, the productivity improvement by AI was only 0.5 percent every year. This signals a higher risk in investing in AI. The research paper estimates that it would require 3 percent annual productivity growth for it to be profitable. DeepSeek revolutionized and changed many economists’ theory about the usefulness of AI.

3. Cloud Computing & Data Centers

It is believed that cloud computing requires intensive energy consumption. It is estimated that a single cloud data center would require thousands of servers. About 40% of this energy normally comes from cooling systems, which is costly and leaving a carbon footprint behind. Using an AI-driven algorithm, the power usage could be adjusted depending on workloads to reduce wasteful energy. Due to the prediction’s workload caused by users’ time preferences, AI can potentially predict demand spikes and reallocate the amount of work to different servers based on geographic location.

The ARM-based chips consume much less power than the traditional x86 processors, which made big data applications more energy-efficient.

4. Automotive

China’s electric car sales in 2024 are more than the rest of the world combined (IEA, 2024). By investing in energy-efficient computing, Chinese EV car manufacturers would benefit greatly due to autonomous cars heavily relying on AI-driven perceptions. Companies had been using a compatible chip for each of their car. This chip allows the vehicle to process data locally, which delivers a faster response and better energy saving. This also reduces EV battery drain, which increases the total mileage of each vehicle. This delivers a faster response time, especially with safety features. With the advance of Vehicle-to-everything (V2X) communication, electric vehicles could interact with road infrastructures and other vehicles to ensure better safety and potentially shorter routes for every vehicle. Since EVs’ functions are getting significant improvement every year, it could be seen that the petroleum and diesel vehicle industry is significantly influenced by it. Correspondingly, many governments implement net zero targets; therefore, the rise of EVs can continue.

In a short-term foreseeable future, with the ARM-based processors replacement, energy management would be even more efficient. The Advanced Driver Assistant Systems (ADAS) use cameras and sensors to make real-time decisions. By localizing the system, ADAS does not have to rely on powerful GPUs or cloud servers. All of these advantages already let EVs have more and more benefits than petroleum and diesel vehicles.