In a past article, I briefly discussed how specific improvements in targeted infrastructure updates could create a project boom for logistics. Going into the next five years, one such area to add to this boom is Artificial Intelligence (AI) and the power needed to sustain it. The AI revolution requires power, and with an increase in power for technology comes an increase in projects for logistics surrounding AI data centers. This need is becoming more and more, especially with the recently released Google AI video generator Veo 3 in May of 2025.
When we think of AI, the first concepts that come to mind for most people are ChatGPT or Amazon’s Alexa.
AI also has been rapidly evolving around AI video generation, which could greatly impact Hollywood by lowering not only financial and technical barriers, but by allowing even independent creators to produce high-quality videos. This in return can create job displacement, copyright issues, ethical concerns, and authenticity scares. China had released their own AI video generator, but now the U.S. is leading the race with Google’s Veo 3. This innovative AI video generator can make realistic and detailed video clips from text and video prompts. This is said to revolutionize how scripts are visualized, and scenes created thus transforming the Hollywood industry with high-quality lifelike videos, with audio and sound effects included.
Googles Veo 3 is available in seventy-three countries and can be found on the Gemini App or Google AI Pro. It has crushed every other competitor and is now ahead in the AI revolution race. However, with drastically high-quality creations comes the need for increased processing power from AI data power centers. Robust power infrastructure is needed to support these powerful technologies and AI data centers are becoming increasingly crucial as these dedicated facilities are designed to handle the high-power demands of AI workloads, density racks, and cooling solutions. AI in general has been driving a massive increase in data center power consumption without even adding onto it the AI video generator Veo 3. Current projections state that there will be a significant rise in energy usage by 2030 because of this. Goldman Sachs in a February 2025 article stated that AI is set to drive a 165% increase in data center power demand by 2030 and the explosion of AI generation has created an arms race to further develop the technology. This means more high-density data centers as well as more electricity to power them. McKinsey & Company back in October 2024 even agreed that the global demand for data center capacity could triple by 2030.
Typically, an AI data center can use around 100,000 homes worth of electricity, with some of the largest centers using 20 times more.
As mentioned above, by 2030, this demand is expected to reach around 1050 TWh, which is more than Japan’s current power usage. Why? Well, the Veo 3 released by Google needs significant processing power thus leading to higher energy consumption compared to tasks like texting and image generation. For instance, a study found that the energy needed to create a low-quality video that only lasted 5 seconds was higher than generating a chatbot response. Yet, Google and other AI companies have not yet released any specific energy consumption figures for these advanced AI generators pushing out high-quality videos that are longer. What is known suggests that high quality and high-definition videos like Veo 3 will have much larger energy footprints when compared to the low quality 5 second videos in the study.
This increase affects power centers and as these AI models improve and are being used widespread, more strain will be placed on these AI specific power centers. Thus, the requirements for more and larger data centers to manage the massive energy workloads so that rapid growth does not put a strain on power grids and hinder its reliability. This power demand has data center owners and investors calculating and exploring a range of energy sources like renewables, natural gas, and nuclear to help run the computing power. However, the transformers store this power and are a commodity in the project logistics field where players such as MTS Logistics could have a hand in transporting.
Transformers play a crucial role for converting incoming high voltage power to lower voltage.
For AI data centers, transformers adjust voltage levels to ensure proper power needs for equipment like servers and cooling systems, as well as power distribution and integration to the power grid. Right now, Asia Pacific and North America are seen as major hubs for data center power. Places like Austin and San Antonio Texas, Iowa, and Montreal are being looked at for new AI data center developments. New means more opportunities for Project Logistics.
Expansion of current data centers are also being looked at as companies are investing heavily in expansion of existing centers as well as new ones. This helps with the need for growing compute power. With compute power emerging as one of this decade’s most critical resources, and with millions of servers running 24/7, overinvesting in datacenter infrastructure is a risk for standing assets, but underinvesting means potential for falling behind. This will be a calculated decision for investors in what is said to be a $7 trillion race to scale data centers. The United States is currently home to more than 3,600 data centers with 80% of them concentrated in fifteen states. The majority of these being housed in Virginia, the state considered to be a global hub with its nearly 600 data centers. Texas follows with 336 and California with 307 data centers.
Whether current AI data centers are expanded or new ones built, project logistics will certainly play a role in the mobilization and transportation of the transformers needed.
Yet, these plans do take time. In September 2024, $30 billion was invested into new and expanding data centers according to Forbes primarily in the U.S., with a total of $100 billion being sought after. Although even with funding, there is still a lot of planning, purchasing, land acquisition, material acquisition, engineering, and so on before the project is put into a stage of transport bids for freight forwarders that handle project cargo. As mentioned though, this is a global AI arms race the U.S. is in and currently winning, so who’s to say things won’t move a little quicker because according to one article interviewee, “If you own the data, you have the power.”
Project Logistics will play a vital role in moving these high value components for power, regardless of what Hollywood wishes for. Drawbacks? Transporting AI and data center equipment like servers, GPUS, racks, and transformers require specialized care due to their sensitive nature. There could also be supply chain constraints such as labor shortages, bottlenecks and regulatory hurdles that delay grid connections and data center expansions. For example, certain transformers needed that are refrigerator sized are in short supply as of 2024 with potential wait times for new units as high as four years. That won’t stop the growing demand though as businesses and individuals increase their usage and rely on AI, the demands on data infrastructure, or the dabbling with Google’s Veo 3 which are accelerating and creating new project logistics opportunities.
