Why AI's environmental footprint won't disappear into orbit

Morgan Stanley has positioned itself at the forefront of financing the rapid expansion of artificial intelligence infrastructure, leveraging its expertise in debt markets to support the construction of data centers essential for AI operations. According to reports from financial media, the Wall Street bank has played a key role in structuring the capital needed to fund these facilities, which are critical for powering the computational demands of AI technologies. This strategic move comes amid a surge in investment in AI-related ventures, driven by the sector’s potential to revolutionize industries ranging from healthcare to finance. The push for AI infrastructure has intensified as companies seek to expand their capabilities, often requiring vast amounts of computing power. This has led to a growing reliance on data centers, massive facilities housing servers that process and store enormous volumes of data. However, the environmental consequences of expanding these facilities have drawn increasing scrutiny. Environmental advocates warn that the energy consumption of data centers, particularly those supporting AI, contributes significantly to carbon emissions and resource depletion. In the United States alone, data centers accounted for approximately 5% of total electricity usage in 2023, and projections suggest this figure could nearly triple by 2028. Some companies are exploring alternatives to traditional terrestrial data centers, including the possibility of moving operations into space. A concept developed by NASA in the 1970s envisioned satellite-based solar farms capable of transmitting energy to Earth, though the project was abandoned due to economic and political factors. Today, similar ideas are resurfacing, with firms like SpaceX proposing to deploy millions of AI-powered satellites and eventually construct data centers on the Moon. Visionaries such as Jeff Bezos, Elon Musk, and Google’s parent company Alphabet have expressed interest in relocating polluting industries off-world, aiming to reduce Earth’s ecological burden. However, experts caution that the environmental benefits of such initiatives remain uncertain when considering the entire lifecycle of space-based operations. Space-based data centers present unique challenges. Unlike Earth-bound counterparts, which can rely on water and air for cooling, orbital facilities must employ specialized systems to manage heat. This includes massive radiators filled with ammonia, which release excess heat as infrared radiation. These radiators, however, are significantly larger than the computing equipment they service, making them difficult to transport and maintain. Moreover, launching these components into orbit requires substantial amounts of rocket fuel, contributing to greenhouse gas emissions. For example, each Starship launch by SpaceX consumes over 1,000 metric tons of liquid methane, resulting in approximately 80,000 metric tons of CO₂ emissions annually. Soot emissions from rocket launches also pose risks to the stratospheric ozone layer, potentially weakening Earth’s natural defense against harmful ultraviolet radiation. In parallel, resistance to terrestrial data centers has grown, particularly in regions where such facilities threaten local environments and communities. In the United Kingdom, the proposed Norwich Apex Data Center near Keswick has faced strong opposition from residents concerned about its impact on the surrounding landscape, wildlife, and quality of life. The facility, designed to house high-performance computing equipment, would consume enough energy to power 10,000 homes and utilize local water resources for cooling. While developers claim the project will implement a closed-loop water system to minimize consumption, critics argue that the long-term sustainability of such operations remains unclear. Similar concerns have fueled a nationwide movement in the U.S., where data centers have become a focal point of public discontent, prompting calls for regulatory oversight and a temporary halt to new developments. As the debate over AI infrastructure intensifies, both terrestrial and extraterrestrial, the challenge lies in balancing technological advancement with environmental responsibility. Whether through innovations in energy efficiency, policy reforms, or alternative approaches to computation, the path forward will require careful consideration of the broader implications of scaling AI capabilities. The next steps will likely involve increased dialogue between stakeholders, including governments, corporations, and environmental groups, as they navigate the complex trade-offs inherent in this rapidly evolving landscape.