**AI’s Environmental Footprint: A Growing Concern**

The rapid advancement of artificial intelligence (AI) technology, while revolutionary, is casting a shadow on our environmental landscape. Imagine producing as much carbon pollution as one of the largest cities in the world and consuming as much water as the global population does through bottled water. In 2025, that’s the kind of environmental impact AI is generating. A startling new study challenges us to reconsider the unseen environmental costs of AI, calling for more transparency from tech companies about their resource consumption.

**The Weight of AI’s Carbon Footprint**

According to research published in 2025, AI’s carbon emissions this year are on par with those of New York City, one of the largest and densest urban areas in the world. New York City annually emits around 50 million tons of carbon dioxide, a staggering figure that AI systems now reportedly match. This underlines a major environmental challenge as AI increasingly integrates into our daily lives. The researcher, Alex de Vries-Gao, a PhD candidate at the VU Amsterdam Institute for Environmental Studies, warns through his study that although precise figures are elusive due to data limitations, the impact is undeniably huge. “In the end, everyone is paying the price for this,” he notes.

This environmental toll stems largely from the energy consumption required to power AI systems. De Vries-Gao’s research builds on earlier studies that project the power demand for AI could surpass 23GW in 2025, overshadowing even the electricity used for activities such as Bitcoin mining in 2024. Much of this energy is tied to data centers, massive facilities that house the servers and hardware essential for AI operations.

**The Underestimated Water Consumption**

While AI’s carbon footprint is alarming, its water usage could be even more surprising. De Vries-Gao estimates AI could be using between 312.5 and 764.6 billion liters of water in 2025, surpassing previous estimates and raising essential questions about sustainability. This immense water consumption is primarily linked to cooling systems in data centers. To prevent servers from overheating, significant quantities of water are utilized in cooling systems, while power plants, essential for generating the electricity required by data centers, also gulp substantial amounts of water to cool their equipment.

Shaolei Ren, an associate professor at the University of California, Riverside, co-authored a 2023 study that highlighted similar concerns. He acknowledges the timeliness of de Vries-Gao’s study amidst increasing debates about AI and water. Ren further emphasizes that de Vries-Gao’s analysis is conservative since it doesn’t fully capture the environmental impact across AI’s entire lifecycle, from hardware production to eventual disposal.

**The Call for Transparency**

A core issue exacerbating AI’s environmental impact is the lack of transparency from tech companies. While many produce annual sustainability reports that disclose total carbon emissions and water usage, they often fail to specify the resources dedicated to AI projects. This leaves a significant gap in the data, making it challenging to address AI’s real environmental footprint comprehensively.

De Vries-Gao was forced to rely on a patchwork of analyst estimates, earnings calls, and other publicly available data to gauge AI’s resource usage. He points out that while sustainability reports are a step in the right direction, they frequently omit crucial details, such as indirect water consumption linked to electricity demand and specifics about AI resource allocation.

Without accurate data, it’s difficult to hold companies accountable or to craft policies that adequately address these environmental impacts. Emissions and water usage can vary dramatically depending on factors like the geographical location of data centers and the cleanliness of local power grids. Therefore, transparency about where companies operate or intend to establish new data centers is crucial in understanding AI’s growing environmental consequence.

**Reflecting on the Future of AI and the Environment**

The findings from this study urge us to reflect deeply on our priorities and responsibilities. As de Vries-Gao questions, “We can really ask ourselves, is this how we want it to be? Is this fair?” Transparency is the vital ingredient that could spark important discussions about sustainability and the future path of AI development.

How do we balance technological advancement with environmental stewardship? Can we afford not to insist on greater transparency from the companies at the forefront of this AI revolution?

These pressing questions encourage deliberations on potential strategies for mitigating AI’s environmental impact. Solutions might include developing more energy-efficient AI technology, implementing sustainable practices in data centers, or even reevaluating the necessity of certain AI applications in favor of environmentally friendlier alternatives. The path forward requires concerted effort from tech companies, policymakers, and the public to ensure that AI, while advancing our world, does not compromise the health of our planet.

Are we ready to demand the transparency needed to make informed decisions about the proliferation and use of AI technologies? The conversation must begin now, as time and resources are not infinite.