AI uses less water than the public thinks

The News

Recent reporting and analysis reveal a significant disconnect between public perception and the actual water consumption of artificial intelligence (AI) infrastructure [1]. While concerns about AI's environmental impact are valid, the narrative frequently overstates its water footprint compared to other industries and even some traditional computing tasks [1]. The California Water Blog’s editorial board recently published an article highlighting this discrepancy, arguing that focusing on AI’s water use can divert attention from more urgent water management challenges facing California and globally [1]. This comes as a newly discovered Linux exploit, CopyFail (CVE-2026-31431), threatens data center security, further complicating the operational landscape for AI deployments [2]. The exploit, though patched, highlights ongoing vulnerabilities in large-scale computing infrastructure, impacting both water and energy consumption [2]. The timing of this revelation, alongside the rise of ethically questionable biotech startups like R3 Bio, which is promoting “brainless clones” [3], adds another layer of complexity to discussions about technological advancement and its societal implications.

The Context

The misconception about AI's water usage stems from a misunderstanding of the infrastructure required for training and deploying large language models (LLMs) and other AI systems [1]. The water footprint primarily arises from data center cooling, which generates significant heat during computationally intensive tasks [1]. Traditional cooling systems often rely on evaporative cooling, using water to dissipate heat [1]. However, the editorial board notes that water use per computation is often less than commonly perceived [1]. The process isn’t simply a matter of “AI uses X gallons of water”; it’s a function of data center design, geographic location (access to water resources), and cooling efficiency [1]. For example, data centers in arid regions often use more water-intensive cooling methods due to limited alternatives, exacerbating the perceived impact [1].

The rise of LLMs, with their massive parameter counts and demanding training requirements, has fueled concerns [1]. Training a single LLM can require substantial computational resources, increasing energy consumption and, consequently, water use for cooling [1]. However, model efficiency is improving through techniques like quantization and pruning, which reduce computational load without sacrificing performance [1]. Alternative cooling methods, such as dry cooling (air-based heat dissipation) and liquid cooling (non-evaporative fluid heat transfer), are gaining traction as more sustainable options [1]. These technologies are driven by environmental concerns and rising water costs in many regions [1]. The emergence of CopyFail underscores the fragility of AI infrastructure, as the vulnerability allows attackers to gain root access, potentially disrupting operations and impacting energy and water consumption through system instability [2]. Patching the vulnerability is critical, but the fact that many machines remain at risk highlights ongoing security challenges in complex, distributed environments [2]. The situation is further complicated by ethically questionable ventures like R3 Bio, which are exploring radical biotechnological solutions without fully addressing ethical implications [3].

Why It Matters

The misrepresentation of AI’s water footprint has significant implications across sectors. For AI developers and engineers, the inaccurate narrative can create unnecessary pressure to adopt less efficient but perceived as “greener” solutions [1]. This can stifle innovation and hinder the development of truly sustainable practices [1]. Enterprises and startups in the AI space face similar challenges. While sustainability is becoming a key differentiator for attracting investment and customers, distorted perceptions of water usage can lead to misallocated resources and misguided marketing [1]. For instance, a startup might prioritize a less performant, water-efficient model over a more powerful alternative, ultimately harming competitiveness [1].

The winners and losers in this ecosystem are increasingly defined by their ability to balance performance and sustainability [1]. Companies investing in efficient data center design, alternative cooling technologies, and model optimization are likely to gain a competitive edge [1]. Conversely, organizations relying on outdated infrastructure and perpetuating the myth of AI’s excessive water consumption risk being viewed as environmentally irresponsible [1]. The Birdfy smart bird feeder deal, offering discounts on AI-powered bird identification devices [4], illustrates growing consumer demand for environmentally conscious technology [4]. This trend extends to AI infrastructure, creating market incentives for sustainable practices [4]. The CopyFail exploit adds complexity, as compromised systems can lead to inefficient operations and increased energy/water consumption due to instability or malicious activity [2].

The Bigger Picture

The debate over AI’s water usage reflects a broader trend: oversimplifying the environmental impact of complex technologies [1]. While climate change and resource depletion are valid concerns, attributing blame solely to AI distracts from systemic issues driving environmental degradation [1]. The focus should shift to a holistic assessment of the entire technology lifecycle, from resource extraction to manufacturing, operation, and disposal [1]. This requires a nuanced understanding of trade-offs in technological choices [1]. Competitors in the AI space are increasingly recognizing sustainability’s importance, with some companies investing in renewable energy for data centers and exploring carbon offsetting programs [1]. However, these efforts are often overshadowed by the narrative of AI’s excessive water consumption [1].

The emergence of ethically questionable biotech startups like R3 Bio, promoting “brainless clones” [3], highlights a broader societal trend: pursuing radical technological solutions without fully considering ethical implications [3]. This parallels the AI debate, where attention often fixates on immediate, quantifiable impacts (like water usage) while neglecting long-term societal consequences [3]. Over the next 12–18 months, increased scrutiny of AI’s environmental footprint is expected, with greater emphasis on transparency and accountability [1]. The adoption of alternative cooling technologies and model optimization techniques is likely to accelerate as water and energy costs rise [1]. The ongoing threat of exploits like CopyFail will also necessitate renewed focus on cybersecurity and data center resilience [2].

Daily Neural Digest Analysis

Mainstream media coverage of AI’s environmental impact often leans into sensationalism, amplifying anxieties without offering balanced perspectives [1]. The focus on water usage, while not entirely unwarranted, risks diverting attention from more pressing issues like e-waste generation and the carbon footprint of manufacturing AI hardware [1]. The CopyFail exploit serves as a stark reminder of vulnerabilities in complex computing systems, with potential for malicious actors to disrupt operations and exacerbate environmental impacts [2]. The ethical considerations raised by ventures like R3 Bio further underscore the need for a more holistic and responsible approach to technological development [3]. The narrative around AI’s water usage is a distraction; the real challenge lies in fostering sustainability across the entire technology ecosystem, from design and manufacturing to deployment and disposal. Given the growing sophistication of cyberattacks and the pressure to develop ethically sound AI solutions, how can we ensure technological advancement doesn’t inadvertently create new environmental and societal risks?

---

References

[1] Editorial_board — Original article — https://californiawaterblog.com/2026/04/26/ai-water-use-distractions-and-lessons-for-california/

[2] Wired — Dangerous New Linux Exploit Gives Attackers Root Access to Countless Computers — https://www.wired.com/story/dangerous-new-linux-exploit-gives-attackers-root-access-to-countless-computers/

[3] MIT Tech Review — Exclusive eBook: Inside the stealthy startup that pitched brainless human clones — https://www.technologyreview.com/2026/04/30/1136684/exclusive-ebook-inside-the-stealthy-startup-that-pitched-brainless-human-clones/

[4] The Verge — Birdfy’s smart bird feeder is down to its best-ever price for Mother’s Day — https://www.theverge.com/gadgets/922165/netvue-birdfy-smart-bird-feeder-bath-mothers-day-deal-sale

[5] ArXiv — AI uses less water than the public thinks — related_paper — http://arxiv.org/abs/1002.1160v1

[6] ArXiv — AI uses less water than the public thinks — related_paper — http://arxiv.org/abs/2601.16513v1

[7] ArXiv — AI uses less water than the public thinks — related_paper — http://arxiv.org/abs/1303.0042v1