Kevin O'Leary Challenges AI Water Fears with Golf Course Analogy

When headlines scream about the environmental toll of artificial intelligence, the conversation often centers on energy consumption. But another resource is quietly entering the debate: water. Critics warn that the massive cooling systems required for AI data centers could strain local water supplies, especially in drought-prone regions. Kevin O'Leary, the blunt-speaking investor known from Shark Tank and now deeply involved in AI infrastructure projects, has a different take. He argues the fear is overblown, comparing the water use of modern data centers to something familiar and seemingly benign—a well-maintained golf course.

O'Leary’s comment isn’t just a casual analogy. It reflects a growing effort by industry leaders to reframe the narrative around AI’s environmental footprint. As data centers multiply to support everything from large language models to real-time analytics, understanding their actual resource demands becomes crucial. The reality, he suggests, is more nuanced than the alarmist headlines imply.

Modern cooling systems are far more efficient than many assume

One of the core misunderstandings, according to O'Leary, lies in how people imagine data centers operate. The image of racks upon racks of servers guzzling water like industrial plants persists, but that doesn’t match today’s engineering. Most large-scale facilities now rely on closed-loop cooling systems or advanced air-based methods that minimize continual water loss. Evaporative cooling, which does use water, is often designed to recycle and treat it on-site, drastically reducing net consumption.

He points out that a typical hyperscale data center might use several million gallons of water annually—a figure that sounds large until you compare it to other commercial users. A single 18-hole golf course in a warm climate can easily consume that much or more each year just to keep the grass green. Unlike golf courses, however, data centers are often located in areas chosen specifically for their access to renewable energy and sustainable water sources. Site selection now weighs environmental impact heavily, not just proximity to fiber lines or power grids.

The comparison isn’t meant to dismiss concerns but to recalibrate them. Water use in tech infrastructure isn’t negligible, but it’s also not inherently wasteful when measured against other large-scale land uses. Context matters, and O'Leary believes the public debate lacks that perspective.

Investment in AI infrastructure continues despite scrutiny

O'Leary’s stance comes as he backs several AI-focused data center ventures, betting that demand for computing power will keep accelerating. He’s not alone. Major tech firms and private equity players are pouring billions into new facilities, driven by the insatiable appetite for AI training and inference workloads. Even as questions about sustainability grow, the financial momentum shows little sign of slowing.

This creates a tension. On one hand, the industry needs to expand to support innovation in healthcare, climate modeling, and logistics. On the other, communities near proposed sites are rightly asking about long-term impacts on aquifers and municipal supplies. Some projects have faced delays or modifications after local resistance, particularly in regions like Arizona or Spain where water scarcity is a daily reality.

O'Leary acknowledges that not all data centers are created equal. Older facilities or those in less regulated markets might still rely on inefficient practices. But he insists that new developments, especially those backed by serious investors, are built to higher standards. Transparency, he argues, should be paired with recognition of progress—not just criticism of intent.

The broader conversation needs nuance

The water debate around AI data centers mirrors earlier concerns about cryptocurrency mining or even traditional manufacturing. New technologies often face scrutiny over resource use before efficiency gains catch up. What’s different now is the scale and speed of AI’s growth, which amplifies both opportunity and apprehension.

Rather than dismissing worries outright, O'Leary suggests shifting the focus toward solutions. Innovations in liquid cooling, AI-driven optimization of facility operations, and even using treated wastewater for non-potable needs are already emerging. Some companies are experimenting with locating centers near desalination plants or in colder climates where free-air cooling works year-round.

He also notes that public perception often lags behind technological reality. Just as few people picture a modern data center as a sleek, quiet facility humming with efficiency, many overestimate its thirst. Bridging that gap requires better communication from the industry—not just defensive comparisons, but clear data on sourcing, recycling, and conservation efforts.

A balanced view serves everyone

Ultimately, the goal isn’t to win an argument about golf courses versus servers. It’s to ensure that as AI reshapes the economy, it does so responsibly. Water stewardship is a legitimate concern, especially as climate change intensifies pressure on freshwater systems worldwide. But solutions come from informed dialogue, not caricature.

O'Leary’s analogy may simplify a complex issue, but it serves a purpose: it challenges assumptions. If we’re going to build the infrastructure that powers the next wave of innovation, we need to understand its real impact—not just its most sensationalized version. Sometimes, that means looking past the headline and checking the numbers. And occasionally, it means admitting that a data center might not be all that different, in water terms, from the perfectly manicured fairway down the road.