As enthusiasm surrounding artificial intelligence intensifies, a recent analysis published by MarketWatch suggests that widespread assumptions regarding data centers require reassessment. The opinion column, authored by Robert Ross on September 18, 2026, indicates that the current narrative often misses key operational realities.
Ross highlights recent events within the AI sector, noting that an Anthropic researcher resigned after raising alarms about existential risks posed by advanced AI models. Shortly thereafter, prominent figures including Dario Amodei from Anthropic, Sam Altman of OpenAI, and Elon Musk of SpaceX AI endorsed a pause in the development of frontier models. Ross interprets these moves with skepticism, characterizing them as potential examples of regulatory capture where established labs seek rules they can afford, thereby creating compliance hurdles for smaller competitors.
The article also points to technological shifts in infrastructure. Modern data centers are increasingly adopting closed-loop liquid cooling systems, a development that significantly reduces water waste compared to older methods. This operational evolution underscores the complexity of the sector, which Ross argues is frequently misunderstood by the public.
Ross is the founder of TikStocks and the author of “A Beginner’s Guide to High-Risk, High-Reward Investing.” He previously served as chief equity analyst at Mauldin Economics and currently writes the investment newsletter “Let’s Analyze” on Substack.
Another opinion piece dressed as analysis. I’d like to see actual energy consumption stats before accepting either side.
As an engineer in this field, the public still thinks we’re drowning GPUs in tap water. It’s exhausting to correct this.
Wait, Dario Amodei resigned THEN endorsed a pause? The timeline seems convenient for the incumbents.
Ross’s take on regulatory capture is terrifyingly plausible. Is a pause really about safety or just market control?
I never realized data centers were moving to closed-loop cooling. That water usage myth really needs to die.