Behind the soaring computing power of AI, the enormous consumption of water and electricity is triggering strong backlash from communities around the world. To shed the label of data centers as "water monsters," Google earlier...AnnounceA brand-new data center using "advanced air-cooling technology" will be built in Texas, USA, which is said to reduce industrial water consumption to a minimum, even limiting it to the needs of the employee cafeteria and the daily operation of the park.
With the popularization of generative AI, major tech giants are not only vying for GPU computing power, but also competing for extremely precious resources such as electricity and water. According to reports, Google is attempting to...Infrastructure technology upgradeTo alleviate this resource crisis.
This plan, which Google announced at the end of last year, is expected to be completed by 2027.Increase investment in Texas by $400 billionThis is part of the construction of two new AI data centers.
Abandoning water cooling? Texas's new data center emphasizes "advanced air cooling technology".
Google officially stated that the new data center in Texas will fully implement "advanced air-cooling technology." Unlike traditional server rooms that heavily rely on large amounts of fresh water for evaporative cooling, the new design aims to strictly limit water consumption to "critical campus operations," such as employee kitchens and restrooms, completely decoupling the cooling system from water resources.
This concrete environmental commitment builds on Google's previously announced massive two-year, $400 billion investment plan in Texas. Furthermore, to avoid crippling the local power grid, Google emphasized that it has reached agreements with utility providers to add approximately 7800 megawatts (MW) of green energy generation and storage capacity to the Texas grid.
AI's energy consumption sparks complaints: Sam Altman counters with "human energy consumption theory," Elon Musk looks to space.
Google's commitment to resource control comes at a time when backlash against data center construction is at its peak in communities across the United States. A growing number of people are concerned that these data centers will not only drive up electricity prices and exacerbate global greenhouse gas emissions, but will also drain local freshwater resources at an alarming rate.
However, faced with fierce criticism from the outside world, the attitudes of tech giants varied greatly:
• Sam Altman's strong counterattack:OpenAI CEO Sam Altman recently publicly refuted these concerns, even stating that the allegations about water consumption were "fake news," and offered a highly controversial defense: "Training a 'human' also requires consuming a lot of food."
• Elon Musk's interplanetary dream:Another group of Silicon Valley opinion leaders, including Tesla CEO Elon Musk, reiterated their idea of "moving data centers directly into space," attempting to solve the water, electricity, and land disputes on Earth once and for all. However, many environmental experts are pessimistic, believing that the potential environmental damage from frequent heavy rocket launches could have far more devastating consequences.
Analysis of viewpoints
In this public relations battle over the "AI resource crisis," Google, OpenAI, and Elon Musk have demonstrated three completely different approaches.
Sam Altman chose to use philosophical sophistry (comparing AI training to the energy consumption of human survival) to withstand external attacks; Elon Musk chose to use his forte, "first principles" and science fiction vision (space server room) to shift the focus; while Google offered a pragmatic, but also the most technically challenging, solution: building a server room in Texas, the most water-scarce state, where you don't drink water.
From an engineering perspective, Google's move is quite bold. Currently, high-end AI servers (such as racks equipped with NVIDIA Blackwell or AMD Helios platforms) generate extremely high levels of heat, and the mainstream industry trend is towards more efficient liquid cooling. Google's move against this trend by pushing for advanced air cooling means they must achieve significant breakthroughs in data center heat dissipation channel design and airflow guidance technology to suppress the raging heat from AI chips without relying on water resources.
If this air-cooling solution proves successful in hot Texas, it will undoubtedly set a new standard for green data centers worldwide.
