Data centers are highly resource-intensive, and their water use is a subject of growing concern and debate, especially with the rapid expansion of artificial intelligence (AI).

To understand the scale of the issue and why it warrants close attention, it helps to separate the raw numbers from the broader context.

1. How Much Water Do Data Centers Use?

The water footprint of a data center is split into two categories: direct use (evaporative cooling to keep servers from overheating) and indirect use (the water consumed by power plants to generate the massive amounts of electricity the data centers pull from the grid).

• The Big Picture: Researchers at the Lawrence Berkeley National Laboratory estimated that U.S. data centers alone consumed roughly 228 billion gallons of water. Of that, about 17 billion gallons were sucked up directly for onsite cooling, while the remaining 211 billion gallons were tied to electricity generation.

• On a Facility Level: A typical data center consumes around 300,000 gallons of water a day (roughly what 1,000 homes use). However, massive "hyperscale" facilities can guzzle up to 5 million gallons a day—matching the water footprint of a city of 50,000 people.

• The "Per-Query" Cost: You may have seen headlines claiming a brief ChatGPT session or an AI-written email "uses a bottle of water". Tech companies like Google and OpenAI argue it is much lower—closer to 5 drops or a fraction of a teaspoon per query. The reality depends heavily on where the data center is located, the local climate, and what kind of energy grid it relies on.

2. Put in Context: Is it the Biggest Threat?

While those numbers are massive, data centers are not the primary driver of water scarcity on a macroeconomic scale. To put things in perspective, everyday human activities consume vastly more water:

Data centers use less water nationally than what Americans lose to leaky pipes or use to water their lawns.

3. Why We Absolutely Must Continue to Watch It

Even though data centers aren't the largest water consumers globally, we absolutely need to monitor them closely. The issue isn't the aggregate global number; it's the localized impact and the staggering trajectory of AI growth.

The Localized "Soda Straw" Effect

Data centers don't spread their water usage evenly across the globe; they pull it intensely from single, specific watersheds. Localized building booms mean data centers act like a "giant soda straw" dropping into a local community's water basin. Shockingly, roughly two-thirds of the data centers built since 2022 have been located in water-stressed or drought-hit regions like Arizona and Texas. When a massive facility opens in a drought-prone area, it directly competes with local residents and agriculture for scarce drinking water.

The AI Boom Trajectory

AI models require much more computational muscle—and therefore generate much more heat—than standard cloud storage or web searches. Projections indicate that global data center water demand could hit 5 billion cubic meters annually by 2027. A United Nations University report warned that by 2030, the global water footprint of AI data centers could equal the basic annual domestic water needs of 1.3 billion people.

The Carbon vs. Water Paradox

Tech companies are under immense pressure to achieve "net-zero" carbon emissions, prompting a massive shift toward renewable energy. However, environmental scientists have pointed out a troubling paradox: some "green" energy choices (like certain forms of bioenergy or nuclear power) require drastically more water to generate than fossil fuels. If we evaluate tech sustainability solely based on carbon footprints, we risk severely worsening local water crises.

The Path Forward

The tech industry is actively pivoting to address this. Many newer facilities are transitioning to closed-loop cooling systems (which recycle the same water repeatedly and cut freshwater use by up to 70%) or utilizing treated wastewater instead of pristine drinking water. Major operators are also pledging "water-positive" goals, meaning they intend to replenish more water into local basins than they consume. Keeping a watchful eye on these facilities ensures that companies are held accountable to these promises and that local infrastructure isn't overwhelmed.  Unquote.

You’ve hit on a fascinating quirk of international bureaucracy!

The reason CPV doesn't change to CBV comes down to how the International Organization for Standardization (ISO) handles country codes, mixed with a little bit of historical linguistics.

1. Where does "CPV" actually come from?

The code CPV isn't actually a typo for Cabo Verde; it is an abbreviation of the country's French name: Cap-Vert (or Cap-Verde).

When the ISO 3166-1 standard was first established in 1974, many international codes were heavily influenced by French, which was (and is) a primary diplomatic and administrative language for international bodies like the UN.

2. The Golden Rule of ISO Codes: Stability

In 2013, the government of Cabo Verde officially requested that all countries stop translating their name into English (Cape Verde) or French (Cap-Vert), declaring that the Portuguese autonym Cabo Verde should be used globally. The UN and the ISO updated the country’s official name in their records accordingly.

However, changing a country's actual code is incredibly disruptive. The ISO intentionally avoids changing code elements unless absolutely necessary because those two- and three-letter codes are deeply embedded in:

• Banking and currency networks (SWIFT)

• Shipping, logistics, and customs databases

• Machine-readable passports and international border control systems

• Domain names (.cv)

If the ISO changed CPV to CBV, thousands of software systems worldwide would suddenly break or require manual updates.

3. Other Countries in the Same Boat

Cabo Verde isn't alone in this. Because of the ISO's strict rules on keeping codes static, several countries have names that no longer perfectly align with their codes due to language differences or historical name changes:

• Switzerland: Uses CHE and CH because its codes are based on its Latin name, Confoederatio Helvetica.

• Germany: Uses DEU and DE because it's based on the endonym Deutschland.

• Spain: Uses ESP and ES (España).

• Czechia: Kept CZE and CZ even after transitioning away from the name Czech Republic.

So while the name Cabo Verde is now the official global standard, the ghost of Cap-Vert lives on in every passport and airport routing tag as CPV.　unquote.

ドラマチックなことはまるでありませんね。ちなみに、アフリカで10以上の国が植民地から独立国になったのが1960年です。「アフリカの年」と呼ばれます。ですが、ポルトガルは海外の領土をかたくなに維持しました。3，4世紀に渡って植民地支配から莫大な利益を得てきたのでそれは理解できなくはありません。しかし、20世紀に入ってからは、香辛料の取引における支配力ははどんどん縮小し、もはやポルトガルが独占的利益を得ることはできなくなっていました。それよりも海外領土を維持するための軍事費などの支出の方がはるかに大きくなっておりました。調べるとCPVの独立は1975年です。その15年の間、ポルトガルは抵抗してきたのです。他にもマカオ、ゴアなどの植民地があちこちにありました。なお、ブラジルの独立は1822年ですね。

ちなみに、ポルトガル、スペインは両方とも20世紀の途中まで民主主義国ではなく、独裁統治下にありました。統治者個人の好みが政治に強く反映されますので、植民地の独立を認めるか認めないかは独裁者が決めますが、経済的に赤字になる植民地経営を維持する経済的理由はありませんね。非経済的理由ということになります。過去の栄光でしょう。

サウジアラビアのコードはKSAです。SAはわかりますね。ではKとは？これはkingdomのことです。House of Saudが統治する国家ですので、国名にkingdomがあっても違和感はありませんね。万が一将来民主化されることがあれば、その時にはISOはKを取ることに同意するのでしょうかねえ。気になります。