The water war does not look like a conventional war across most of the world. There are no tanks guarding every reservoir, nor a global exchange where a barrel of water is traded like oil. But that is precisely what makes it deceptive: water is becoming strategic not because it is scarce everywhere, but because it is local, difficult to transport, impossible to replace and needed simultaneously for food, energy, industry, cities and technology.
In 2026, that pressure is already visible. United Nations University has introduced the term “global water bankruptcy” for situations in which some river basins and aquifers have been pushed so far beyond their limits that drought, falling water levels and ecosystem degradation can no longer be treated as temporary episodes followed by a simple return to normal.
The World Resources Institute estimates that 25 countries, representing roughly a quarter of the world’s population, already face extremely high water stress every year. About 4 billion people experience water stress for at least one month a year, and by 2050 the share of the population exposed could approach 60%.
Water is therefore starting to resemble less a cheap utility from the tap and more a strategic infrastructure. Without it, farms, power plants, semiconductor factories, data centres, mines, refineries and cities cannot operate.
The water war begins with a simple reality: we cannot make freshwater wherever we want
The planet holds vast quantities of water, but most of it is saline. Usable freshwater is distributed very unevenly across space and time, while a large share is stored in glaciers, aquifers or ecosystems that cannot be exploited without limit.
Electricity can travel hundreds or thousands of kilometres. Gas can move through pipelines, or be liquefied and shipped by sea. Copper can be loaded into a container. Water is bulky and heavy: one cubic metre weighs about one tonne.
That is why a water-rich region cannot simply solve shortages in a region thousands of kilometres away. Desalination, pipelines and inter-basin transfers are possible, but they require energy, infrastructure and capital.
This geography makes water a profoundly local resource. A company may have access to global capital, globally sourced chips and electricity bought through international contracts, but a data centre must find water near the place where it is built.
Agriculture uses more water than any other sector
FAO estimates that agriculture accounts for about 72% of global freshwater withdrawals, followed by industry and services. It is the figure that immediately changes the discussion around the water crisis.
When a country faces drought, the first serious competition is not between a resident’s shower and a data centre. It is often between urban consumption, irrigation, energy production and the flows ecosystems require.
Irrigated farming produces a disproportionately large amount of food relative to the land area it covers, but in many regions it uses aquifers faster than they can recharge. In the short term, extra pumping sustains crops. In the long term, water levels fall, pumping energy costs rise and soils may become more vulnerable to salinisation.
FAO estimates that around 1.2 billion people live in agricultural areas facing severe water constraints. At the same time, by 2050 agriculture must produce substantially more food, feed and fibre for a global population approaching 10 billion.
Water is already built into food prices, even if it is not visible on the label
A kilogram of agricultural produce carries more than calories. It indirectly carries the water used to grow it — a concept known as “virtual water”. When a country imports cereals, meat or fruit, it also imports the water required to produce them.
This reality creates an invisible geopolitics. A highly arid country can save domestic water by importing agricultural products from a wetter region. Conversely, large-scale exports of highly water-intensive crops from an already stressed basin can amount to exporting a scarce local resource.
In future, agricultural advantage will depend not only on soil and climate, but also on water productivity: how much food is obtained from every cubic metre used.
Water is also becoming a raw material for artificial intelligence
Data centres turn electricity into computing power, but servers convert some of that electricity into heat. That heat must be removed.
Some data centres use evaporative cooling, which can be highly energy-efficient but consumes water. Others use closed-loop systems, air, dry cooling or direct-to-chip systems that reduce local water use.
In 2025, Google reported that its water stewardship projects had replenished about 4.5 billion gallons, equivalent to 64% of the company’s total freshwater consumption in 2024. In its 2026 environmental report, Google said replenishment had risen to 7.7 billion gallons in 2025, around 78% of total freshwater consumption.
These figures do not mean that AI alone “drinks” all that water. Google operates offices, infrastructure and data centres, and consumption depends heavily on location and cooling technology. But the scale of its replenishment programmes shows that water has become a strategic infrastructure issue for hyperscalers.
InfoBuz has separately examined how AI’s energy consumption is already reaching the scale of cities and putting pressure on power grids. Water is the less visible side of the same infrastructure.
Microsoft is seeking to build data centres that use no water for cooling
The pressure is not merely reputational. In water-stressed regions, access to water can determine whether a data centre receives approval at all, and how quickly it can expand.
Microsoft has announced a generation of data centres with direct-to-chip cooling and closed loops that can avoid water consumption for cooling. The company estimates that the new design can avoid about 125,000 cubic metres of water a year for each facility.
In Quincy, Washington, Microsoft and local authorities built water-reuse infrastructure that reduced the company’s potable-water use in the region by 97% and released about 1.5 million m³ annually for community needs.
This points to the industry’s likely direction: evaporative cooling may remain competitive in water-rich areas; in arid regions, access to recycled water, wastewater and non-evaporative cooling may become a technological advantage.
Chip factories turn ultrapure water into an industrial raw material
Semiconductors have an even more direct relationship with water. Manufacturing a modern chip involves hundreds or thousands of processing steps, and wafers must be repeatedly cleaned of particles and residues at microscopic scale.
Those operations use ultrapure water — water treated far beyond drinking-water standards, with almost all minerals, ions, particles and contaminants removed.
For 2025, TSMC reported total water use of about 151 million m³ across its fabs and subsidiaries. At the same time, it expanded its use of recycled water: its Taiwan fabs used more than 67,000 m³/day of reclaimed water, while cumulative recycled-water consumption had exceeded 23.5 million m³ by the end of 2025.
This is one reason a semiconductor factory’s location is not chosen solely on the basis of wages and energy. It requires a highly stable power grid, quality water, treatment infrastructure and a local authority able to guarantee these resources for decades.
Water may determine where the next AI factory is built
In the AI era, a competitive location must offer more than cheap electricity. It must answer three questions at once: is there enough power, is there enough water, and is there infrastructure to remove heat?
That could alter the investment map. A region offering cheap land and energy may lose a project if the local basin cannot support the additional consumption in drought years.
In February 2026, UC Berkeley published an analysis of data-centre water consumption in California, noting that the AI boom is rapidly increasing demand while public data on the precise consumption of many facilities remain insufficient.
Water is therefore becoming part of investment due diligence, just like grid access.
The Panama Canal shows how water can disrupt global trade without conflict
The Panama Canal uses freshwater to operate its locks. The same rain-fed lakes also supply water to a large share of the country’s population.
In September 2026, the Panama Canal Authority temporarily reduced the number of available daily slots because rainfall in the basin was below expectations. The measure affected both Neopanamax and Panamax vessels and was accompanied by adjustments to draught limits and the booking system.
It is an almost perfect example of water’s strategic value. Insufficient water in a single basin can alter shipping times for goods moving between oceans and raise logistical costs thousands of kilometres away.
In other words, there need not be a “water war” for water to have geopolitical effects.
Sixty per cent of global freshwater flows cross borders
UNESCO says transboundary waters account for more than 60% of global freshwater flows, and 153 countries have territory in a basin or aquifer shared with at least one other state.
That means a dam, irrigation project or water abstraction built in one country can affect the water available in another. Rivers do not observe political borders.
UN-Water lists 313 transboundary river and lake systems, as well as hundreds of shared aquifers. Yet institutional cooperation remains incomplete: many regions have no operational agreements covering an entire basin.
This is where the real geopolitical risk lies. Water does not automatically cause war, but shortages can intensify existing tensions between agriculture, energy, cities and states.
“Water wars” are rarer than headlines suggest
It is important to separate a viral scenario from actual history. Water can cause tensions and aggravate conflicts, but transboundary basins also generate extensive cooperation.
UNESCO stresses that agreements governing rivers, lakes and aquifers can create economic and political benefits large enough to make water a reason for cooperation rather than war.
The Danube is a European example. Dozens of uses — energy, navigation, water supply, agriculture and ecosystems — are coordinated through regional institutions and agreements precisely because no country controls the river on its own.
InfoBuz recently examined what would happen to Romania if the Danube remained at extremely low flows for an entire year. That scenario shows how quickly water scarcity can shift from a hydrological issue to an energy, agricultural and economic one.
Water is not the new oil — it is more complicated than oil
The comparison that “water is the new oil” is appealing, but imperfect. Oil is consumed and can be moved through a global market. Water participates in a natural cycle: it is reused, evaporated, infiltrated and returned to basins.
Moreover, the value of one cubic metre depends enormously on its location. A cubic metre in a wet basin does not have the same economic and social value as the same volume in a desert city during drought.
Water is closer to infrastructure than to a commodity. Its value comes from abstraction, storage, treatment, distribution, recycling and the right to use it in a particular place and at a particular time.
The most valuable water companies of the future may not “sell water”
The economic opportunity lies around water: desalination membranes, efficient pumps, leak-detection sensors, wastewater treatment, industrial reuse, precision irrigation, digital twins for networks and storage infrastructure.
Microsoft already uses AI to identify water losses in networks through acoustic analysis. Industrial companies are investing in zero-liquid-discharge plants, while chip factories are recovering an increasing share of their process water.
In this sense, the strategic “asset” is not always the right to a lake. It may be the technology that enables a factory to use the same water more than once.
Desalination can create freshwater, but not for free
For highly arid coastal countries, desalination radically changes the equation. Saudi Arabia, the United Arab Emirates, Israel and other states can produce freshwater from the sea on an industrial scale.
Reverse-osmosis technology has become much more efficient, but it requires energy, membranes, costly plants and management of the resulting saline concentrate.
Desalination is well suited to municipal water and certain high-value industrial uses. It is far more difficult to justify economically for the enormous volumes needed by low-value agricultural crops.
For that reason, the most efficient solutions will differ by sector: desalination for cities, reuse for industry, more precise irrigation for agriculture and protection of aquifers for drought periods.
Recycled water is becoming a premium industrial resource
Urban wastewater is among the most undervalued resources. It is already collected at a single point, and after advanced treatment it can be used for industry, irrigation, cooling or, in certain systems, even to recharge drinking-water reserves.
TSMC uses reclaimed water in Taiwan, Microsoft reuses water for data centres, and water-stressed cities are investing in schemes where treated wastewater replaces potable water in industrial uses.
The circular water economy has one straightforward advantage: every cubic metre reused is a cubic metre that does not have to be taken from a river or aquifer.
Network losses can be the “cheapest reservoir” a city builds
In many cities, a significant share of treated water is lost before reaching consumers through ageing pipes, faulty connections and inaccurate meters.
Reducing losses can be cheaper than building a new dam or desalination plant. Pressure sensors, smart meters and acoustic analysis can locate leaks without digging up an entire neighbourhood.
In a water-stressed world, network efficiency becomes, quite literally, a new source of water.
Climate change does not reduce water everywhere. It makes it less predictable
One of the most important points to understand is that global warming does not produce a uniform decline in rainfall across the planet. Some regions will receive more water, others less.
The economic problem is variability: longer droughts, intense rainfall that runs off quickly without recharging aquifers, less mountain snow naturally stored for summer, and higher temperatures that increase evapotranspiration.
WRI projects that global water demand could increase by about 20–25% by 2050, while the number of basins with high year-to-year variability may rise.
In other words, a reservoir built for a 20th-century climate may become less reliable in a climate where rainfall arrives differently.
$70 trillion in GDP could be exposed to water stress in 2050
WRI estimates that about $70 trillion, equivalent to 31% of projected global GDP, could be exposed to high water stress by 2050, compared with about $15 trillion in 2010.
That figure does not mean 31% of the world economy will stop. “Exposed” means economic activity is taking place in basins where demand is high relative to available supply.
For investors, however, exposure becomes a tangible risk: factory shutdowns, restrictions, higher insurance costs, mandatory investment in recycling and competition with urban demand.
Tunisia shows how water can become a macroeconomic issue
In September 2026, the World Bank estimated that Tunisia has only about 380 m³ of renewable freshwater per person per year, well below the threshold used for “absolute water scarcity”.
Without investment and reforms, the institution estimates that water stress could reduce Tunisia’s GDP by about 6.4% by 2050 and eliminate at least 30% of agricultural jobs.
The example matters because it shows that water does not remain an issue for the environment ministry. It becomes an issue of public finances, employment, migration, exports and social stability.
What “water bankruptcy” means in UNU’s 2026 report
United Nations University proposes an intentionally forceful term: water bankruptcy. The idea is not that the planet is running out of water, but that some regions are living off accumulated hydrological “capital”, rather than only the renewable annual “interest”.
If an aquifer is pumped faster than it recharges, its level falls. If a lake loses more than it receives, it shrinks. If soils and wetlands degrade, the basin retains less water for dry periods.
After long enough, a return of rainfall for one year does not restore the system to its original state. That is the report’s central point: in some places, a “temporary crisis” has become structural.
Water is one of the few assets whose low price can conceal enormous value
In many cities, users mainly pay for abstraction, treatment and distribution, rather than for the economic value of raw water. That model is necessary for social affordability, but it can send too weak a price signal to industrial or agricultural consumers.
In a water-abundant region, this problem may seem theoretical. During a severe drought, it immediately becomes real: which use takes priority — households, farms, power plants, factories or ecosystems?
A mature economy does not answer this question through price alone. It uses water rights, drought plans, ecological thresholds and priorities for essential consumption.
Romania is not a water-scarce country, but it is not immune
Romania has the Danube, the Carpathian Mountains, inland rivers, lakes and significant aquifers. Compared with many arid regions, its position is favourable.
But availability is not uniform. The south and east are more vulnerable to drought, agricultural irrigation infrastructure has not been fully modernised, and recent episodes have shown that the Danube can fall to extremely low flows.
Romania’s strategy need not be one of panic, but of infrastructure: reducing network losses, restoring irrigation systems, retaining water in soils and landscapes, reusing industrial water and planning abstractions for drought years.
Romania is also part of the Danube basin, one of Europe’s most important transboundary systems. The water entering at Baziaș has crossed several states before reaching Romania. Hydrological cooperation is, quite literally, foreign policy.
For further context, see Dama Solar Romania: one of Europe’s largest solar parks, 1.3 GW and €728 million, which explores an aspect directly connected to this subject.
Who will have the advantage in the water economy?
In the coming decades, the advantage will not necessarily belong to the countries with the greatest raw quantity of water, but to those that manage it most productively.
- Cities that lose less water through pipes will be able to grow without major new abstractions.
- Agriculture that obtains more output from every cubic metre will withstand drought more effectively.
- Industry that recycles water will be able to operate in basins where competitors must cut production.
- Data centres with near-zero-water cooling will be able to locate in more regions.
- Countries with functioning agreements on shared rivers will face fewer geopolitical risks.
The water war will above all be a war of efficiency
The century’s most likely competition is not an army conquering a lake. It is competition among cities, companies and countries to produce more with the same quantity of water.
A chip factory that reuses water, a data centre that uses no water for cooling, a farmer using soil-moisture sensors and a city reducing network losses all gain the same thing: additional economic capacity without an equivalent additional abstraction.
In this sense, water can become one of the century’s most important assets without being privatised and traded like oil. Its value will lie in the ability to produce food, energy, chips and computing in a world where the same resource must be shared among ever more users.
Frequently asked questions about the water war
Will the world run out of water?
No. Water constantly circulates through the hydrological cycle. The issue is the availability of quality freshwater, in the place and at the time when people and the economy need it.
Who uses the most freshwater?
Agriculture. FAO estimates that it accounts for about 72% of global withdrawals.
Does AI use a lot of water?
Data centres can use water directly for cooling and indirectly through electricity generation. The volume varies enormously by location and design; companies are investing in cooling that uses no water and in reuse.
Why do chip factories need water?
Semiconductors must be repeatedly cleaned with ultrapure water during manufacturing. Modern fabs invest heavily in its recycling and recovery.
Can wars over water occur?
Water can intensify tensions and conflicts, particularly in transboundary basins. Historically, however, shared resources have also generated numerous mechanisms for cooperation.
Is desalination the answer?
It is an important solution for cities and industry in arid coastal regions, but it requires energy and infrastructure spending and is not economical for every use.
Sources and transparency
- United Nations University – Global Water Bankruptcy, 20 January 2026
- World Resources Institute – global water stress and 2050 projections
- WRI Aqueduct 4.0 – dataset updated in September 2026
- FAO – SDG 6.4.2 Water Stress: 72% of withdrawals for agriculture
- Google – Environmental Report 2026, water and infrastructure
- Microsoft – Environmental Sustainability Report: data centres and water positive
- TSMC Annual Report 2025 – water consumption and reuse in semiconductor factories
- UC Berkeley CLEE – Regulating Data Center Water Use in California, February 2026
- Panama Canal Authority – restrictions due to reduced rainfall, 2026
- UN-Water – Progress on Transboundary Water Cooperation
- UNESCO – transboundary water cooperation and conflict prevention
- World Bank – Tunisia’s water scarcity and economic impact, 22 September 2026





Discuția cititorilor
Ai o opinie sau o întrebare despre acest articol? Spune-ne ce crezi.
Fii primul care deschide discuția.
Te rugăm să păstrezi un limbaj civilizat și să discuți la subiect. Comentariile spam sunt filtrate automat.