Water and capital liquidity are inexorably linked. Water underpins most value chains, not unlike energy: food, power, industry, transport, and urban development are historically founded on the assumption of water in abundance. However, water is no longer in abundance, particularly in landlocked Eurasia, on both shores of the Caspian Sea. The Caucasus and Central Asia are amongst the most affected regions in the world, facing rising heat, increasing evaporation, melting glaciers, and declining rivers. There is less water to fulfill greater aspirations.

As water declines, aspirations turn into choices that can be achieved, but not all at once. To address the question of how declining water levels affect the capital liquidity available to invest, Caucasus Watch turns to Sobir Kurbanov, an international development professional with over 25 years of experience working with the World Bank Group and the International Monetary Fund, holding roles at both headquarters and in field assignments across Central Asia, Eastern Europe, and the broader Eurasia region. In addition to his policy development work, Sobir has teaching experience in economic policy, industrial policy, and public policy analysis at institutions such as the OSCE Academy in Bishkek and the George Washington University’s Elliott School.

From the Black Sea to Central China you have a series of landlocked regions, many of whom are looking towards a joint development perspective. Armenia and Uzbekistan are moving to invest in data centers; we are told about “closed loop” water systems. Is there a global precedent and is it possible to invest in datacenters without an abundance of water resources? 

Yes. Water scarcity does not necessarily preclude investment in data centers, and there are already technological and policy models that enable such facilities to operate with substantially lower freshwater requirements. However, for water-stressed countries such as those in Central Asia and parts of the South Caucasus, water availability needs to be an integral part of the investment decision, rather than treated simply as an engineering issue. This is particularly relevant as Armenia and Uzbekistan consider further investment in data-center infrastructure. 

Recent World Bank work provides a useful framework. The Bank emphasizes that data centers create both direct water demand, primarily for cooling, and substantial indirect water demand from electricity generation. In fact, indirect water use associated with power generation can account for 80 percent or more of their overall water footprint. The problem is highly location-specific: globally, data centers use relatively little water compared with agriculture, but a concentration of facilities within a water-stressed basin can create significant competition with households, agriculture, and other industries. 

The World Bank therefore proposes a watershed-informed investment model. This means assessing water availability and economic value before deciding where to locate large data centers. It includes watershed-informed siting, transparent monitoring and disclosure of industrial water use, more efficient cooling technologies, water recycling and reuse, and incentives that encourage conservation. Importantly, the Bank argues that governments should consider these issues when designing incentives to attract data-center investment, rather than offering subsidies or tax incentives without assessing the longer-term implications for local water security. 

Closed-loop systems are an important part of the solution. Cooling water can be recirculated rather than continuously replaced with freshwater, and treated wastewater and other non-potable sources can substitute for drinking-quality water. Other technologies can reduce the amount of water required for cooling. The broader principle is water circularity, as data centers should increasingly be designed as part of a system in which water is recovered, treated and reused rather than withdrawn, used once and discharged. The World Economic Forum analysis highlights closed-loop systems, wastewater reuse and advanced cooling as practical responses to rapidly increasing AI-related water demand. 

Malaysia offers a particularly relevant example. Rapid growth in data centers in the state of Johor has increased pressure on local freshwater resources, prompting greater attention to alternative water-supply models. The World Bank’s Scaling Water Reuse work highlights a partnership with the state-owned Johor Special Water to develop what is expected to be Malaysia’s largest recycled-water supply facility for data-center campuses. This demonstrates a potentially useful model for Central Asia and the South Caucasus: rather than allowing new data centers to compete with households and agriculture for scarce potable water, governments can link digital investment to treated wastewater reuse, closed-loop cooling, and dedicated recycled-water infrastructure. Properly structured, this can also mobilize private investment in wastewater treatment and reuse while reducing the water footprint and long-term climate vulnerability of data-center development.

This approach is particularly relevant to the Caspian region. Countries such as Armenia and Uzbekistan should therefore assess proposed data-center investments not simply in terms of electricity availability, land and broadband connectivity, but also against basin-level water availability, competing agricultural and urban demand, future climate scenarios and the opportunity cost of water. Where water stress is high, project approval could require higher water-use efficiency, closed-loop cooling or the use of treated wastewater.

There is also a potentially positive development impact. Investment in data centers does not necessarily have to compete with improvements in water security. Properly structured, large private investments could contribute to wastewater treatment, recycling infrastructure, digital water monitoring and improvements in local utility systems. The World Bank similarly points to AI itself as potentially useful for detecting network leakage, improving irrigation scheduling, forecasting droughts and floods, and optimizing water reuse.

An important consideration, however, is cost. Incorporating the true long-term costs of water, climate resilience, and more water-efficient technologies will inevitably increase the upfront capital requirements of many mining projects. Closed-loop water systems, recycling and treatment facilities, more efficient tailings management, alternative water sources, and climate-resilient infrastructure all carry additional costs. The key question is whether investors are prepared to factor these costs into project economics from the outset. In my view, this should increasingly be part of standard investment appraisal rather than an optional environmental add-on. A project that appears financially attractive only because water is underpriced or future water scarcity is ignored may prove substantially more expensive—or even unviable—over a mine's 20- or 30-year operating life. The challenge for governments and investors is therefore to establish realistic water-pricing and risk-assessment frameworks while ensuring that the additional costs do not simply make otherwise viable critical-mineral investments unbankable. This may require appropriate risk-sharing, blended finance, and targeted IFI support, such as EBRD and IFC, and US DFC, for investments that substantially improve water efficiency and long-term resilience.

The discussion over energy-intensive data centers in the region brings to the discussion the ambition of developing nuclear capacity. The SMRs that can be delivered to market today rely on water for cooling reactors. Given that developing a nuclear plant take 7-to-10 years, are the assumptions we make for a landlocked region today safe for when the plant begins to generate power?

Not necessarily. This is one of the reasons why long-term infrastructure investment needs to incorporate climate scenarios rather than rely primarily on current hydrological conditions.

Nuclear energy presents a potentially important opportunity for Central Asia. It can provide stable, low-carbon baseload generation as electricity demand grows rapidly. Armenia, Kazakhstan and Uzbekistan also have major uranium resources, and additional reliable generation could support industrial development, critical-mineral processing, and potentially data centers. Nuclear generation could also complement the seasonal hydropower systems of countries such as Tajikistan and Kyrgyzstan. These potential benefits are reflected in the background analysis. 

At the same time, these are very long-term investments. Water availability for cooling therefore needs to be assessed not only against today's conditions but against possible conditions in 2040, 2050 and beyond. Projects should be stress-tested against higher temperatures, drought, changes in river flows and growing competition for water from cities and agriculture.

There are other considerations as well. Much of the region is seismically active, which requires very rigorous site selection and safety standards. Nuclear projects also involve substantial upfront costs and long-term technological relationships with external suppliers. SMRs may eventually reduce some of these constraints, but many designs still have relatively limited commercial operating experience. 

Mining in the Caucasus and Central Asia is attractive not only due the abundance of critical mineral resources but also because the region is sparsely populated. As mining is water intensive, are banks and investors willing to commit capital without knowing whether the water required is in place? 

For well-structured international investment, water availability increasingly needs to be treated as a material project risk rather than simply an environmental issue.

This is particularly important for critical minerals because mines and processing facilities can operate for decades and some processes are highly water intensive. An investor therefore needs to understand not only whether sufficient water exists when the project starts, but the reliability and economic cost of that water throughout the life of the investment.

There is also an important financing issue in the Caucasus and Central Asia. Many domestic financial institutions remain relatively underdeveloped in long-term project-finance capabilities. Bank portfolios have traditionally been concentrated in shorter-term commercial lending, while mining requires appraisal of geological, commodity-price, environmental, infrastructure and political risks over much longer periods. Water and climate risk add another layer to this challenge. 

Azerbaijan has joined he C5 to potentially create a single pathway from China to the Black Sea. Billions have been spent in infrastructure in Azerbaijan, Turkmenistan, and Kazakhstan, not least in port infrastructure around the Caspian Sea. Do you see a threat of “stranded assets” as the Caspian Sea evaporates? 

Yes, although I would define the risk somewhat more broadly. The immediate concern is not necessarily that major ports suddenly become unusable, but that declining Caspian water levels progressively increase the cost of operating and adapting infrastructure.

For ports, this could mean more dredging, changes to navigation channels, limitations on vessel loads, and eventually modifications or relocation of some facilities. This matters because Azerbaijan, Kazakhstan and Turkmenistan are investing substantially in infrastructure supporting the Trans-Caspian or Middle Corridor. 

These investments have long economic lives. Infrastructure being constructed today may still be operating in 2050 or beyond. Climate and water risks should therefore be incorporated systematically into feasibility studies and economic appraisal.

The World Bank, EBRD and other IFIs increasingly incorporate climate risk directly into infrastructure project appraisal. This goes considerably beyond a conventional environmental assessment. The World Bank, for example, uses climate and disaster stress testing to examine how a project's economic returns its NPV, benefit-cost ratio and operating costs perform under different climate scenarios. EBRD similarly screens investments for physical climate risks such as water stress, drought, flooding, erosion and extreme temperatures and, where those risks are material, incorporates resilience measures into project design.

For Caspian and the Middle Corridor infrastructure projects, this approach could be particularly important. Instead of assuming one future Caspian water level, a port investment could be tested against several plausible scenarios of sea-level decline. The appraisal could estimate what each scenario means for dredging costs, navigable depth, vessel loading capacity, port throughput, maintenance requirements and potentially the need to modify or relocate infrastructure. Those impacts can then be translated into changes in operating costs, revenues and ultimately the economic and financial rate of return.

The Caucasus and Central Asia are amongst the most affected regions in the world from Climate Change. Is the sum of national policies adequate in responding to this regional challenge?

I do not think national policies alone are sufficient, because many of the principal climate risks in this region are inherently transboundary. At the same time, national policies themselves are often not sufficiently comprehensive, well-funded or effectively implemented. Governments face significant institutional and technical capacity constraints, as well as limited fiscal space to finance the scale of climate adaptation and mitigation investment required. Another major constraint is the poor bankability of many climate projects. In a relatively high-risk financial environment, with shallow domestic capital markets and limited availability of long-term finance, potentially viable green and adaptation projects often struggle to attract private investment. 

That said, there has been progress in several cases. Governments have adopted green-economy strategies, renewable-energy programs and climate commitments. There are also interesting practical examples, including the expansion of renewable energy across Central Asia and Dushanbe's rapid transition toward electric taxis. At the same time, governments face difficult development trade-offs. They need affordable electricity, industrialization, infrastructure and employment, while many existing development models remain dependent on fossil fuels and other carbon-intensive industries. 

The larger problem is institutional and regional. Climate change affects systems that cross national borders. Glacier retreat in Tajikistan and Kyrgyzstan affects downstream water availability. Hydropower operations influence irrigation. Drought affects regional agricultural markets. Changes in the Caspian Sea affect several countries simultaneously. And Afghanistan's increasing use of Amu Darya water adds another dimension to regional water management.

National climate strategies therefore need to be complemented by much stronger regional coordination. This is particularly important for water, energy and infrastructure because the physical systems involved are inherently cross-border. In Central Asia, for example, upstream hydropower decisions in Kyrgyzstan and Tajikistan affect downstream irrigation and water availability in Kazakhstan, Uzbekistan and Turkmenistan. Climate-driven changes in glaciers, snow accumulation and river flows therefore cannot be managed effectively through national adaptation strategies alone.

A first priority should be common hydrological and climate information. Countries need compatible systems for monitoring river flows, glacier and snow conditions, reservoir levels, drought and extreme-weather risks, together with agreed mechanisms for exchanging this information. Without a common evidence base, governments can be making investment and water-allocation decisions using different assumptions about the same river basin.

Second, there needs to be stronger basin-level coordination of water allocation, reservoir management and infrastructure investment. Central Asia already has institutions such as the ICWC and IFAS, so the priority should be strengthening and modernizing existing government-endorsed mechanisms rather than creating parallel structures. Cooperation also needs to extend below the national level—to basin authorities, irrigation systems, utilities and cross-border facilities—because many practical water-management problems occur at this level.

Third, the water-energy nexus needs to be managed explicitly. An upstream reservoir may be economically valuable for winter electricity generation, but downstream countries may need the same water for irrigation in the summer. Climate change makes these trade-offs more difficult as hydrological patterns become less predictable. Regional planning should therefore increasingly integrate electricity generation and trade, reservoir operations, irrigation demand, and urban water requirements rather than addressing each sector separately.

Finally, countries need more consistent approaches to climate-risk appraisal for major regional infrastructure. A port on the Caspian, an HPP, an irrigation system, a mine, or a data center may be financed nationally, but its climate and water risks can be regional. World Bank and EBRD-type climate stress testing could therefore be applied more systematically to major cross-border investments, using common scenarios for water availability, drought, extreme heat, and future sea levels in the Caspian. This would allow governments and investors to compare risks more consistently and incorporate adaptation costs into project design and economic appraisal.

The financing dimension is equally important. Many national climate strategies identify substantial investment needs yet have limited fiscal resources and weak mechanisms to mobilize private capital. Regional coordination can help create larger, more bankable investment programs, supported by common technical standards, clearer water pricing and cost-recovery frameworks, better utility governance, and appropriately structured PPPs or blended-finance mechanisms. That is where institutions such as the World Bank and EBRD can add considerable value not only by providing financing but also by linking climate adaptation to institutional reform and investment bankability.

International financial institutions can play an important role here. The World Bank, EBRD and others can help translate climate strategies into economically viable investments by bringing together climate-risk assessment, infrastructure planning, institutional reform, PPP frameworks and private financing.

Interview conducted by Ilya Roubanis for Caucasus Watch