MINEX Kazakhstan 2026, held in Astana on April 15-16, brought together roughly 1,000 participants from the mining industry, government ministries, diplomatic missions, financial institutions and media for the largest extractive-sector gathering in Central Asia. The forum’s theme, “Kazakhstan’s Mineral Resources: Reforming for Value in a Multi-Vector Reality,” framed two days of discussion around how the country positions its critical minerals endowment against a tightening global competition for supply. 

Much of the debate focused on the investment climate, royalty reformm and processing incentives. Less of it focused on the constraint that is becoming a first-order concern for the sector: water. The analysis below draws on a presentation delivered at the forum on April 16 by Nightingale Int. Fellows Vlad Paddack and Sobir Kurbanov on water security risks facing Central Asia’s mining sector and the broader industrial strategy built around it.

In January 2026, Kazakhstan broke ground on a one-gigawatt data center valley in Ekibastuz, with a planned $30 billion investment envelope and negotiations underway with Microsoft, OpenAI, G42 and Mubadala. A month later, the government approved the site of its second nuclear power plant on the shores of Lake Balkhash, alongside the first, with Russia’s Rosatom and China’s CNNC as lead contractors. 

Uzbekistan, meanwhile, has signed a $5 billion data center partnership with Saudi Arabia’s DataVolt for up to 500 MW of capacity, anchored in a new special regime for Karakalpakstan that offers electricity at five cents per kilowatt-hour. A separate $2 billion AI center involving NVIDIA, Freedom Holding and OpenAI was announced in November 2025. These projects represent the strategic direction Tashkent and Astana want to be known for: critical minerals, nuclear power, hyperscale compute, the industries of the next decade. 

Each of these industries is water-intensive, and each nation sits in a basin already under hydrological stress.

Three Forces Converging on the Same Basins

Central Asia’s water supply is shrinking, fragmenting and being repriced at the same time. Glaciers in the Tien Shan and Pamir ranges, which feed the Syr Darya and Amu Darya, have lost roughly 27 percent of their mass since the 1970s and are projected to lose at least another half of the remaining volume by the 2050s. Annual streamflow in the region’s two main river systems is projected to fall by up to 30 percent by mid-century. Feasibility studies for mining projects with 20-to-30 year operational horizons are therefore being written against a hydrological baseline that will not hold for the length of the mine.

Transboundary allocation is a second pressure point. Upstream Kyrgyzstan and Tajikistan hold the sources and the hydropower; downstream Kazakhstan and Uzbekistan depend on them for water supply. There is no binding regional agreement on seasonal allocation. The risk this creates for industrial users is specific and often mispriced. When drought occurs, Central Asian governments apply a priority order regardless of prior commercial agreements: human consumption first, agriculture second, industry last. A mining operator or data center holding a water use license or long-term supply contract does not automatically rank ahead of that hierarchy in a severe drought year.

The transboundary problem extends beyond Central Asia’s internal borders. China sits upstream of Kazakhstan on the Ili and Irtysh rivers, has built more than 40 reservoirs and 250 water intake facilities on the Ili alone, and has reduced its annual flow from 17.8 to 12.7 cubic kilometers over two decades. Kazakhstan’s Ecology Ministry has documented a 21.5 percent reduction in Irtysh basin runoff attributable to Chinese withdrawals. Beijing has declined to sign the U.N. Watercourse Convention, and Astana has so far absorbed the loss in order to protect the wider economic relationship. Ekibastuz, the location of the flagship data center valley, sits in the Irtysh basin.

Further south, Afghanistan’s Qosh-Tepa canal is already diverting an estimated 10 cubic kilometers per year from the Amu Darya, and Taliban officials moved in October 2025 to accelerate construction of a dam on the Kunar river, explicitly invoking a “right to water” following India’s suspension of the 1960 Indus Waters Treaty. The weaponization of water agreements as political instruments is becoming a pattern across a wider arc in Asia, and the pattern is spreading upstream of Central Asia’s most ambitious industrial projects.

The third force is regulatory. Kazakhstan’s new Water Code, in force since June 10, 2025, is the country’s first major water overhaul in more than two decades and arrived alongside amendments to the Subsoil Use Code. Together, they bar most subsoil operations on water fund lands and near drinking-water aquifers, require formal approval from the water authority for activities near any surface water body, and for the first time give regulators a statutory definition of “ecological flow” to restrict industrial withdrawals during drought. Fines have been materially increased.

Water Risk Is Transmitting Into Balance Sheets

Cost is part of the story but secondary. The primary risk is supply security, whether an operation can be guaranteed sufficient water for its full project life. International lenders including the EBRD, IFC and ADB now require demonstrated water supply certainty as a condition of financing, and if a company cannot show that water access is secured for the duration of mine life or facility life, financing is not available.

Where operations can be financed, water transmits into unit economics through four channels. Direct operating cost rises when surface supply becomes unreliable, forcing a shift to groundwater extraction, longer pumping distances, or trucking. In water-stressed mining jurisdictions globally, water accounts for 10 to 20 percent of total operating expenses in copper and iron ore segments. A 2024 World Bank assessment put climate-related water losses in Europe and Central Asia at around 1.7 percent of regional GDP, with the annual investment required to meet water security needs in Central Asia alone estimated at 2.2 percent of GDP. 

Unbudgeted capital expenditure is the second channel: when public water infrastructure cannot be relied on, operators invest in their own storage, treatment and recycling systems, and this rarely appears in original feasibility studies. 

Tailings compliance is the third. Tailings facilities are one of the largest water sinks in mining and the point of greatest environmental and regulatory exposure, and tightening requirements under Kazakhstan’s new Water Code, the Environmental Code and international lender standards raise the cost of getting this wrong. 

The fourth channel is energy cost amplification. Hydropower dominates electricity generation in Kyrgyzstan and Tajikistan, drought years reduce output and trigger rationing, and for energy-intensive mining, critical minerals processing, data centers and AI training clusters, a low-water year compresses margins directly. Water scarcity and energy insecurity are the same risk in this region, concentrated on the same sites.

Industrial Geography is Concentrating the Exposure

The strategic locations chosen for the region’s next-generation industries concentrate water risk rather than spread it. Ekibastuz sits in the Irtysh basin, already losing a fifth of its runoff to Chinese withdrawals, and the data center valley is planned to draw on coal-fired and eventually nuclear power, both water-intensive generating technologies. Karakalpakstan, where Uzbekistan is anchoring its data center hub, is the region most affected by the Aral Sea disaster. Authorities have acknowledged the water constraint by requiring low-water or water-free cooling in new facilities, which is the right response, but the energy inputs powering those facilities will still draw on water-stressed generation assets.

Kazakhstan’s two planned nuclear plants are both to be built on Lake Balkhash. Lake Balkhash is fed primarily by the Ili, the same Ili whose annual flow has been cut from 17.8 to 12.7 cubic kilometers by Chinese withdrawals. Two large pressurized water reactors alongside a third plant under consideration will add a significant new cooling demand to a lake already on a declining trajectory. The energy required to support Ekibastuz, if fed partly by the Balkhash plants, connects one basin under external stress to another facing internal drawdown, with the region’s flagship AI ambitions sitting on top of both.

Three Scenarios for the Next Decade

Three trajectories are plausible. 

In a managed adaptation scenario, companies model water in feasibility studies, invest in closed-loop systems, engage regulators early and access blended finance from development institutions, and costs rise moderately and predictably. 

In a reactive self-insurance scenario, operators invest in pumps, storage and backup generation ad hoc as shortages develop, costs rise faster than modeled, some projects turn marginal and financing terms tighten as lenders reprice risk they previously underweighted. 

In a climate shock and institutional drift scenario, a severe drought year coincides with a transboundary dispute and a regulatory enforcement push, electricity rationing hits processing operations and data centers, water access becomes subject to government discretion, some operations face suspension, and international investors reprice the entire Central Asian mining and infrastructure sector.

The difference between these scenarios is is whether water is treated as a first-order strategic variable before the shortage arrives, by operators, by lenders and by the governments that are currently writing commitments into the region’s growth story without a matching water security architecture.

The $30 billion at Ekibastuz, the $5 billion DataVolt commitment, the two Balkhash reactors and the wider critical minerals pipeline represent a coherent economic strategy. Whether that strategy survives contact with the region’s hydrology will depend on choices being made now, in project design, in lender conditionality and in whether Central Asian governments can convert summit-level cooperation into binding regional water arrangements that industrial users can price. Power can be generated. Water cannot.