What happened
Nvidia has teamed with US infrastructure company Firebird and Kazakhtelecom to build “Data Center Valley” in Kazakhstan, billed as the region’s largest AI‑computing hub. The initial rollout is said to cost $10 billion and could eventually host up to 5 GW of server capacity, according to statements made by then‑Nvidia CEO Jensen Huang during an October visit to Astana. The project is planned near the coal‑fired Ekibastuz GRES‑1 plant and would receive a fixed, low‑price electricity contract for 20 years. Critics, including geospatial analyst Marat Beksultanov, argue the venture is mostly financed by the Kazakh government, lacks private investment, and may struggle to attract European customers because of its reliance on coal power and opaque contractual arrangements. Similar AI‑data‑center ambitions are emerging in Uzbekistan, Tajikistan, Kyrgyzstan and Turkmenistan, each facing their own financing, energy‑supply and transparency hurdles.
Nvidia, Firebird and Kazakhtelecom disclosed plans for a $10 billion AI data‑center complex called Data Center Valley in Kazakhstan. The project is positioned as the largest AI infrastructure effort in Central Asia and is intended to leverage the country’s abundant natural‑resource base.
During a televised interview with Kazakh Deputy Prime Minister Zhaslan Madiyev, then‑Nvidia CEO Jensen Huang said the initial phase could reach 1 GW of capacity, with long‑term ambitions of up to 5 GW. He framed the initiative as a way for Kazakhstan to convert its energy surplus into “intelligence” for export.
Geospatial analyst Marat Beksultanov, CEO of GeoBox, raised three primary concerns: (1) the project is predominantly financed by the Kazakh government with little private capital; (2) the power source is the nearby coal‑fired Ekibastuz GRES‑1 plant, which could deter European customers subject to carbon‑footprint reporting; and (3) the government has not published the power‑price contract, lease terms with Firebird, or the source of the 20‑year electricity discount.
Beksultanov also highlighted Kazakhstan’s electricity balance: in 2025 the country generated 123.1 billion kWh but consumed 124.6 billion kWh, covering the shortfall with imports from Russia. Adding a multi‑gigawatt data‑center could exacerbate this deficit and increase the risk of grid instability.
Parallel AI‑data‑center initiatives are unfolding across the region: Uzbekistan’s Digital Uzbekistan 2030 strategy offers tax and customs exemptions for data‑center investors, with Saudi‑based DataVolt planning a $3 billion, 12 MW liquid‑cooled facility in Tashkent; Tajikistan announced an AI hub under Nvidia’s Global AI Nations program; Kyrgyzstan is courting Chinese firms such as Inspur and Huawei; and Turkmenistan is exploring proposals from Singapore’s Bitdeer Technologies despite its aging internet infrastructure and high ambient temperatures.
Why it matters
The project illustrates how AI infrastructure expansion is moving beyond traditional tech hubs into resource‑rich but less transparent regions. If successful, Kazakhstan could become a major exporter of AI , reshaping global AI supply chains and potentially lowering costs for AI developers worldwide. However, the heavy reliance on coal‑generated electricity conflicts with growing ESG expectations in Europe and the United States, where cloud‑service buyers must report carbon footprints. The lack of disclosed financing and power‑price contracts raises governance concerns that could deter foreign investors and technology firms wary of corruption risks. Moreover, the announced capacity—up to 5 GW—would significantly increase regional electricity demand in a country already importing power to cover a modest shortfall, potentially straining the grid and prompting outages. The broader Central Asian push, mirrored in Uzbekistan’s tax‑exempt data‑center incentives and Kyrgyzstan’s courting of Chinese hardware firms, signals a regional race to capture AI compute markets, but also highlights systemic challenges around energy sustainability, funding transparency and regulatory oversight that could limit the long‑term viability of these projects.
The scale of the Kazakhstan project—potentially 5 GW of AI —could make the country a significant node in the global AI hardware supply chain, offering an alternative to data‑center locations in the United States, Europe and East Asia.
European and North American cloud providers are increasingly required to disclose the carbon intensity of the they purchase. A coal‑adjacent data center may be unattractive, limiting the market for Kazakhstan’s AI services and reducing the likelihood of securing high‑value contracts.
The opaque financing and lack of publicly available contracts raise governance and corruption concerns, especially given Kazakhstan’s 96th ranking on Transparency International’s 2025 Corruption Perceptions Index. This opacity could deter foreign investors and technology firms that demand clear legal and financial frameworks.
The projected electricity demand of up to 5 GW would represent a substantial increase relative to Kazakhstan’s current net import of electricity. Without additional generation capacity or renewable integration, the project could strain the national grid, leading to reliability issues or higher electricity costs for other consumers.
The broader regional trend of AI‑focused data‑center development underscores a strategic shift toward capturing AI as a commodity. However, the same challenges—energy sustainability, financing transparency, and regulatory clarity—are recurring across Central Asian states, suggesting that the success of these initiatives will depend on how each government addresses these systemic issues.
Interactive Mechanism: How It Actually Works
Explore the underlying technology behind this development interactively.
Which component of an AI application is the machine-learning model itself?
What to watch next
Future announcements on actual power‑supply contracts, private‑sector investment commitments and customer sign‑ups will indicate whether the Kazakhstan hub can overcome its financing and ESG hurdles. Monitoring European cloud‑service providers’ procurement policies for AI will reveal if they will consider a coal‑adjacent data center. Additionally, policy developments from the United Nations Economic Commission for Europe and national regulators on data‑center water and electricity usage could shape the regulatory environment for Central Asian AI infrastructure.
Release of detailed power‑purchase agreements and lease contracts for Data Center Valley, which would clarify the cost structure and potential subsidies involved.
Announcements of private‑sector investors or technology partners committing capital to the Kazakhstan project, indicating market confidence beyond state backing.
Statements from major European cloud providers (e.g., AWS, Azure, Google Cloud) about willingness to source from Kazakhstan, especially in light of ESG reporting requirements.
Policy updates from the United Nations Economic Commission for Europe or regional bodies concerning water and electricity usage limits for data centers in Central Asia.
Progress reports from Uzbekistan, Tajikistan, Kyrgyzstan and Turkmenistan on their own data‑center projects, which could either reinforce the regional AI hub narrative or expose further feasibility gaps.