Overview
Per China Daily on October 20, Inner Mongolia has become the world’s largest region by new-type energy-storage capacity, leading the build-out of a “green reservoir.” The flagship is the Zhongjin Data zero-carbon computing base in Ulanqab—China’s first “source-grid-load-storage” demo directly supplying green power to a data center: 300,000 kW of wind/solar paired with 45,000 kW of storage, with an intelligent platform linking wind, solar, storage and compute in millisecond real time, lifting the data center’s green-power share to 85%.
To feed gigawatt-scale AI clusters, Inner Mongolia’s parks are rolling out “incremental distribution grid + direct green power” zero-carbon models, holding direct green supply above 80% via centralized renewables and long-duration storage. Per Science and Technology Daily, by mid-October Ulanqab’s operational compute reached 190,000 P (up 112% YoY), with intelligent compute over 95% and renewable capacity at 20.29 million kW (67% green). Storage is shifting from “ancillary” to the “energy base” of AI.
Background & Interpretation
1. Industry Background and Policy Context
“Compute ends at power.” In the global AI compute race, Western giants turn to gas turbines or small nuclear plants; China leverages its world-leading grid, renewables and storage to pioneer “compute-electricity coordination.” Inner Mongolia’s cheap green power, ~4.3°C average temperature (near 10 months of free cooling) and low-latency Beijing-Tianjin-Hebei links make it the top relocation destination.
2. Core Drivers and Underlying Mechanisms
Storage is the linchpin. The Zhongjin base “computes more when wind is strong, storage fills the gap when nature pauses,” guaranteeing zero-flicker training while absorbing curtailment. Duration is shifting from ~2.5h (end-2025) to a mainstream 4h, with 8h long-duration moving from demo to commercial—enough to cover the entire evening peak.
3. Market Structure and Industrial Chain Impact
For storage and renewable firms, Inner Mongolia’s scale means large orders and system-solution demand. For cloud and AI firms, stable cheap green power (¥0.32–0.36/kWh, 30%–50% below eastern cities) is a core edge. The China Industrial Association of Power Sources expects 25%–30% annual storage growth through 2030. The competitive high ground has moved from cell shipments to “reliable system solutions + scenario fit.”
Implications & Outlook
Short-to-medium term, compute-electricity coordination scales from demos; direct green-power and source-grid-load-storage approvals accelerate (Ulanqab has cleared 14 direct-green-power projects). Watch long-duration storage cost curves, power-price stability, and the climb from “selling electricity” to “selling tokens” (Inner Mongolia already launched a green-compute token-trading platform).
Takeaways for Industry Participants
Storage and power-equipment makers should focus on system solutions and extreme-condition reliability. Cloud/AI firms should treat direct green power as a siting core metric and build multi-energy redundancy. Industrial parks can copy the “distribution grid + direct green power” zero-carbon model. Investors should note compute-electricity coordination as a high-certainty cross-sector theme.
This column compiles industry information and shares technical perspectives; it does not constitute investment advice.