GLP Buys an Option on Inner Mongolia's Power

GLP China signed a strategic cooperation agreement with the Ulanqab government on 10 July, a framework for multi-gigawatt hyperscale digital infrastructure across a planned 1.33 million square meters in Inner Mongolia, delivered in phases, that GLP presents as a scalable pipeline and a reinforcement of its long-term growth strategy in AI-ready capacity. A framework is priced by the terms it leaves open.

The Ulanqab text leans on an integrated energy model, presented as enabling efficient local consumption of renewable power and supporting high-density infrastructure for hyperscale and AI workloads. Inner Mongolia's renewable resources are the reason the framework exists, and holding one with the local government is closer to a queue position than to a building permit; consent is the product in every data center trade, and Ulanqab secures it up front, with the government cast as counterparty rather than as the permitting risk everything else has to clear.

The resources make a multi-gigawatt, high-density campus plausible, and a developer who can point to them can raise against the pipeline before an electron moves. Angela Zhao, GLP China's chief executive, described the agreement as combining data center development and operations, renewable energy integration, and capital formation—three capabilities in one package—and called data centers a core growth engine as demand from AI and cloud computing accelerates. Grid permission is the underwriting asset in this cycle: queue positions and interconnection rights price before the electrons they carry.

GLP's capital sits behind the third leg of Zhao's package. GLP Capital Partners, the asset-management arm, ranked second in PERE's 2026 APAC Fund Manager Guide, having raised roughly US$8.9 billion for APAC-focused real estate strategies from the start of 2021 to the end of 2025. That is a real estate franchise pointed at compute, and the Ulanqab framework is the clearest recent statement of where the firm wants its next pool of capital to work.

No disclosed anchor tenant, no offtake agreement, and no power price are attached to the 1.33 million square meters, and the announcement does not say who will consume the compute or who funds the phases. The blank price column has migrated from projects to fund managers, and merchant risk is now an ownership-level problem rather than only a developer's financing problem. Ulanqab fits that diagnosis: the framework is less a project than an option on Inner Mongolia's land and renewable capacity, and option value is not infrastructure value until a tenant signs.

That gap puts the framework on the speculative side of the pricing line. Hyperscaler-anchored digital assets get infrastructure treatment; pipelines without named counterparties compete for capital on a different curve. A phased, tenant-less framework, however large its intended footprint, is why GLP describes a pipeline and not a project. If the Ulanqab phases eventually reach a fund, limited partners should expect the early ones to carry merchant exposure rather than the anchored, low-risk profile a multi-gigawatt headline implies.

A contract would change the read. An anchor tenant, a disclosed offtake, or a first-phase energization date would let the megawatts price as contracted capacity rather than as ambition. Until one of those appears, the deal's most concrete content is the figure GLP chose to lead with: 1.33 million square meters of intended, phased, renewable-powered build in the province whose power makes it possible. The 10 July agreement bought the queue position; the next signing sets its price.