Envision's Galaxy Campus Delivers 120 MW in Under a Year
Envision's Galaxy AI campus in Inner Mongolia delivered 120 MW in under a year, per SemiAnalysis. The "fifth of US costs" claim remains behind a paywall, and key figures conflict.
By Nathan Brooks
4 min read

What's News
- Envision completed the first 120 MW of its Galaxy AI campus in Ulanqab in under a year, per SemiAnalysis (published November 7)
- The full campus is planned for up to 800 MW of IT capacity across four phases by end-2027
- SemiAnalysis claims first-phase delivery at "a fifth of US costs"; the per-MW figure is paywalled
- Chinese data center capex has fallen ~25% over five years to under $3 million per MW, per SemiAnalysis estimates
- The 800V DC solid-state transformer deployment is a 5 MW pilot in one hall, not campus-wide
Envision Energy completed the first 120 MW of its Galaxy AI data center campus in Ulanqab, Inner Mongolia, in less than a year, according to research firm SemiAnalysis, which visited the site in September and published its findings on November 7. The full campus is planned for up to 800 MW of IT capacity across four phases by the end of 2027 — roughly the scale of the ~820 MW of critical IT capacity at the Stargate campus in Abilene, Texas.
SemiAnalysis says the first phase was delivered in August 2026 at "a fifth of US costs," though the detailed cost breakdown sits behind its paywall, and the per-megawatt figure was never published. The firm estimates Chinese data center capex has fallen about 25% over five years, to under $3 million per MW. That's a model output, not an audited benchmark — a distinction practitioners should keep in mind.
What does the campus actually include?
The primary source is SemiAnalysis's report, titled "We Visited China's Stargate in Inner Mongolia." It describes a project vertically integrated from wind generation to rack-level power electronics. Envision owns the campus and:
- Operates two dedicated wind farms totaling 420 MW of installed capacity
- Supplies four-hour battery storage through affiliate AESC
- Designs its own low-voltage power modules, solid-state transformers, and prefabricated substation
The campus runs under China's "source-grid-load-storage" model on a dedicated transmission line rather than the public grid. Envision targets more than 60% of the first two phases' 240 MW load from its wind farms, sited roughly 80 kilometers apart to reduce correlation during low-wind periods. Ulanqab's wind capacity factor runs 35–40%, and all-in electricity cost in the region is approximately $0.055/kWh, per SemiAnalysis.
The numbers that don't reconcile
Several figures in the public record conflict. SemiAnalysis describes a four-phase plan to 800 MW of IT capacity. Envision's own August commissioning announcement said the campus was designed to scale beyond 2 GW, and a November release gave 3 GW. The 800 MW covers IT load only; the larger numbers may reflect total facility power or different boundaries. The sources don't reconcile them.
The milestone itself differs too. SemiAnalysis says 120 MW came online in August 2026; a November 5 Caixin report, citing an Envision executive, described a 200 MW first phase handed over in late October. It's unclear whether the two use the same power definition or mark different acceptance stages.
All four phases are built-to-suit for two internet companies SemiAnalysis did not name. Their identities and lease terms are undisclosed, so the revenue certainty behind the buildout can't be assessed independently.
Does vertical integration win?
Envision has installed more than 100 GW of wind turbines globally and manufactures its own battery cells through AESC. SemiAnalysis notes that factory pre-configuration of the full power chain can compress on-site commissioning from 3–6 months to under three for the substation.
Whether that translates into a durable cost or schedule advantage is open. Hyperscalers still choose vendors mainly on resources and track record, SemiAnalysis says. The Chinese market is brutal: rents have roughly halved over five years to about $40/kW/month excluding power. Owning more of the value chain differentiates — but hasn't yet proven decisive.
Power reliability: managed, not solved
A four-hour battery shifts surplus from windy hours into the next calm period; it cannot cover multi-day lulls. SemiAnalysis's modeling shows wind share rising from 40% at 1:1 wind-to-load to about 55% with 1.75x oversizing, higher still with storage and better load matching. Envision targets over 60% for the first two phases. Diesel generators cover emergencies, per a China Securities Journal site report in October. No source provides a measured hourly renewable share.
The 800V DC story deserves similar caution. In one 5 MW hall, solid-state transformers convert 10 kV AC directly to 800V DC for server racks, eliminating low-voltage transformers, UPS units, and rectifiers. Envision said on November 2 the system runs live AI workloads, with a larger deployment planned in later phases. But it's a 5 MW pilot — not campus-wide — and no independent efficiency or reliability measurements exist. Nvidia's roadmap includes AC-compatible 800V racks in the second half of 2026, row-level power centers in 2027, and facility-scale medium-voltage-to-800V DC later.
Cooling follows a design common in northern China: liquid- and air-cooled servers share one water loop with no central chiller plant, enabled by local climate and permitted water temperatures. No independently measured annual cooling-efficiency result is available.
The bigger pipeline
Inner Mongolia's announced data center pipeline totals around 25 GW, including several gigawatt-scale campuses, per SemiAnalysis's China Datacenter Model, which tracks over 1,000 facilities across 70+ operators. The Chahar High-tech Zone alone had nearly 30,000 construction workers on site across 22 data center projects at the time of the visit.
What does this change for operators? Envision is presenting a competing model for AI infrastructure — vertically integrated, renewables-anchored, and fast. Whether it transfers beyond Inner Mongolia's specific regulatory, geographic, and supply-chain conditions is a question the current evidence doesn't answer.
Source: epyhia.com
More from Nathan Brooks
Show full bio
News editor covering regulation, policy and their impact on entrepreneurship, startups and venture funding at Business Bearings.
1,794 articles