Envision has deployed what is described as the world’s first commercial 800V DC power system for AI data centers, located at its Ulanqab campus in Inner Mongolia. The system uses solid-state transformers to convert 10kV AC grid power directly to 800V DC for server racks, reducing conversion stages and supporting high-density AI workloads.
Overview
Envision Energy has announced the commercial operation of an integrated 800-volt direct-current power architecture at an artificial intelligence facility in Inner Mongolia, a deployment the company describes as a world first for this configuration. The site is the Galaxy AI campus in Ulanqab, where Envision says trials of its Dubhe controller and supporting energy model began in August 2026. The announcement positions the system as a departure from conventional data center power designs that typically pass incoming electricity through multiple conversion stages before delivering power suitable for servers and accelerators.
The significance lies less in the voltage number alone than in the integration: Envision is presenting a packaged power path from medium-voltage grid connection to rack-level DC distribution, with storage and fast control folded into the same architecture.
Technical Architecture
At the core of Envision’s design is a solid-state transformer that converts 10-kilovolt medium-voltage alternating current directly into an 800-volt DC distribution line. In a conventional facility, power would ordinarily be transformed and rectified across several stages before reaching information-technology equipment. Envision states that its approach reduces the number of conversion steps, improving electrical efficiency and allowing power equipment to occupy substantially less physical space within data centers.
The architecture also incorporates battery storage intended to respond rapidly to processor demand changes and to provide temporary backup during unexpected interruptions. Envision claims the system can provide immediate support over seconds and maintain backup capability for several minutes. A controller coordinates the transformer and batteries within milliseconds, aiming to manage sudden fluctuations in electricity consumption from graphics processing units.
The Ulanqab Deployment and Galaxy AI Campus
The installation is located at the Galaxy AI campus in Ulanqab, Inner Mongolia. Envision says the facility is expected to grow to 3 gigawatts of capacity, and that its eventual scale could serve one million AI accelerators. Those figures are company projections rather than independently confirmed operating data, but they indicate the intended class of deployment: not a small pilot rack, but a campus-scale AI power environment.
The choice of Inner Mongolia is consistent with the article’s description of renewable-powered computing. Envision intends to introduce the technology through Mission Gobi, an initiative seeking 5 gigawatts of renewable-powered computing capacity by 2030. That program focuses on desert locations, where Envision expects renewable generation and energy storage to support large computing developments.
Efficiency, Materials, and Space Claims
Envision cites reduced material usage as a key advantage. A higher-voltage distribution network requires less copper for cabling, a factor the company says could lower costs as computing installations expand rapidly. The same higher-voltage DC approach is said to reduce the footprint of power equipment, freeing floor area for computing capacity. Envision also notes that the system can accommodate higher-density server racks, allowing operators to install more computing capacity within a given floor area.
These are meaningful claims for AI data centers, where power distribution and cooling infrastructure can consume space and capital that operators would rather devote to accelerators. However, the article does not provide quantified efficiency gains, copper savings, or floor-space reductions. The argument is therefore architecturally plausible but not yet evidenced in the public record.
Battery Storage and Controller Coordination
The integrated battery storage is intended to serve two roles: smoothing the fast-changing demand of AI processors and providing ride-through during interruptions. Envision claims immediate support over seconds and backup capability for several minutes. The controller’s millisecond coordination of transformer and batteries is central to that promise, because GPU workloads can produce substantial swings in demand that place additional pressure on electrical infrastructure supplying densely packed computing equipment.
This is a different value proposition from conventional uninterruptible power supplies or backup generators. The system is presented as an active participant in power quality, not merely a passive backup layer. If it works as described, it could reduce the stress that AI load volatility places on upstream grid equipment. The article, however, does not establish how the architecture performs under sustained full-load conditions or across facilities with substantially different hardware configurations.
Market Context and Industry Impact
The deployment arrives as AI data centers face growing scrutiny over power availability, grid interconnection delays, and the cost of electrical infrastructure. Envision’s pitch is that a simpler, higher-voltage DC path can improve efficiency and reduce material intensity at exactly the moment when scale matters. The commercial launch also signals that solid-state transformer technology is moving from demonstration into data center power distribution, at least in one vendor’s implementation.
For the wider industry, the most important question is whether the approach can be replicated across different hardware ecosystems and utility regimes. A 10-kilovolt to 800-volt DC architecture may suit a campus with a controlled grid interface and renewable generation, but data centers vary widely in incoming voltage, redundancy requirements, and server power shelves. The article does not address those variations, nor does it provide installation costs or savings compared with conventional power systems.
Future Outlook and Verification Gap
Envision’s roadmap ties the Galaxy deployment to Mission Gobi and to a projected 3-gigawatt campus serving up to one million AI accelerators. If those plans proceed, the architecture will be tested at a scale that can reveal whether integrated 800-volt DC distribution, solid-state transformation, and millisecond battery control deliver the promised operational benefits. The company has designed the system to work alongside renewable electricity sources, with batteries intended to balance supply and changing computing requirements.
Yet the announcement leaves a significant verification gap. Envision has not provided independently verified measurements quantifying efficiency improvements, installation costs, or savings compared with conventional power systems. Nor does it establish how the architecture performs under sustained full-load conditions or across facilities with substantially different hardware configurations. The deployment is therefore a commercial milestone, but the scale of its claimed advantages awaits detailed independent evidence.