🔍 Read the full analysis: AI’s Power Shift: Electricians, 800-Volt DC, And Batteries on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Two power-sector developments highlight how AI data centers are changing their electricity plans: Vantage reserved more than a gigawatt of factory capacity for an 800-volt DC transformer, and NiSource agreed to supply Microsoft with about 1,100 MW of generation and a 600 MW battery. The deals point to growing demand for grid connections, power equipment and storage, but do not establish that 800-volt systems or batteries will lower electricity costs for customers.
Data center power infrastructure is becoming a central constraint for AI expansion: Vantage reserved more than 1 gigawatt of factory capacity for an 800-volt DC solid-state transformer on October 21, while NiSource signed a 20-year agreement with Microsoft that adds about 1,100 megawatts of generation and a 600 MW battery. The announcements show companies planning for much larger, more flexible power needs, though they do not by themselves show how much the equipment will improve efficiency or affect electricity bills.
The Vantage reservation concerns manufacturing capacity for a transformer designed to convert grid electricity directly into 800-volt direct current. The source material does not specify when the equipment will be delivered, where it will be installed, or how much of the reserved capacity is tied to a particular data center project. A reservation of factory capacity is not the same as a completed installation or operating power supply.
NiSource’s agreement with Microsoft is a separate development. Reported on October 16, the 20-year power deal adds about 1,100 MW of generation and a 600 MW battery. The available information does not identify the generation technology, project locations, construction timetable or the precise terms for allocating costs. The agreement therefore cannot establish whether the new resources will reduce pressure on other utility customers.
The engineering case for 800-volt distribution is partly based on reducing the number of power conversions between the grid and AI servers. NVIDIA’s reference design calls for converting 13.8-kilovolt grid power to 800 V DC near a facility and distributing it to high-power racks. NVIDIA says its design could use 45% less copper than 415-volt AC distribution and deliver up to 5% better end-to-end efficiency than current 54-volt systems. These are vendor projections, not independently verified results from a broad deployment; full-scale production with its Kyber systems is planned for 2027.
Infographic 01 · From grid to rack
AI racks near a megawatt are outgrowing the AC power chain.
Why the industry is moving to 800-volt direct current: fewer conversions, less copper, batteries on the main path. The numbers are vendor figures with different baselines, so they cannot be added up.
Today: many conversions
Typical AC chain from the grid to the rack
- Medium-voltage AC from the grid
- Transformer steps it down to low-voltage AC
- UPS converts to DC for the batteries and back to AC
- Distribution units feed the racks
- Every server power supply converts AC to DC again
Each stage loses energy, and almost all of the loss becomes heat the cooling system must remove.
800 V DC: convert once, outside
NVIDIA reference design, full-scale production planned with Kyber racks in 2027
- 13.8 kV AC from the grid
- Converted once to 800 V DC at the edge of the site, batteries on the same DC bus
- 800 V DC busway to the racks
- Stepped down inside the rack, for 1 MW racks and beyond
Bridge for existing halls: a “sidecar” cabinet next to the rack converts AC to 800 V DC, while the old chain stays in place.
Power lost per MW delivered to the rack
kW, SMA white-paper model vendor model
What NVIDIA says 800 V DC changes
NVIDIA’s engineering claims claim
Sources: NVIDIA technical blog (20 May 2025, updated 31 Jul 2025); SMA America, sponsored content in Utility Dive (26 Oct 2026). Rack density: IEA, Key Questions on Energy and AI (Apr 2026) (AI server power density 11× from 2020 to 2025, a further 4× by 2027).
Powered by Thorsten Meyer AI
Power Equipment Sets AI Buildout Pace
The developments matter because AI data centers need not only chips but also the equipment and grid access required to supply them. Higher rack power can raise demand for conductors, transformers, cooling and electricity connections. If those systems take longer to obtain than computing hardware, power infrastructure can set the pace of a project.
Higher-voltage DC could reduce copper use and conversion losses in some facility designs, while batteries may help manage short-lived changes in demand and provide backup. But projected benefits are not the same as measured performance. The cost and reliability of an installation depend on its design, operating conditions, grid connection and how storage is used.
The impact reaches beyond data center operators. New generation, transmission and distribution work can affect local planning and utility spending. Whether households or other ratepayers pay any portion of those costs depends on project financing, utility regulation and contract terms; the supplied details do not settle those questions. National figures also mask local differences in grid capacity and electricity prices.
Products Worth Considering
LG UltraGear 27-Inch QHD OLED Gaming Monitor, 240Hz, 0.03ms, G-SYNC Compatible, FreeSync Premium Pro, HDMI 2.1
Apple MacBook Pro 2021 16‑inch with M1 Max, 32 GB RAM, 2 TB SSD, Space Gray (QWERTY English)
Logitech MX Vertical Wireless Ergonomic Mouse – Graphite
As an affiliate, we earn on qualifying purchases.
From AC Server Rooms to Dense AI Racks
Data centers have traditionally received alternating current and converted it several times before servers use low-voltage DC. Each conversion can waste some energy as heat. That arrangement was built for older rack loads, but the source material describes AI systems that concentrate far more power into a smaller space.
The International Energy Agency reported that AI server power density rose 11-fold from 2020 to 2025 and could quadruple again by 2027. It estimates global data center electricity use could grow from 485 terawatt-hours in 2025 to 950 TWh in 2030. These are forecasts, not confirmed future consumption. NVIDIA says a 1 MW rack supplied at 54 volts could require up to 200 kilograms of copper busbar, illustrating the challenge of moving very high power at low voltage.
Grid access is another constraint. The IEA says transformers can take two to three years to arrive and grid connections five to ten years in many locations. Berkeley Lab reported that the U.S. interconnection queue held about 2,300 GW at the end of 2024, while only 13% of earlier requests were ultimately built. Those figures describe a broad project backlog, not a measure of data center projects alone.
An October 26 Utility Dive article discussing 800 V DC was sponsored content from inverter maker SMA America. Its modeled comparison—about 80 kilowatts of losses for a 1 MW traditional AC chain versus about 20 kW for an external conversion and DC design—is a vendor-sponsored model, not an independent field measurement. SMA also describes a “sidecar” cabinet that could let 800 V racks operate in halls designed around AC equipment.
“Leaving no room for compute”
— NVIDIA, describing its 800 VDC design
Costs, Deployment and Bill Impacts
Several key details remain unavailable. The source material does not say when Vantage’s reserved transformer capacity will become equipment in service, or provide the project’s cost, site or expected performance. It also does not provide independent operating data for NVIDIA’s projected copper and efficiency savings.
For the NiSource-Microsoft deal, the generation source, construction schedule, exact battery role and cost allocation are not specified. It is also unclear whether the 600 MW figure refers to the battery’s power output alone; the battery’s energy capacity and duration are not provided. The agreement’s effect on local prices and other utility customers cannot be determined from the reported terms.
The IEA estimates that data centers could have 20–25 GW of battery capacity by 2030, but that is a forecast, not a confirmed buildout. The extent to which batteries smooth demand, support the grid or reduce the need for other infrastructure will depend on actual deployments and operating rules.
Projects Must Move From Plans to Power
The next test is whether reserved manufacturing capacity becomes installed equipment and whether data center operators publish performance from 800 V DC systems under real workloads. NVIDIA has said full-scale production with Kyber systems is planned for 2027; that is a stated plan, not confirmation that the systems will be deployed on schedule.
For the NiSource-Microsoft agreement, construction, permitting and project details will clarify when the added generation and battery can serve the arrangement. More information on the battery’s duration, dispatch and cost allocation will also help show how it fits into the utility system. Until those details emerge, the deals signal investment in power infrastructure but do not establish final costs, operating benefits or household bill effects.
Key Questions
What happened with Vantage’s transformer reservation?
On October 21, Vantage reserved more than 1 GW of factory capacity for a solid-state transformer designed to convert grid power to 800 V DC. The reported information does not say when or where the equipment will be installed.
What does the NiSource-Microsoft agreement include?
The 20-year agreement adds about 1,100 MW of generation and a 600 MW battery. The available details do not specify the generation technology, battery duration or cost allocation.
Why are data centers considering 800-volt DC?
Higher rack power can make low-voltage distribution require substantial copper and equipment space. NVIDIA says its 800 V design could reduce copper use and improve efficiency, but those figures are vendor claims and are not presented here as independently measured results.
Will the projects lower household electricity bills?
The reported announcements do not establish that. Effects on household bills depend on project costs, utility regulation, grid upgrades and how expenses are assigned among customers.
Are data center batteries already a standard grid resource?
The source material describes battery use as growing and cites an IEA estimate of 20–25 GW in data centers by 2030. That is a forecast; how batteries support the grid will depend on deployment, operating arrangements and incentives.
Source: ThorstenMeyerAI.com
