🔍 Read the full analysis: Why AI Data Centers Are Turning To Electricians, 800-Volt DC, And Batteries on ThorstenMeyerAI.com
Get movie nights 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 October developments highlight how AI data center growth is shifting attention from chips to electricity infrastructure: Vantage reserved more than 1 gigawatt of factory capacity for an 800-volt DC transformer, and NiSource agreed to supply Microsoft with about 1,100 megawatts of generation and a 600-megawatt battery. The technology and contracts address power delivery and flexibility, but grid connection delays, costs and the eventual effect on electricity bills remain uncertain.
Vantage reserved more than 1 gigawatt of factory capacity for a solid-state transformer designed to convert grid power directly to 800-volt direct current, while NiSource signed a 20-year agreement to provide Microsoft with about 1,100 megawatts of generation and a 600-megawatt battery. The October deals reflect a shift in AI data center planning toward power equipment, storage and grid supply as computing facilities demand more electricity, a trend explored in capacity questions facing U.S. data centers.
The Vantage reservation, reported as made on October 21, is for factory capacity associated with a solid-state transformer from Heron Power. The equipment is intended to convert incoming grid electricity into 800-volt DC, which can then serve high-demand data center systems. A reservation of manufacturing capacity is not the same as confirmation that the equipment has been delivered, installed or brought into service; the source material does not provide those milestones.
On October 16, NiSource signed a 20-year power deal with Microsoft that adds about 1,100 megawatts of generation and a 600-megawatt battery, according to the source report. The agreement illustrates a separate part of the response: securing power supply and storage alongside data center development amid growing power bottlenecks for AI data centers. The reported figures describe the planned deal, not measured output from facilities already operating.
The engineering case for 800-volt DC is tied to rack density. The International Energy Agency says AI server power density rose 11-fold from 2020 to 2025 and is projected to quadruple again by 2027. NVIDIA’s proposed design converts 13.8-kilovolt grid power to 800-volt DC near the facility perimeter, then distributes DC to high-power racks. NVIDIA says this could use 45% less copper than 415-volt AC distribution and improve end-to-end efficiency by up to 5% compared with current 54-volt systems. Those are vendor estimates, not independent measurements.
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’s Pace
For data center operators, the constraints are increasingly about getting electricity to a site and distributing it inside the building, not simply buying computing hardware. Higher rack loads require heavier conductors and more electrical capacity; a design that reduces conversion steps or copper could affect construction space, energy losses and cooling demand. Whether the proposed efficiencies hold in deployed facilities remains to be demonstrated independently.
Storage could help manage rapid changes in demand as AI workloads start and pause. The IEA says batteries may become important for handling these swings and estimates that 20–25 gigawatts of batteries could be installed in data centers by 2030. It also says storage could support the wider grid if incentives are in place. That is a forecast, not a commitment by operators to build that amount.
The consequences extend beyond operators. New generation, grid connections and transmission equipment require investment, while costs and benefits may be allocated differently across utilities, data center customers and other ratepayers. The source material says responsibility is being settled contract by contract; it does not establish how any particular project will affect household bills.
Popular right now
iRobot Roomba Max 715 Pet Robot Vacuum with AutoEmpty Dock
Dyson Car+Boat Cordless Handheld Vacuum
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
As an affiliate, we earn on qualifying purchases.
From AC Conversion to DC Racks
Most data centers have historically received alternating current, stepped it down through transformers and backup-power equipment, and converted it to low-voltage DC inside servers. Each conversion can lose energy, which becomes heat that cooling systems must remove. The existing approach has served lower-density racks, but the power needs of newer AI systems are pushing operators to reconsider where and how conversion happens.
NVIDIA’s reference design targets 1-megawatt racks and a planned rollout with Kyber systems in 2027. NVIDIA says supplying that much power at 54 volts could require up to 200 kilograms of copper busbar. The company’s proposed higher-voltage distribution aims to reduce conductor demands. Separately, an October 26 Utility Dive article sponsored by inverter maker SMA America described a “sidecar” conversion cabinet as a possible bridge: it would let 800-volt racks operate in facilities built around AC systems. SMA’s modeled energy-loss comparisons are sponsor-provided claims, not independent test results.
Power availability is a broader bottleneck. The IEA says transformers can take two to three years to arrive and grid connections can take five to ten years in many places. Berkeley Lab reported about 2,300 gigawatts of generation and storage capacity in US grid queues at the end of 2024, while only 13% of earlier queue requests were ultimately built. Queue totals represent proposed projects, not power that is ready to serve data centers.
“NVIDIA says its design can move 85% more power through the same conductor size.”
— NVIDIA, in its 800 VDC technical material
Deployment, Costs and Grid Effects
The reports do not establish when Vantage’s reserved transformer capacity will translate into delivered equipment or operating data centers, or how much of the NiSource-Microsoft power arrangement is already built. Details such as project locations, construction schedules and the terms governing battery dispatch are not provided.
It is also unclear how 800-volt DC systems will perform across different facilities once installed. NVIDIA’s efficiency, copper and maintenance figures are projections from the company; SMA’s loss estimates come from sponsored modeling. Independent comparisons under real operating conditions are not included in the source material.
Finally, the deals do not settle who will pay for new generation and grid upgrades, or whether batteries will reduce peak demand and costs for other customers. The effect on household electricity bills will vary by utility and region and cannot be inferred from the announced capacity figures alone.
Projects Must Reach Operation
The next milestones are equipment production and delivery, construction of the associated power facilities, and confirmation that the planned generation and storage can serve data center demand. NVIDIA has said full-scale production of its 800-volt design is planned with Kyber systems in 2027; actual deployment and independently measured performance will show whether the projected benefits materialize.
For utilities and customers, attention will turn to grid connection approvals, construction timelines and contract terms that define who bears infrastructure costs. Further project disclosures and operating data should clarify whether batteries are used mainly for backup, to smooth data center load swings, or to provide services to the grid.
Key Questions
What happened in the Vantage and NiSource deals?
Vantage reserved more than 1 gigawatt of factory capacity for a solid-state transformer designed for 800-volt DC. NiSource signed a 20-year agreement with Microsoft that includes about 1,100 megawatts of generation and a 600-megawatt battery.
Why are data centers considering 800-volt DC?
Higher-powered AI racks need substantial electricity. Distributing power at a higher voltage can reduce current for a given power level and may reduce conductor requirements and conversion losses. The specific savings cited by NVIDIA are company estimates and still require independent validation in operating facilities.
Why add batteries to a data center power system?
Batteries can provide backup and may help manage rapid changes in electricity demand from AI workloads. The IEA forecasts 20–25 gigawatts of batteries in data centers by 2030, but that is a projection rather than a confirmed buildout.
Will these projects lower or raise household electricity bills?
The reported deals do not establish how costs will be allocated or what effect they will have on local rates. Bill impacts depend on the utility, grid upgrades, project contracts and regulatory decisions.
Source: ThorstenMeyerAI.com
