🔍 Read the full analysis: How Electricians Are Meeting The Power Demands Of AI Data Centers on ThorstenMeyerAI.com
Get the latest gadgets 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 show how data center operators are addressing rising AI electricity demand: Vantage reserved more than 1 GW of transformer factory capacity, and NiSource signed a 20-year power deal with Microsoft that includes about 1,100 MW of generation and a 600 MW battery. The shift toward higher-voltage DC distribution and on-site storage is intended to manage dense, variable loads, but claimed efficiency gains and the costs to electricity customers remain unsettled.
Data center power infrastructure is drawing new investment as AI workloads strain conventional designs: Vantage reserved more than 1 gigawatt of factory capacity for Heron Power solid-state transformers on October 21, and NiSource signed a 20-year power deal with Microsoft five days earlier that adds about 1,100 megawatts of generation and a 600-megawatt battery. The deals point to a growing focus on transformers, direct-current distribution and storage, as access to electricity becomes a constraint on AI data center expansion.
Heron Power’s solid-state transformer is designed to convert grid electricity directly into 800-volt direct current, with battery storage integrated into the system. Vantage’s reservation is for manufacturing capacity, not confirmation that the full amount of equipment has been installed or that it will supply a particular data center, amid the broader US data center power puzzle. The source material does not give a delivery schedule or the specific sites involved.
The NiSource-Microsoft agreement is a separate development in the data center sector. The source says it runs for 20 years and adds around 1,100 MW of generation alongside a 600 MW battery. It does not identify the generation technology, provide project costs, or explain how expenses and benefits will be allocated among the companies and electricity customers.
Industry designs are also changing at the rack level. NVIDIA’s reference design would convert 13.8-kilovolt grid power to 800-volt DC at a facility’s perimeter and distribute it to high-demand racks. NVIDIA says its design 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. These are company engineering claims, not independently verified results. 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
Why Power Design Is Changing
AI servers are concentrating more electricity demand into each rack, putting pressure on equipment designed for less dense computing. The International Energy Agency says AI server power density increased 11-fold from 2020 to 2025 and is projected to quadruple again by 2027. It estimates a rack could then draw as much power at peak as 65 households. That comparison describes peak demand, not typical continuous use.
Higher-voltage DC distribution could reduce the number of conversion stages between the grid and computing equipment, while batteries may help buffer rapid changes in demand. Whether the resulting savings justify new equipment depends on real-world performance, installation costs and the design of each site. The practical constraint is not only how much electricity a data center uses, but whether utilities and equipment suppliers can deliver the needed power and hardware on time.
The consequences extend beyond operators. The IEA projects global data center electricity use will rise from 485 terawatt-hours in 2025 to 950 TWh in 2030. In the United States, Berkeley Lab estimates data centers used 4.4% of electricity in 2023 and could reach 6.7% to 12% by 2028. Those projections make decisions about new generation, grid upgrades and cost allocation relevant to businesses and households as well as technology companies.
Products Worth Considering
Jackery HomePower 1000 Plus v2 Portable Power Station
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Bose S1 Pro+ Portable PA Speaker
As an affiliate, we earn on qualifying purchases.
From AC Conversion to 800-Volt DC
Most data centers use alternating current from the grid, stepping the voltage down and converting power through equipment such as transformers, uninterruptible power supplies and distribution units. Servers ultimately use direct current, so power is converted again inside or near the equipment. Each conversion can lose energy, much of which becomes heat that cooling systems must remove.
Traditional rack designs were built around loads of roughly 10 to 20 kilowatts. The source says emerging AI racks may approach a megawatt, increasing the amount of copper and power equipment needed to deliver electricity at lower voltages. NVIDIA has said a 1 MW rack supplied at 54 volts could require up to 200 kilograms of copper busbar. That figure is from the company’s technical material, rather than an independent measurement.
Operators may not replace every existing system at once. A Utility Dive article published October 26, sponsored by inverter maker SMA America, describes a “sidecar” conversion cabinet that could let 800-volt racks operate in halls designed around AC. Its model estimates losses of about 80 kilowatts when delivering 1 MW through a traditional chain, compared with about 20 kW in an 800-volt DC setup. Those figures are a vendor-sponsored model, not reported field measurements.
“Up to 5% better end-to-end efficiency”
— NVIDIA
Costs, Delivery and Measured Savings
The announced capacity reservations and power agreement do not establish when the equipment and generation will be available, or which data centers will use the new systems. The source material does not provide project-level costs, contract terms beyond the 20-year duration, or a breakdown of who will pay for associated grid infrastructure.
Efficiency figures also need careful interpretation. NVIDIA’s savings estimates and SMA’s modeled conversion losses are not presented as independently measured operating results. Actual performance will depend on site design, equipment, workload and cooling. The source does not provide results from large-scale deployments of the proposed 800-volt systems.
Grid access remains a major unknown for proposed projects. 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 roughly 2,300 GW of projects in U.S. interconnection queues at the end of 2024, while only 13% of requests from earlier queue cohorts had been built. Those figures describe queues and past outcomes, not a forecast that every current proposal will proceed.
Deployment and Rate Decisions
The next test is whether reserved transformer capacity and announced generation translate into equipment installed and power delivered. NVIDIA has said full-scale production of its 800-volt design is planned alongside Kyber systems in 2027. Operators and utilities will also need to show how batteries perform under real data center loads and whether they reduce grid strain or simply shift when electricity is drawn.
Regulators and utilities face decisions about how to pay for generation, transmission and distribution upgrades. The source reports that the average U.S. residential electricity price is 18.21 cents per kilowatt-hour in 2026, up from 17.30 cents in 2025, and that wholesale prices in the PJM region are up 41%. It does not establish that data centers caused those changes. Future rate cases and contract details will help clarify how new power demand affects customers in different regions.
Key Questions
What did Vantage announce?
Vantage reserved more than 1 GW of factory capacity for Heron Power solid-state transformers. The announcement concerns manufacturing capacity; the source does not say that this amount has been delivered or installed.
What is 800-volt DC power for data centers?
It is a proposed way to distribute electricity at higher-voltage direct current closer to AI racks, reducing some conversion steps and the amount of copper needed. The efficiency and material savings cited in the source are company claims or vendor estimates, not independently verified operating results.
What does the NiSource-Microsoft deal include?
The source describes a 20-year agreement adding about 1,100 MW of generation and a 600 MW battery. It does not specify the generation technology, project costs or how expenses will be divided.
Will these projects raise household electricity bills?
The source does not establish that the announced projects will raise household bills. Costs depend on contract terms, utility regulation and who pays for new generation and grid upgrades; those details are not provided.
Source: ThorstenMeyerAI.com
