🔍 Read the full analysis: Electricians Step Into The AI Data Center Power Story 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 deals in October put power infrastructure at the center of AI data center expansion: Vantage reserved more than 1 gigawatt of transformer factory capacity, and NiSource signed a 20-year power agreement with Microsoft that includes about 1,100 megawatts of generation and a 600-megawatt battery. The developments arrive as operators examine 800-volt direct-current systems and storage, but efficiency figures cited for new designs are largely vendor claims, and the timing and cost of grid connections remain major constraints.
Data center power infrastructure moved into focus this month as operator Vantage reserved more than 1 gigawatt of factory capacity for Heron Power solid-state transformers, and Indiana utility NiSource signed a 20-year agreement with Microsoft that adds about 1,100 megawatts of generation and a 600-megawatt battery. The deals show how AI expansion is driving investment in electricity supply, conversion equipment and storage, even as independent evidence for some efficiency claims remains limited, a theme explored in approaches to transparent AI.
Vantage made its reservation on October 21. The source report says the capacity is for a Heron Power transformer designed to convert grid power into 800-volt direct current, amid broader data center expansion. A factory-capacity reservation is not the same as a confirmed delivery schedule or a report that the equipment is already operating at data centers; those details were not provided.
Five days earlier, NiSource and Microsoft signed a 20-year power deal. According to the source material, it adds about 1,100 megawatts of generation and a 600-megawatt battery. The reported figures describe the deal’s planned power resources; the material does not specify the generation technology, project locations, construction timetable or how costs will be allocated.
A Utility Dive article published October 26 raised the design question behind the deals: whether data centers built for conventional alternating-current distribution can keep pace with high-density AI racks. The article was sponsored by SMA America, an inverter maker. Its modeled comparison says an 800-volt DC arrangement could reduce conversion losses, but that is a vendor-backed model, not an independent measurement of performance across operating facilities.
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 Limits Shape AI Expansion
For data center developers, electricity access and equipment delivery can determine when a campus is built as much as the availability of processors. The source material, citing the International Energy Agency, says transformers can take two to three years to arrive and grid connections can take five to ten years in many locations. Those estimates point to a bottleneck that can delay new capacity even when companies have secured land and computing equipment.
Power demand also affects utilities and electricity customers. The IEA estimates global data center electricity use could rise from 485 terawatt-hours in 2025 to 950 TWh in 2030. Berkeley Lab estimates US data centers used 4.4% of national electricity in 2023 and could account for 6.7% to 12% by 2028. These are projections and estimates, not a guarantee of future consumption.
Who pays for new generation and grid upgrades matters to households and businesses. The source reports average US residential electricity prices of 18.21 cents per kilowatt-hour in 2026, up from 17.30 cents in 2025, and a 41% increase in wholesale prices in PJM. Those figures do not by themselves establish that data centers caused the increases. The effects and cost allocation vary by region and contract.
Popular right now
Roborock Robot Vacuum and Mop with LiDAR Navigation and 10,000 Pa Suction
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
KitchenAid Artisan 5-Quart Tilt-Head Stand Mixer, Almond Cream
As an affiliate, we earn on qualifying purchases.
From AC Distribution to 800-Volt DC
Most data centers receive power as alternating current, which is stepped down and converted through several stages before servers use low-voltage DC. Each conversion loses some energy, much of it as heat. That design was developed for lower-density server racks; the source report says AI server power density rose 11-fold from 2020 to 2025 and could quadruple again by 2027, citing the IEA’s April 2026 report.
NVIDIA has proposed converting 13.8-kilovolt grid power to 800-volt DC near a facility’s perimeter and distributing DC to high-power racks. The company says the 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. It also says a 1-megawatt rack supplied at 54 volts could require up to 200 kilograms of copper busbar. These are NVIDIA’s engineering claims, not independently verified results, and its full-scale production plan is tied to Kyber systems in 2027.
Existing buildings may adopt the technology incrementally. SMA describes a “sidecar” conversion cabinet that could let 800-volt racks operate in halls designed around AC distribution. Batteries are another part of the shift: the IEA projects 20 to 25 gigawatts of data center battery capacity by 2030, while noting that whether storage becomes a grid asset depends on incentives. Heron Power’s product is described as integrating storage with its transformer, but the source gives no deployment results.
“Leaving no room for compute”
— NVIDIA, in its 800 VDC technical blog
Costs and Deployment Timelines Remain Open
The available source material does not say when Vantage’s reserved transformer capacity will be delivered, how much equipment it covers beyond the stated capacity, or whether the units have entered service. It also does not provide project-level schedules or a breakdown of the costs and ownership arrangements in the NiSource-Microsoft agreement.
Performance comparisons require caution. NVIDIA’s savings figures and SMA’s modeled conversion-loss estimates come from companies with a commercial interest in the technology. The source does not cite independent, operating-site measurements that would show how the designs perform under different data center loads, cooling systems or grid conditions.
It is also unclear how much new generation and storage will be built specifically to meet data center demand, how utilities will recover associated network costs, and what share may fall on other customers. National price and demand figures do not resolve those local questions.
Watch Equipment Orders and Grid Plans
The next evidence to watch is whether Vantage’s factory reservation turns into equipment deliveries and operating installations, and how NiSource and Microsoft implement the generation and battery resources described in their agreement. The source material does not give dates for those milestones.
In parallel, operators and equipment makers are expected to test whether 800-volt DC distribution and batteries can support higher rack loads in new facilities and retrofits. Independent measurements of energy losses, reliability, maintenance costs and battery performance would help distinguish field results from vendor projections. Utility filings, construction schedules and rate decisions will show how new supply and grid costs are divided.
Key Questions
What did Vantage announce?
Vantage reserved more than 1 gigawatt of factory capacity for Heron Power solid-state transformers, according to the source report. It did not provide a delivery schedule or confirm that the equipment is operating at data centers.
What is included in the NiSource-Microsoft agreement?
The reported 20-year power deal adds about 1,100 megawatts of generation and a 600-megawatt battery. The source material does not identify the generation technology, locations or cost allocation.
Why are data centers considering 800-volt DC?
Higher rack power can make conventional low-voltage distribution bulky and increase conversion losses. An 800-volt DC design aims to move power through fewer conversion stages, but published savings figures cited here are vendor claims or models, not independent operating results.
Do these deals mean household electricity prices will rise?
Not by themselves. New data center demand can require additional generation and grid investment, but the effect on bills depends on local supply, utility rules and who pays for infrastructure. The figures cited in the source do not establish that data centers caused recent price increases.
Source: ThorstenMeyerAI.com
