Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Nvidia says a data center can devote no more than 70% of the electricity it draws from the grid to AI processing. In less efficient conditions, overhead can reach 40%, leaving only about 60% for useful AI work. The estimate comes from Nvidia and was reported by TechRadar on 14 October 2026; it is not an independently audited industry average.
In Nvidia’s one-gigawatt example, roughly 20% is consumed by cooling, power distribution and conversion, while another 10% is lost through misconfigured racks, failures and restarts. Those losses explain why a facility can have a large GPU fleet yet deliver substantially less computing than its grid connection suggests.
What the 60% and 70% figures mean
The two headline figures describe different efficiency conditions in Nvidia’s estimate, rather than a contradiction. TechRadar reported that Nvidia sees a maximum of 70% of grid electricity reaching AI processing. The same report says inefficiency can reach 40%, which leaves as little as 60% for AI work.
Quick wins for a faster PC:
Scan for outdated or missing drivers - takes under a minuteDriver Scan →Clear out junk files and repair common Windows errorsFree Scan →| Condition described by Nvidia | Electricity used for AI processing | Overhead | What the estimate includes |
|---|---|---|---|
| Main breakdown reported by TechRadar | Up to 70% | About 30% | Cooling, distribution, conversion, rack issues, failures and restarts |
| Low-efficiency case reflected in the headline | About 60% | Up to 40% | Nvidia’s broader estimate of maximum total inefficiency |
| Illustrative one-gigawatt allocation | About 700 megawatts | About 300 megawatts | Approximately 200 megawatts for cooling, distribution and conversion, plus 100 megawatts for rack, failure and restart losses |
The one-gigawatt row is arithmetic applied to Nvidia’s percentages, not a measurement from a named facility. Actual proportions will vary with workload, cooling design, power architecture, rack configuration and outage history.
#1 Best Overall
- POWER AND CHARGE: This rack mount power strip provides an additional 8 NEMA 5-15 outlets (120V/15A) and features a 6ft (1,8m) long cord so you can plug your devices in while leaving the rack mobile
- 1U RACK DESIGN: Compatible with all 19" server racks 4 inches or deeper, this horizontal-mount power distribution unit fits many network racks and has an integrated power cord; ANSI/EIA RS-310-D standard
- EASY INSTALLATION: This IT-grade rackmount PDU features a rugged steel chassis, LED indicators for ground and surge protection, and lets you control the power state with power and reset switches
- PROTECTS YOUR EQUIPMENT: This rack mountable 8-outlet (120V) power strip features a built-in circuit breaker and reset switch, ensuring a dependable performance of your networking equipment
- THE IT PRO'S CHOICE: Designed and built for IT Professionals, this rack PDU is backed for 2-Years, including free lifetime 24/5 multi-lingual technical assistance
Where the missing electricity goes
Cooling, distribution and conversion
About one-fifth of the power in Nvidia’s example is consumed before it can support computation. Cooling equipment removes heat from densely packed accelerators, while electrical distribution and conversion equipment moves grid power through the facility and changes it into the forms servers can use. These loads rise with the thermal and electrical density of modern AI racks.
Rack configuration problems
Nvidia attributes another part of the loss to incorrectly configured racks. A rack may be installed but unable to operate at its intended capacity because of power limits, cooling settings, network dependencies or other configuration constraints. Electricity committed to that infrastructure does not produce a matching amount of AI processing.
Rank #2
- Versatile and Space-Saving: This 1U Rack mount PDU features a compact design that allows for efficient use of space in standard 19-inch racks. 16 rear-facing plug outlets and three front-facing outlets provide ample connectivity for your devices
- Efficient USB Power: Featuring four USB ports, it enables simultaneous power supply to your favorite devices, ensuring convenience and productivity
- Built-in Circuit Breaker: The PDU is equipped with a built-in 12-Amp circuit breaker that protects against circuit overloads. This ensures reliable performance and helps prevent damage to your equipment
- Heavy-Duty Construction: The power distribution unit is designed with heavy-duty components and a sturdy metal housing for durability and long-lasting use
- Convenient Mounting: The PDU features mounting ears on the back panel for easy installation in a standard 19-inch rack
Failures and restarts
The report attributes about 10% of consumption in the one-gigawatt example to failures and restarts as well as rack configuration. “Ten percent of every watt” should be read as a facility-level share in this estimate, not as a claim that each individual GPU wastes 10% of its own energy. Restarting a large job can repeat work that has already consumed electricity, and repeated faults can leave powered equipment idle while operators recover service.
Why a 100-megawatt contract may deliver less
Rod Evans, Nvidia’s vice president of supercomputing, AI cloud infrastructure, and AI/HPC for EMEA, described the gap between a contracted connection and usable power: “You sign a contract for 100 megawatts, but you never actually get that. With a bit of luck, you’ll get 93 megawatts, and at certain times of the day, even less.”
Rank #3
- 10 NEMA 5-15R Outlets with Long Power Cord: This rack mount power strip provides 10 NEMA 5-15 outlets (15A) and features a 6ft long cord to conveniently plug in devices, keeping the mounting more easily and flexible
- Universal 19-Inch Rack Compatibility: Compatible with all 19-inch server racks 4 inches or deeper, this horizontal-mount power distribution unit fits many network racks and has an integrated 6ft/1.8m power cord
- Fireproof Heavy-Duty Construction: This server rack mountable power distribution unit features whole housing fireproofed construction that will support long-life working performance
- Built-In Circuit Breaker Protection: This rack mountable 10-outlet (AC100-240V) power strip features a built-in circuit breaker and reset switch, ensuring dependable performance of your power equipment and safety for using power
- Industrial-Grade Materials and Design: Features industrial equipment pure copper wire material for high power capacity, industrial-grade metal housing, and cord retention tray for enhanced durability
The shortfall can occur in the grid connection, on-site electrical conversion and distribution, or during operating conditions when equipment cannot run at its design limit. A contracted figure therefore describes available supply, not guaranteed AI compute.
Why facilities are built for peaks instead of normal workloads
Evans also said, “A data center is actually designed for peak load and outages, not for what it typically uses.” Designing for the worst expected load and for resilience leaves capacity that is necessary for spikes, maintenance and failures but sits underused during ordinary operation.
Rank #4
- Efficient Power Distribution: Delivers 12A (US Listed 12A) power through 8 NEMA 5-15R ports, ideal for server racks, network setups, and office environments
- Comprehensive Indicators: Monitors status with Power (yellow), Earth (green), and Lightning Protection (red) indicators, ensuring secure power for server rooms and warehouses
- Compact Design: Sleek 1U horizontal profile (19" x 1.7" x 1.7") maximizes rack space efficiency in IT cabinets
- Versatile Compatibility: 1 NEMA 5-15P input for seamless integration in diverse setups, providing flexibility for network installations
- Safety Assurance: UL listed for compliance, guaranteeing safety and quality in data centers and office spaces
Nvidia’s proposed response is to tap that unused capacity and revise power and cooling configurations. In practice, operators must preserve enough reserve for reliability while finding safe ways to increase the share of installed capacity doing useful work. Removing resilience margins indiscriminately could make outages more damaging, so capacity recovery has to be tied to measured thermal, electrical and workload limits.
The infrastructure bottleneck is moving to the grid
The reported constraint is no longer only whether enough AI accelerators can be manufactured. Vertiv EMEA president Paul Ryan described prospective sites where power, permits or practical equipment delivery are missing. A facility can have a ready customer and available GPUs yet remain unable to operate at scale because its grid connection, approvals or supporting equipment are not ready.
Best Value
- 【Heavy-Duty 9 Outlet PDU】 Designed for standard 19" server racks, this 1U rack mount power strip provides 9 US standard outlets (15A/125V/1875W), ideal for data centers, network cabinets, and audio-visual setups needing reliable power distribution.
- 【Individual Switch Control】 Each outlet is equipped with its own illuminated on/off switch, so you can manage connected devices individually instead of unplugging them. The switch modules are fully independent: if one outlet trips, only that outlet shuts down while all remaining outlets keep running normally — no whole-strip shutdown, no interruption to your other equipment. A tripped switch also tells you exactly which device has reached its load limit, giving you faster, more sensitive overload protection and a clear visual cue for troubleshooting.
- 【Overload Protection & Power Monitoring】 Equipped with overload protection and a digital power monitoring display, this PDU safeguards your equipment from overloads while providing real-time voltage and current data for secure operation. The switch will automatically trip if the current exceeds 15A. Simply having wires or cables touch the switch will not cause it to trip — the switch only responds to an overload condition.
- 【Durable Metal Construction】 Built with a sturdy metal housing and a 14AWG heavy-duty 6.5FT power cord, ensuring durability and stable performance even in high-demand environments like professional server rooms and industrial settings.
- 【Versatile Installation】 Ideal for studios, labs, and data centers, ensuring peak performance and reliability. Designed for 1U rackmount for hassle-free cable management. Supports horizontal installation in server racks with included mounting brackets.
Nvidia and Vertiv have reportedly collaborated since early 2026 on next-generation data-center designs. Vertiv supplies power and infrastructure-management systems for Nvidia GPU deployments. Their stated design aim, as quoted by TechRadar, is “to build sustainable AI data centers where the focus isn’t on using as much power as possible, but on getting the most out of every watt.”
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to assess an AI data center’s real efficiency
The percentages alone cannot rank facilities. A useful comparison should separate compute output from the infrastructure and reliability costs that support it.
| Measure | Question to ask | Why it matters |
|---|---|---|
| Useful-AI-work share | What fraction of metered facility electricity produces completed AI processing? | Shows the gap between grid demand and productive output. |
| Cooling and conversion overhead | How much power is used by cooling, distribution and conversion under the stated workload? | Identifies losses that occur even when GPUs are operating normally. |
| Rack and restart losses | How often are racks unavailable, jobs restarted or work repeated? | Captures the operational waste behind Nvidia’s 10% estimate. |
| Recoverable idle capacity | How much installed capacity can be used without violating thermal, electrical or outage limits? | Distinguishes genuinely stranded capacity from necessary reserve. |
| Grid-connection lead time | When will the site receive its dependable connection, and at what delivered capacity? | A contracted supply number may not equal usable power. |
| Permits and site logistics | Are approvals, transformers, switchgear and other equipment available? | Missing practical infrastructure can delay an otherwise complete project. |
| Resilience versus peak provisioning | Which capacity is reserved for outages and demand spikes? | Prevents efficiency claims from disguising a reduction in reliability. |
What the numbers mean for U.S. electricity demand
The issue is significant beyond individual facilities. The Congressional Research Service, summarizing Lawrence Berkeley National Laboratory, reported approximately 176 terawatt-hours of U.S. data-center electricity use in 2023, equal to about 4.4% of national electricity consumption that year. The same report notes projections reaching up to 12% of U.S. electricity use by 2028.
Those figures cover U.S. data centers broadly, not AI workloads alone. They also are not a conversion of Nvidia’s efficiency percentages into national savings. A credible estimate of avoided cost or emissions would require a facility’s actual load profile, electricity tariff and grid mix.
How to interpret Nvidia’s claim
- Treat 70% useful power and up to 40% overhead as Nvidia estimates reported by TechRadar, not universal measured averages.
- Read the 10% failure-and-restart figure as the share cited for Nvidia’s one-gigawatt example, not a guaranteed value for every AI site.
- Separate electricity delivered to a facility from electricity that reaches productive AI processing.
- Account for resilience: some idle capacity is intentional protection against peak demand and outages.
- Require facility-specific metering before translating percentages into money, carbon emissions or promised performance.
Nvidia’s figures make the central point clear: expanding the grid connection or GPU count does not automatically produce the same increase in AI work. Cooling, electrical conversion, configuration errors, failures, restarts and reserved peak capacity determine how much of each incoming watt becomes useful computation.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
