{"id":88662,"date":"2026-09-23T13:37:10","date_gmt":"2026-09-23T11:37:10","guid":{"rendered":"https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/"},"modified":"2026-09-23T13:37:10","modified_gmt":"2026-09-23T11:37:10","slug":"custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure","status":"publish","type":"news","link":"https:\/\/fhi.nl\/en\/news\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/","title":{"rendered":"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure"},"content":{"rendered":"<header id=\"header\" class=\"header header--low\">\n\n\t\n\t\t\t<div class=\"header__background header__background--graphic\"><\/div>\n\t\n\t<div class=\"container\">\n\t\t<div class=\"header__content\">\n\t\t\t<div class=\"header__first header__first--alone\">\n\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t<h1 class=\"header__title\" >\n\t\t\t\t\tCustom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure\t\t\t\t<\/h1>\n\n\t\t\t\t<div class=\"header__dots-line\">\n\t\t\t\t\t<svg width=\"431\" height=\"9\" viewbox=\"0 0 431 9\" fill=\"none\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\"><path d=\"M430.799 4.192a1.136 1.136 0 1 1-2.272-.001 1.136 1.136 0 0 1 2.272 0Zm-27.272 0a1.135 1.135 0 1 1-2.27 0 1.135 1.135 0 0 1 2.27 0Zm-27.27 0a1.136 1.136 0 1 1-2.272-.001 1.136 1.136 0 0 1 2.272 0Zm-27.272 0a1.39 1.39 0 1 1-2.78 0 1.39 1.39 0 0 1 2.78 0Zm-27.78 0a1.645 1.645 0 1 1-3.29 0 1.645 1.645 0 0 1 3.29 0Zm-28.29 0a1.9 1.9 0 1 1-3.799 0 1.9 1.9 0 0 1 3.799 0Zm-28.799 0a2.154 2.154 0 1 1-4.308 0 2.154 2.154 0 0 1 4.308 0Zm-29.308 0a2.41 2.41 0 1 1-4.819 0 2.41 2.41 0 0 1 4.819 0Zm-29.819 0a2.663 2.663 0 1 1-5.326.001 2.663 2.663 0 0 1 5.326-.001Zm-30.327 0a2.919 2.919 0 1 1-5.837 0 2.919 2.919 0 0 1 5.837 0Zm-30.837 0a3.173 3.173 0 1 1-6.345.001 3.173 3.173 0 0 1 6.345 0Zm-31.346 0a3.428 3.428 0 1 1-6.856 0 3.428 3.428 0 0 1 6.856 0Zm-31.856 0a3.683 3.683 0 1 1-7.365 0 3.683 3.683 0 0 1 7.365 0Zm-32.365 0a3.937 3.937 0 1 1-7.875 0 3.937 3.937 0 0 1 7.875 0Zm-32.874 0a4.192 4.192 0 1 1-8.384 0 4.192 4.192 0 0 1 8.384 0Z\" fill=\"#FFF960\"\/><\/svg>\t\t\t\t<\/div>\n\n\t\t\t\t\n\t\t\t\t\n\t\t\t<\/div>\n\n\t\t\t\n\t\t<\/div>\n\t<\/div>\n<\/header>\n\n\n\n<div class=\"text bg--white\">\n\t<div class=\"container\">\n\t\t\t\t\t\t\t\t\t\t\t<div class=\"text__content text__content--1-col\">\n\t\t\t\n<h2 class=\"wp-block-heading\">What is a custom PDU for AI hardware?<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">A custom PDU for AI hardware is a rack Power Distribution Unit configured specifically around the power requirements of GPU servers, AI accelerators, high-density compute, storage and networking equipment within an AI or HPC rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Instead of adapting the rack design to the limitations of a standard PDU, the PDU is configured around the infrastructure.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This can include:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>single- or three-phase power;<\/li>\n\n\n\n<li>10A, 16A, 32A or 63A input;<\/li>\n\n\n\n<li>specific input plug and cable requirements;<\/li>\n\n\n\n<li>custom cable lengths and cable entry positions;<\/li>\n\n\n\n<li>outlet type, quantity and positioning;<\/li>\n\n\n\n<li>phase-balanced outlet layouts;<\/li>\n\n\n\n<li>branch protection;<\/li>\n\n\n\n<li>input-, branch- and outlet-level energy metering;<\/li>\n\n\n\n<li>remote outlet switching;<\/li>\n\n\n\n<li>environmental monitoring;<\/li>\n\n\n\n<li>communication and network architecture;<\/li>\n\n\n\n<li>integration with DCIM, BMS and energy-management software.<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer PDU 5.0 supports single- and three-phase configurations and current ratings from 10A through 63A, with configurable metering at input, branch and outlet level and optional remote switching.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The objective is simple:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Design the rack PDU around the AI hardware, not the AI hardware around the PDU.<\/strong><\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Why does AI hardware require a different approach to rack power?<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">AI infrastructure is increasing rack power density.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Traditional enterprise racks may contain relatively diverse and independently operating equipment. Modern AI racks can instead contain highly concentrated GPU compute operating as a coordinated system.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That changes the importance of the rack power architecture.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">NVIDIA, for example, documents designed rack power of <strong>120 kW for GB200 NVL72 and 135 kW for GB300 NVL72<\/strong> in its current power management documentation.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">At the same time, NVIDIA is developing its infrastructure towards 800 VDC architectures for future AI factories, specifically because increasing compute performance and rack density are putting greater demands on traditional power distribution.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This does <strong>not<\/strong> mean every AI rack requires a 120 kW rack PDU, nor that every AI deployment has the same electrical architecture.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It means that the term <strong>\u201cAI rack\u201d no longer describes one standard power requirement<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The correct PDU depends on the actual hardware and power architecture being deployed.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">AI rack power starts with the hardware specification<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The first question when specifying a PDU for AI hardware should therefore not be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Which PDU model do we normally use?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It should be:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What does this rack actually need?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Important inputs include the maximum and expected power consumption of the IT equipment, number of power supplies, redundancy strategy, required supply voltage, number of feeds, available upstream capacity, connector types and physical rack layout.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The PDU specification can then be built around those requirements.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This becomes increasingly important as AI infrastructure evolves quickly. A fixed PDU catalog designed around yesterday&#039;s typical rack density may not provide the best fit for tomorrow&#039;s GPU infrastructure.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">32A or 63A PDU for AI hardware?<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">There is no universal answer.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A 32A PDU may be appropriate for one AI rack architecture, while another may require multiple 32A feeds, 63A three-phase PDUs or a different distribution architecture entirely.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For example, NVIDIA&#039;s earlier DGX H100 SuperPOD design guidance specified <strong>415 VAC, 32A, three-phase<\/strong> as the preferred power configuration for high-density deployment patterns, while allowing designs to be modified for other supply schemes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The correct question is therefore not whether <strong>32A or 63A is \u201cbetter\u201d for AI<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It is whether the complete combination of:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>voltage \u00d7 current \u00d7 phases \u00d7 number of feeds \u00d7 redundancy<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">provides the required usable capacity for the specific rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer can configure PDU 5.0 in 10A, 16A, 32A and 63A versions, including three-phase configurations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For higher-density racks, a <strong>three-phase 63A PDU<\/strong> can provide up to approximately <strong>43.5 kVA<\/strong> at a 400 V supply.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Multiple feeds can then be designed around the required rack capacity and redundancy architecture.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Why three-phase power is important in high-density AI racks<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">As rack power increases, three-phase distribution becomes particularly relevant.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Instead of concentrating the load on a single phase, power can be distributed across:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>L1 \u2192 L2 \u2192 L3<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The PDU outlet configuration can then be designed to distribute connected loads across the available phases.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For AI infrastructure, phase balance is not merely a design consideration at installation.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It should also be <strong>measurable during operation<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is where an intelligent PDU becomes valuable: the operator can see how the actual load develops instead of relying only on the original design calculation.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">AI workloads are dynamic, not static<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Nameplate power and average consumption do not tell the entire story.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">AI compute can produce significant changes in electrical demand as workloads move between training, inference, idle states and other operating conditions.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">NVIDIA specifically notes that GPU clusters can create sudden power-demand bursts and that infrastructure based purely on static provisioning may struggle with these runtime fluctuations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This makes <strong>headroom and real-time measurement<\/strong> increasingly important.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A rack may appear to have sufficient capacity based on average consumption while individual phases, branches or feeds approach their operational limits during workload peaks.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For data center operators, the question therefore becomes:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>How much capacity do we have \u2014 and where is that capacity actually available?<\/strong><\/p>\n<\/blockquote>\n\n\n\n<h1 class=\"wp-block-heading\">Measure AI power where it is actually consumed<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">The closer measurement gets to the IT equipment, the more useful the data becomes.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">An intelligent PDU can provide several layers of visibility:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Rack \u2192 PDU \u2192 Phase \u2192 Branch \u2192 Outlet \u2192 IT device<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Depending on the selected Schleifenbauer PDU functionality, measurements can be provided at input, branch and outlet level. The PDU 5.0 platform supports power monitoring with stated 0.5% accuracy at input and outlet level.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For an AI rack, this can help answer practical questions such as:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How much power is the rack consuming?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How is the load distributed across L1, L2 and L3?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Which branch is approaching its capacity?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How much power is an individual connected device consuming?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How much headroom remains before additional AI hardware can be deployed?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That is considerably more actionable than knowing only the total facility load.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Outlet-level monitoring for AI and GPU infrastructure<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Outlet-level monitoring adds another level of granularity.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Instead of knowing only how much energy enters the PDU, operators can see consumption closer to individual connected loads.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This becomes useful in racks containing combinations of:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">GPU compute systems, CPU servers, network switches, storage appliances, management equipment and other supporting IT hardware.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The PDU effectively becomes the <strong>measurement boundary between rack power distribution and individual IT loads<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For Managed PDU configurations, outlet-level monitoring can be combined with remote outlet switching.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Custom outlets for AI hardware<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Power capacity alone does not determine whether a PDU fits an AI rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The physical connections must fit as well.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Different IT systems can require different plug types and outlet arrangements. High-density racks can also create pressure on the amount of usable rack space available for power distribution.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer therefore allows the <strong>type, quantity, position and layout of outlets to be configured around the project<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">One available option is the <a href=\"https:\/\/www.schleifenbauer.eu\/en\/cx-lock-outlet\/\"><strong>CX Combination Outlet C13\/C15\/C19\/C21<\/strong>.<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A single CX outlet can accept C14, C16, C20 and C22 plugs, while IEC Lock helps prevent accidental disconnection. Custom outlet combinations and load-balancing layouts can also be specified.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For AI racks containing different device types, this can simplify the physical power design considerably.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Redundancy matters because an AI rack is a system<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">A high-density AI rack should not necessarily be considered as a collection of independent servers.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">In clustered AI environments, the failure of one component can affect a much larger workload.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">NVIDIA makes this particularly clear in its DGX SuperPOD design guidance: a failure of a single system within a multi-node AI workload can cause the complete job to stop. Its H100 design therefore incorporates specific power-source redundancy requirements.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The precise redundancy architecture depends on the AI platform.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That can mean:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>A\/B feeds<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">or<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>multiple independent power paths<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">or another architecture prescribed by the hardware vendor.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The rack PDU design needs to follow that redundancy strategy rather than undermine it.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Do not confuse available PDU capacity with usable AI capacity<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">This distinction becomes increasingly important at high power densities.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A rack may have several power feeds, but their combined nameplate capacity is not automatically the amount of power that can safely be allocated to IT equipment.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Redundancy, breaker ratings, phase loading, upstream infrastructure and required failover headroom all affect usable capacity.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">NVIDIA&#039;s current power-management documentation similarly distinguishes between physical power paths and the usable power envelope that remains after redundancy requirements are considered.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For AI deployments, capacity management therefore needs to answer two different questions:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What is installed?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">and<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What can be safely used?<\/strong><\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Power Quality becomes more relevant as rack density increases<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">AI power management is not only about kW and kWh.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">High-density electronic loads also make the characteristics of the electrical load increasingly interesting.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer PDU 5.0 and EnerTree can provide <strong><a href=\"https:\/\/www.schleifenbauer.eu\/en\/data-centre-power-quality-monitoring\/\">Power Quality Monitoring<\/a><\/strong>, including measurements such as Total Harmonic Distortion (THD), Crest Factor and voltage\/current peaks.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This adds context that ordinary energy-consumption monitoring cannot provide.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For an AI infrastructure operator, energy monitoring answers:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How much electricity is being consumed?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Power Quality Monitoring helps answer:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>What is happening electrically while that power is being consumed?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The two should be considered complementary.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">From custom AI PDU to Data Center Energy Management<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">A single intelligent PDU provides detailed information about one rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">An AI data center may contain hundreds or thousands of PDUs.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That creates the next challenge:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>How do you turn thousands of electrical measurements into usable operational information?<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is where <strong>EnerTree<\/strong> becomes part of the architecture.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer PDU 5.0 separates the required PDU functionality from the way the intelligent PDU environment is managed.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The communication architecture can be selected independently using a <strong>Controller Module, Gateway Module or Daisy Chain Module<\/strong>.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Controller Module + EnerTree Lite<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">The Controller Module provides a conventional intelligent PDU architecture.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>EnerTree Lite<\/strong> runs embedded on the Controller Module, providing a local web interface and direct access to PDU measurements.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It can also integrate directly with an existing DCIM environment.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">EnerTree Lite supports environments of up to <strong>100 PDUs per IP address<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For smaller AI installations or environments that already have an established management platform, this may provide the required architecture.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Gateway Module + EnerTree Platform<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">For larger AI deployments, the Gateway architecture moves intelligence away from each individual PDU and centralizes it within <strong>EnerTree Platform<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">EnerTree Platform runs as a virtual machine.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The PDU becomes a real-time measurement and control endpoint, while processing, analysis, configuration and management are centralized.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A single EnerTree Platform environment can scale to <strong>10,000 PDUs<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For large GPU clusters and AI data centers, this provides a very different operating model from independently managing thousands of intelligent rack PDUs.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">EnerTree functionality includes real-time energy monitoring, alerts, historical analysis, reporting, hierarchical infrastructure views, environmental monitoring and PUE calculation.<\/p>\n\n\n\n<h2 class=\"wp-block-heading\">Daisy Chain for scalable PDU connectivity<\/h2>\n\n\n\n<p class=\"wp-block-paragraph\">Not every PDU in an AI rack or row needs its own Ethernet connection.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A Controller or Gateway can connect downstream PDUs through a Daisy Chain architecture.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Up to <strong>100 PDUs can operate within one ring on a single IP address<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A dedicated Daisy Chain Module functions as a cost-efficient data relay without its own Ethernet port, reducing network hardware and the Ethernet attack surface inside the rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This can become particularly relevant when deploying intelligent PDUs at scale.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">From one GPU outlet to complete AI data center visibility<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">The value of this architecture becomes clearer when looking at the complete hierarchy:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>AI device<\/strong><br>\u2193<br><strong>Outlet<\/strong><br>\u2193<br><strong>Branch<\/strong><br>\u2193<br><strong>Phase<\/strong><br>\u2193<br><strong>PDU<\/strong><br>\u2193<br><strong>Rack<\/strong><br>\u2193<br><strong>Row<\/strong><br>\u2193<br><strong>Data center<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">At the bottom of that hierarchy, the PDU distributes electricity to the AI hardware.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">At the same time, measurement data travels in the opposite direction.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>Device \u2192 Outlet \u2192 Branch \u2192 Phase \u2192 PDU \u2192 Rack \u2192 EnerTree<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The physical power infrastructure therefore becomes a source of structured energy data.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That is the difference between simply <strong>delivering power to AI hardware<\/strong> and actually <strong>understanding AI rack power<\/strong>.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Integrating AI rack power with DCIM and BMS<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">AI power data should not become another isolated dataset.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">EnerTree Platform and EnerTree Lite support integration with external data center systems through interfaces and protocols including <strong>HTTP\/HTTPS, REST API, MODBUS\/TCP, SNMP v1\/v2c\/v3, IPv4\/IPv6, SMTP, Syslog and command-line interfaces<\/strong>. Data can also be exported to databases including MS SQL, MySQL and MariaDB.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This allows rack-level electrical information to become part of a broader DCIM, BMS or data analytics environment.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Custom AI PDUs should also be physically customizable<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">AI infrastructure changes quickly.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A PDU that has the correct electrical rating but does not fit the rack layout is still the wrong PDU.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer therefore builds its PDUs to order.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Configuration options include cable length, connector type, outlet layout and position, metering functionality, controller position, identification and labelling, mounting adaptations and other mechanical requirements.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">PDU 5.0 can also be configured with different protection options, residual current sensing, environmental sensors and additional features depending on the project.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">This is particularly relevant for OEMs, system integrators, data center designers and operators deploying AI hardware where the rack itself is increasingly engineered as an integrated system.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">What should you specify when ordering a custom PDU for AI hardware?<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Before selecting the PDU, define the rack.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A useful AI PDU specification should establish at minimum:<\/p>\n\n\n\n<figure class=\"wp-block-table\"><table class=\"has-fixed-layout\"><thead><tr><th>Requirement<\/th><th>What to determine<\/th><\/tr><\/thead><tbody><tr><td>AI hardware<\/td><td>GPU\/server platform and configuration<\/td><\/tr><tr><td>Maximum rack load<\/td><td>Expected and design power<\/td><\/tr><tr><td>Supply<\/td><td>Voltage and frequency<\/td><\/tr><tr><td>Phases<\/td><td>Single-phase or three-phase<\/td><\/tr><tr><td>Current<\/td><td>16A, 32A, 63A or project-specific requirement<\/td><\/tr><tr><td>Redundancy<\/td><td>A\/B, N+1, 2N or vendor-prescribed architecture<\/td><\/tr><tr><td>Number of feeds<\/td><td>Requires independent rack power paths<\/td><\/tr><tr><td>Connectors<\/td><td>Input plug and IT equipment plug types<\/td><\/tr><tr><td>Outlets<\/td><td>Type, quantity, position and grouping<\/td><\/tr><tr><td>Phase layout<\/td><td>Required load distribution across L1\/L2\/L3<\/td><\/tr><tr><td>Branch protection<\/td><td>Required breaker\/fuse architecture<\/td><\/tr><tr><td>Metering<\/td><td>Input, branch and\/or outlet<\/td><\/tr><tr><td>Switching<\/td><td>Whether remote outlet control is required<\/td><\/tr><tr><td>Monitoring<\/td><td>Energy, capacity, Power Quality, environment<\/td><\/tr><tr><td>Networking<\/td><td>Controller, Gateway or Daisy Chain<\/td><\/tr><tr><td>Integration<\/td><td>DCIM\/BMS\/API\/SNMP requirements<\/td><\/tr><tr><td>Physical design<\/td><td>0U\/19\u2033\/21\u2033, dimensions, mounting, cable entry<\/td><\/tr><\/tbody><\/table><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">The most important principle is:<\/p>\n\n\n\n<blockquote class=\"wp-block-quote is-layout-flow wp-block-quote-is-layout-flow\">\n<p class=\"wp-block-paragraph\"><strong>Do not specify the AI rack PDU independently from the AI hardware, upstream power architecture and redundancy strategy.<\/strong><\/p>\n<\/blockquote>\n\n\n\n<h1 class=\"wp-block-heading\">Can an existing data center be adapted for AI hardware?<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Often, yes, but the available electrical capacity and distribution architecture need to be assessed first.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Not every AI deployment requires a completely new data center.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Existing facilities can potentially accommodate GPU infrastructure by changing rack layouts, increasing power density in selected areas, modifying distribution or upgrading rack-level power infrastructure.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">NVIDIA itself is developing hybrid approaches intended to allow next-generation AI compute to operate with existing AC facility infrastructure during the industry&#039;s transition towards higher-voltage DC architectures.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For existing rack infrastructure, Schleifenbauer also offers <strong>Inline Meters<\/strong> that can add input-level energy measurement to existing Basic or legacy intelligent PDUs without replacing the complete installed base.<\/p>\n\n\n\n<h1 class=\"wp-block-heading\">What is the best PDU for AI hardware?<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">There is no single best PDU for every AI deployment.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The correct PDU is the one that matches the <strong>electrical, mechanical, redundancy, monitoring and management requirements of the specific AI rack<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For one project, that might be a three-phase 32A Monitored PDU.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another may require multiple 63A feeds, outlet-level monitoring, custom CX outlets and a Gateway Module connected to EnerTree Platform.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Another AI architecture may move beyond conventional AC rack distribution altogether.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">That is precisely why a configurable approach matters.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Rather than asking:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u201cWhich standard PDU model is designed for AI?\u201d<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A better question is:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>\u201cWhat rack power architecture does this AI hardware require?\u201d<\/strong><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The PDU can then be engineered around the answer.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><br>Read also case study Supermicro: <a href=\"https:\/\/www.schleifenbauer.eu\/en\/cases\/case-study-supermicro\/\">https:\/\/www.schleifenbauer.eu\/en\/cases\/case-study-supermicro\/<\/a><\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Read also Nvidia case study: <a href=\"https:\/\/www.schleifenbauer.eu\/en\/cases\/case-study-nvidia\/\">https:\/\/www.schleifenbauer.eu\/en\/cases\/case-study-nvidia\/<\/a><\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Schleifenbauer custom PDUs for AI hardware<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">Schleifenbauer designs and manufactures rack PDUs in the Netherlands and does not rely on fixed catalog configurations.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">For AI and high-density computing environments, this allows the PDU to be configured around the infrastructure: <strong>10A to 63A, single- or three-phase, custom inputs, cable lengths, outlet configurations, protection, input\/branch\/outlet metering, remote switching and different management architectures<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">PDU 5.0 then adds a modular intelligence layer.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">Choose <strong>Controller + EnerTree Lite<\/strong> for conventional local PDU intelligence, <strong>Gateway + EnerTree Platform<\/strong> for centralized Data Center Energy Management, or use <strong>Daisy Chain<\/strong> to create scalable PDU networks with fewer Ethernet connections.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">And because the communication modules are hot-swappable, the management architecture can be changed later without replacing the underlying PDU.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">The result is not simply a high-power rack PDU.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">It is an architecture designed to make AI rack power:<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>configurable \u2192 measurable \u2192 visible \u2192 manageable \u2192 scalable.<\/strong><\/p>\n\n\n\n<h1 class=\"wp-block-heading\">Building an AI rack? Start with the power requirements<\/h1>\n\n\n\n<p class=\"wp-block-paragraph\">If you are designing a new AI, GPU or HPC rack, send Schleifenbauer the <strong>hardware specification, rack design, tender specification or required electrical configuration<\/strong>.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">A custom rack PDU can then be configured around the actual infrastructure requirements rather than forcing the project into a predefined PDU model.<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><strong>From the incoming rack feed to the individual AI device \u2014 design the power distribution around the hardware.<\/strong><\/p>\n\n\n\n<figure class=\"wp-block-gallery has-nested-images columns-default is-cropped wp-block-gallery-2 is-layout-flex wp-block-gallery-is-layout-flex\">\n<figure class=\"wp-block-image size-large\"><figcaption class=\"wp-element-caption\">Supermicro PDU<\/figcaption><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><figcaption class=\"wp-element-caption\">NVIDIA PDU<\/figcaption><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><figcaption class=\"wp-element-caption\">Nvidia PDU<\/figcaption><\/figure>\n\n\n\n<figure class=\"wp-block-image size-large\"><figcaption class=\"wp-element-caption\">CX Combi outlet PDU<\/figcaption><\/figure>\n<\/figure>\n<p>The post <a href=\"https:\/\/www.schleifenbauer.eu\/en\/custom-pdus-for-ai-hardware\/\">Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure<\/a> appeared first on <a href=\"https:\/\/www.schleifenbauer.eu\/en\">Schleifenbauer \u2013 Bespoke PDUs<\/a>.<\/p>\n<p>Source: <a href=\"https:\/\/www.schleifenbauer.eu\/en\/custom-pdus-for-ai-hardware\/\">https:\/\/www.schleifenbauer.eu\/en\/custom-pdus-for-ai-hardware\/<\/a><\/p>\t\t<\/div>\n\t<\/div>\n<\/div>\r\n\t<div class=\"articles bg--offwhite automatic\">\r\n\t\t<div class=\"container\">\r\n\t\t\t<div class=\"articles__header\">\r\n\t\t\t\t\t\t\t\t\t<div class='heading-wrapper'><svg width=\"13\" height=\"13\" viewbox=\"0 0 13 13\" fill=\"none\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\"><circle cx=\"6.394\" cy=\"6.5\" r=\"6.394\" fill=\"#000\"\/><\/svg><h2>Related articles<\/h2><\/div>\t\t\t\t\t\t\t\t\t\t\t\t\t<a href=\"\/en\/profiel\/schleifenbauer-products-bv\/\" class=\"button button--outline\">view profile<\/a>\r\n\t\t\t\t\t\t\t<\/div>\r\n\t\t\t\t\t\t<div class=\"post-grid post-grid--no-padding\">\r\n\t\t\t\t\n<a class=\"single-item single-item__articles\" href=\"https:\/\/fhi.nl\/en\/news\/keysight-ap4021a-fs1-frequency-synthesizer-delivers-low-phase-noise-and-ultra-fast-frequency-switching-up-to-40-ghz\/\" data-id=\"84563\">\n\t<div class=\"single-item__articles-icon\">\n\t\t<svg width=\"103\" height=\"103\" viewbox=\"0 0 103 103\" fill=\"none\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\">\n<path fill-rule=\"evenodd\" clip-rule=\"evenodd\" d=\"M49.5 98.9587V82.3076C33.3844 81.2758 20.6328 67.878 20.6328 51.5019C20.6328 35.1258 33.3844 21.728 49.5 20.6962V4.04134C24.1941 5.08952 4 25.9365 4 51.5C4 77.0635 24.1941 97.9105 49.5 98.9587ZM53.5 78.2983V61.8096H76.3238C72.5238 70.9498 63.8222 77.5397 53.5 78.2983ZM49.5 78.298V61.8096H26.681C30.4804 70.9484 39.1799 77.5377 49.5 78.298ZM47.0728 57.8096H25.3773C24.8906 55.7865 24.6328 53.6743 24.6328 51.5019C24.6328 37.3358 35.5955 25.7301 49.5 24.7058V37.013H34.0466C33.2709 37.013 32.5653 37.4615 32.2359 38.1637C31.9065 38.8659 32.0128 39.6952 32.5086 40.2916L47.0728 57.8096ZM55.9249 57.8096H77.6274C78.1141 55.7865 78.3719 53.6743 78.3719 51.5019C78.3719 37.3342 67.4068 25.7275 53.5 24.7055V37.013H68.9511C69.7268 37.013 70.4324 37.4615 70.7618 38.1637C71.0912 38.8659 70.9849 39.6952 70.4891 40.2916L55.9249 57.8096ZM53.5 82.3079V98.9587C78.8059 97.9105 99 77.0635 99 51.5C99 25.9365 78.8059 5.08952 53.5 4.04134V20.6959C69.6178 21.7254 82.3719 35.1242 82.3719 51.5019C82.3719 67.8796 69.6178 81.2784 53.5 82.3079ZM0 51.5C0 23.7271 21.9843 1.08883 49.5 0.0381213V0H51.5H53.5V0.0381213C81.0157 1.08883 103 23.7271 103 51.5C103 79.9427 79.9427 103 51.5 103C23.0573 103 0 79.9427 0 51.5ZM38.3102 41.013H64.6875L51.4989 56.8765L38.3102 41.013Z\" fill=\"#2DD881\"\/>\n<\/svg>\n\t<\/div>\n\t<div class=\"single-item__articles-title\"><div class='heading-wrapper'><h3>Keysight AP4021A FS1 Frequency Synthesizer Delivers Low Phase Noise and Ultra-Fast Frequency Switching up to 40 GHz<\/h3><\/div><\/div>\n\t<div class=\"single-item__articles-terms\">\n\t\t\n\t\t\n\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Industrial Electronics<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Design Automation &amp; Embedded Systems Event<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Test and measuring equipment<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Digitization<\/span>\n\t\t\t\t\t<\/div>\n\t<div class=\"single-item__articles-author-date-wrapper\">\n\t\t\t\t\t<div class=\"single-item__articles-author\">\n\t\t\t\tCN Rood\t\t\t<\/div>\n\t\t\t\t\t\t\t<div class=\"single-item__articles-date\">\n\t\t\t\t23 June 2026\t\t\t<\/div>\n\t\t\t<\/div>\n<\/a>\n\n<a class=\"single-item single-item__articles\" href=\"https:\/\/fhi.nl\/en\/news\/wurth-elektronik-offers-m12-a-circular-connectors\/\" data-id=\"33426\">\n\t<div class=\"single-item__articles-icon\">\n\t\t<svg width=\"103\" height=\"103\" viewbox=\"0 0 103 103\" fill=\"none\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\">\n<path fill-rule=\"evenodd\" clip-rule=\"evenodd\" d=\"M49.5 98.9587V82.3076C33.3844 81.2758 20.6328 67.878 20.6328 51.5019C20.6328 35.1258 33.3844 21.728 49.5 20.6962V4.04134C24.1941 5.08952 4 25.9365 4 51.5C4 77.0635 24.1941 97.9105 49.5 98.9587ZM53.5 78.2983V61.8096H76.3238C72.5238 70.9498 63.8222 77.5397 53.5 78.2983ZM49.5 78.298V61.8096H26.681C30.4804 70.9484 39.1799 77.5377 49.5 78.298ZM47.0728 57.8096H25.3773C24.8906 55.7865 24.6328 53.6743 24.6328 51.5019C24.6328 37.3358 35.5955 25.7301 49.5 24.7058V37.013H34.0466C33.2709 37.013 32.5653 37.4615 32.2359 38.1637C31.9065 38.8659 32.0128 39.6952 32.5086 40.2916L47.0728 57.8096ZM55.9249 57.8096H77.6274C78.1141 55.7865 78.3719 53.6743 78.3719 51.5019C78.3719 37.3342 67.4068 25.7275 53.5 24.7055V37.013H68.9511C69.7268 37.013 70.4324 37.4615 70.7618 38.1637C71.0912 38.8659 70.9849 39.6952 70.4891 40.2916L55.9249 57.8096ZM53.5 82.3079V98.9587C78.8059 97.9105 99 77.0635 99 51.5C99 25.9365 78.8059 5.08952 53.5 4.04134V20.6959C69.6178 21.7254 82.3719 35.1242 82.3719 51.5019C82.3719 67.8796 69.6178 81.2784 53.5 82.3079ZM0 51.5C0 23.7271 21.9843 1.08883 49.5 0.0381213V0H51.5H53.5V0.0381213C81.0157 1.08883 103 23.7271 103 51.5C103 79.9427 79.9427 103 51.5 103C23.0573 103 0 79.9427 0 51.5ZM38.3102 41.013H64.6875L51.4989 56.8765L38.3102 41.013Z\" fill=\"#2DD881\"\/>\n<\/svg>\n\t<\/div>\n\t<div class=\"single-item__articles-title\"><div class='heading-wrapper'><h3>W\u00fcrth Elektronik offers M12-A circular connectors<\/h3><\/div><\/div>\n\t<div class=\"single-item__articles-terms\">\n\t\t\n\t\t\n\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Industrial Electronics<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Design Automation &amp; Embedded Systems Event<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">19 inch racks<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Active components<\/span>\n\t\t\t\t\t<\/div>\n\t<div class=\"single-item__articles-author-date-wrapper\">\n\t\t\t\t\t<div class=\"single-item__articles-author\">\n\t\t\t\tW\u00fcrth Elektronik NL BV\t\t\t<\/div>\n\t\t\t\t\t\t\t<div class=\"single-item__articles-date\">\n\t\t\t\tApril 21, 2023\t\t\t<\/div>\n\t\t\t<\/div>\n<\/a>\n\n<a class=\"single-item single-item__articles\" href=\"https:\/\/fhi.nl\/en\/news\/workshop-industrial-network-communication-engelstalig\/\" data-id=\"34663\">\n\t<div class=\"single-item__articles-icon\">\n\t\t<svg width=\"35\" height=\"35\" viewbox=\"0 0 35 35\" fill=\"none\" xmlns=\"http:\/\/www.w3.org\/2000\/svg\">\n<mask id=\"mask0_1182_5532\" style=\"mask-type:luminance\" maskunits=\"userSpaceOnUse\" x=\"0\" y=\"0\" width=\"35\" height=\"35\">\n<path d=\"M0 0H35V35H0V0Z\" fill=\"white\"\/>\n<\/mask>\n<g mask=\"url(#mask0_1182_5532)\">\n<path d=\"M5.12695 9.22852H1.02539V31.9238H10.4299C12.2868 31.9238 14.0678 32.6615 15.3809 33.9746H19.6191C20.9322 32.6615 22.7132 31.9238 24.5701 31.9238H33.9746V11.2793H29.873\" stroke=\"#2A5CEE\" stroke-width=\"2\" stroke-miterlimit=\"10\"\/>\n<path d=\"M17.5 9.22852H18.1426C20.0088 7.89544 22.2237 7.17773 24.5615 7.17773H29.873V27.8223H24.5615C22.2237 27.8223 20.0088 28.54 18.1426 29.873H16.8574C14.9912 28.54 12.7763 27.8223 10.4385 27.8223H5.12695V5.12695H9.22852\" stroke=\"#2A5CEE\" stroke-width=\"2\" stroke-miterlimit=\"10\"\/>\n<path d=\"M9.22852 1.02539V21.6699C13.759 21.6699 17.5 25.3425 17.5 29.873V9.22852C17.5 4.698 13.759 1.02539 9.22852 1.02539Z\" stroke=\"#2A5CEE\" stroke-width=\"2\" stroke-miterlimit=\"10\"\/>\n<\/g>\n<\/svg>\n\t<\/div>\n\t<div class=\"single-item__articles-title\"><div class='heading-wrapper'><h3>Workshop Industrial Network Communication (English)<\/h3><\/div><\/div>\n\t<div class=\"single-item__articles-terms\">\n\t\t\n\t\t\n\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Industrial automation<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Industrial Ethernet<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Software for automation<\/span>\n\t\t\t\t\t\t\t\t<span class=\"button button--outline single-item__articles-term--branche single-item__articles-term\">Cybersecurity<\/span>\n\t\t\t\t\t<\/div>\n\t<div class=\"single-item__articles-author-date-wrapper\">\n\t\t\t\t\t<div class=\"single-item__articles-author\">\n\t\t\t\tPhoenix Contact BV\t\t\t<\/div>\n\t\t\t\t\t\t\t<div class=\"single-item__articles-date\">\n\t\t\t\tSeptember 20, 2022\t\t\t<\/div>\n\t\t\t<\/div>\n<\/a>\n\t\t\t<\/div>\r\n\t\t<\/div>\r\n\t<\/div>","protected":false},"excerpt":{"rendered":"","protected":false},"featured_media":0,"template":"","branches":[],"events":[],"secretariat":[],"categories":[],"themes_tax":[],"content_types":[],"class_list":["post-88662","news","type-news","status-publish","hentry"],"acf":[],"yoast_head":"<!-- This site is optimized with the Yoast SEO plugin v28.5 - https:\/\/yoast.com\/product\/yoast-seo-wordpress\/ -->\n<title>Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, federatie van technologiebranches<\/title>\n<meta name=\"robots\" content=\"index, follow, max-snippet:-1, max-image-preview:large, max-video-preview:-1\" \/>\n<link rel=\"canonical\" href=\"https:\/\/fhi.nl\/en\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/\" \/>\n<meta property=\"og:locale\" content=\"en_GB\" \/>\n<meta property=\"og:type\" content=\"article\" \/>\n<meta property=\"og:title\" content=\"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, federatie van technologiebranches\" \/>\n<meta property=\"og:url\" content=\"https:\/\/fhi.nl\/en\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/\" \/>\n<meta property=\"og:site_name\" content=\"FHI, federatie van technologiebranches\" \/>\n<meta name=\"twitter:card\" content=\"summary_large_image\" \/>\n<script type=\"application\/ld+json\" class=\"yoast-schema-graph\">{\"@context\":\"https:\\\/\\\/schema.org\",\"@graph\":[{\"@type\":\"WebPage\",\"@id\":\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\\\/\",\"url\":\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\\\/\",\"name\":\"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, federatie van technologiebranches\",\"isPartOf\":{\"@id\":\"https:\\\/\\\/fhi.nl\\\/#website\"},\"datePublished\":\"2026-09-23T11:37:10+00:00\",\"breadcrumb\":{\"@id\":\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\\\/#breadcrumb\"},\"inLanguage\":\"en-GB\",\"potentialAction\":[{\"@type\":\"ReadAction\",\"target\":[\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\\\/\"]}]},{\"@type\":\"BreadcrumbList\",\"@id\":\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\\\/#breadcrumb\",\"itemListElement\":[{\"@type\":\"ListItem\",\"position\":1,\"name\":\"Home\",\"item\":\"https:\\\/\\\/fhi.nl\\\/\"},{\"@type\":\"ListItem\",\"position\":2,\"name\":\"Nieuws\",\"item\":\"https:\\\/\\\/fhi.nl\\\/nieuws\\\/\"},{\"@type\":\"ListItem\",\"position\":3,\"name\":\"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure\"}]},{\"@type\":\"WebSite\",\"@id\":\"https:\\\/\\\/fhi.nl\\\/#website\",\"url\":\"https:\\\/\\\/fhi.nl\\\/\",\"name\":\"FHI, federatie van technologiebranches\",\"description\":\"Nederlandse branchevereniging voor technologiebranches\",\"publisher\":{\"@id\":\"https:\\\/\\\/fhi.nl\\\/#organization\"},\"potentialAction\":[{\"@type\":\"SearchAction\",\"target\":{\"@type\":\"EntryPoint\",\"urlTemplate\":\"https:\\\/\\\/fhi.nl\\\/?s={search_term_string}\"},\"query-input\":{\"@type\":\"PropertyValueSpecification\",\"valueRequired\":true,\"valueName\":\"search_term_string\"}}],\"inLanguage\":\"en-GB\"},{\"@type\":\"Organization\",\"@id\":\"https:\\\/\\\/fhi.nl\\\/#organization\",\"name\":\"FHI, federatie van technologiebranches\",\"url\":\"https:\\\/\\\/fhi.nl\\\/\",\"logo\":{\"@type\":\"ImageObject\",\"inLanguage\":\"en-GB\",\"@id\":\"https:\\\/\\\/fhi.nl\\\/#\\\/schema\\\/logo\\\/image\\\/\",\"url\":\"https:\\\/\\\/fhi.nl\\\/app\\\/uploads\\\/2024\\\/06\\\/3-e1722349014385.png\",\"contentUrl\":\"https:\\\/\\\/fhi.nl\\\/app\\\/uploads\\\/2024\\\/06\\\/3-e1722349014385.png\",\"width\":732,\"height\":136,\"caption\":\"FHI, federatie van technologiebranches\"},\"image\":{\"@id\":\"https:\\\/\\\/fhi.nl\\\/#\\\/schema\\\/logo\\\/image\\\/\"},\"sameAs\":[\"https:\\\/\\\/www.linkedin.com\\\/company\\\/fhi-federation-of-technology-branches\",\"https:\\\/\\\/www.instagram.com\\\/fhi_nl\\\/\"]}]}<\/script>\n<!-- \/ Yoast SEO plugin. -->","yoast_head_json":{"title":"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, Federation of Technology Industries","robots":{"index":"index","follow":"follow","max-snippet":"max-snippet:-1","max-image-preview":"max-image-preview:large","max-video-preview":"max-video-preview:-1"},"canonical":"https:\/\/fhi.nl\/en\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/","og_locale":"en_GB","og_type":"article","og_title":"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, federatie van technologiebranches","og_url":"https:\/\/fhi.nl\/en\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/","og_site_name":"FHI, federatie van technologiebranches","twitter_card":"summary_large_image","schema":{"@context":"https:\/\/schema.org","@graph":[{"@type":"WebPage","@id":"https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/","url":"https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/","name":"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure - FHI, Federation of Technology Industries","isPartOf":{"@id":"https:\/\/fhi.nl\/#website"},"datePublished":"2026-09-23T11:37:10+00:00","breadcrumb":{"@id":"https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/#breadcrumb"},"inLanguage":"en-GB","potentialAction":[{"@type":"ReadAction","target":["https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/"]}]},{"@type":"BreadcrumbList","@id":"https:\/\/fhi.nl\/nieuws\/custom-pdus-for-ai-hardware-designing-rack-power-for-high-density-ai-infrastructure\/#breadcrumb","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https:\/\/fhi.nl\/"},{"@type":"ListItem","position":2,"name":"Nieuws","item":"https:\/\/fhi.nl\/nieuws\/"},{"@type":"ListItem","position":3,"name":"Custom PDUs for AI Hardware: Designing Rack Power for High-Density AI Infrastructure"}]},{"@type":"WebSite","@id":"https:\/\/fhi.nl\/#website","url":"https:\/\/fhi.nl\/","name":"FHI, federation of technology industries","description":"Dutch trade association for technology industries","publisher":{"@id":"https:\/\/fhi.nl\/#organization"},"potentialAction":[{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https:\/\/fhi.nl\/?s={search_term_string}"},"query-input":{"@type":"PropertyValueSpecification","valueRequired":true,"valueName":"search_term_string"}}],"inLanguage":"en-GB"},{"@type":"Organization","@id":"https:\/\/fhi.nl\/#organization","name":"FHI, federation of technology industries","url":"https:\/\/fhi.nl\/","logo":{"@type":"ImageObject","inLanguage":"en-GB","@id":"https:\/\/fhi.nl\/#\/schema\/logo\/image\/","url":"https:\/\/fhi.nl\/app\/uploads\/2024\/06\/3-e1722349014385.png","contentUrl":"https:\/\/fhi.nl\/app\/uploads\/2024\/06\/3-e1722349014385.png","width":732,"height":136,"caption":"FHI, federatie van technologiebranches"},"image":{"@id":"https:\/\/fhi.nl\/#\/schema\/logo\/image\/"},"sameAs":["https:\/\/www.linkedin.com\/company\/fhi-federation-of-technology-branches","https:\/\/www.instagram.com\/fhi_nl\/"]}]}},"_links":{"self":[{"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/news\/88662","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/news"}],"about":[{"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/types\/news"}],"version-history":[{"count":0,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/news\/88662\/revisions"}],"wp:attachment":[{"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/media?parent=88662"}],"wp:term":[{"taxonomy":"branches","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/branches?post=88662"},{"taxonomy":"events","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/events?post=88662"},{"taxonomy":"secretariat","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/secretariat?post=88662"},{"taxonomy":"categories","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/categories?post=88662"},{"taxonomy":"themes","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/themes_tax?post=88662"},{"taxonomy":"content_types","embeddable":true,"href":"https:\/\/fhi.nl\/en\/wp-json\/wp\/v2\/content_types?post=88662"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}