AI and Energy Storage Competing for PCB Capacity: What It Means

On August 18, 2026, Energy Talk reported that continued expansion of AI computing infrastructure is affecting upstream electronic component supply for the energy storage industry. Lead times for some energy storage chips have extended from approximately four weeks to three months, with memory and power management devices under simultaneous price pressure. On August 19, further industry commentary noted that power conversion systems, battery management systems and energy management systems in storage installations all depend on control chips, power devices and PCBs. When chip and PCB manufacturing resources tighten at the same time, supply chain pressure in the storage sector moves from a single component shortage into complete system delivery.

Two Industries, One Set of Manufacturing Resources

AI and energy storage appear to belong to separate industries, but at the hardware level the overlap is substantial and growing.

An AI server requires GPUs, memory, power management chips, high speed PCBs and large quantities of passive components. Storage power conversion systems, battery management systems and energy management systems require control chips, power semiconductors, power devices, multilayer PCBs and high reliability assembled boards. When AI infrastructure enters a large capital expenditure cycle, upstream wafer fabrication, packaging, electronic materials and high end PCB capacity are all pulled along.

The effect does not stop at the chip level. AI servers are using 16 to 78 layer boards, high speed low loss materials and HDI structures in volume, and producing them consumes high precision drilling, plating, lamination, exposure and inspection capacity. Energy storage does not need to replicate ultra-high layer count structures across the board, but it does have sustained demand for heavy copper, high reliability multilayer boards and power assembly. The two industries ultimately intersect over certain materials, equipment and manufacturing resources.

The consequence for storage supply chains is that the problem is no longer simply longer chip lead times. Chips, PCBs, power devices and assembly capacity are all affected by simultaneous demand expansion. Supply chain management capability is shifting from a cost variable into a delivery capability, and companies that treated procurement as a purely commercial function are discovering that it now determines whether projects ship on schedule.

Power Conversion Pushes Boards Into the Power Path

Rising storage system power ratings are changing what the PCB inside a power conversion system actually does.

Conventionally, the board handled control and signal connection. In a high power density conversion system, high current transmission, power device connection and thermal management become far more significant. Heavy copper boards of 6 ounces and above, high glass transition temperature materials and high voltage withstand structures are entering more power applications.

Heavy copper is not a matter of adding thickness. As copper weight increases, trace etching, resin filling, lamination uniformity and the process window relating copper thickness to achievable line width all change. Local temperature rise under high current and long term thermal cycling reliability must be addressed simultaneously. The upgrade path for storage power conversion is therefore fundamentally about the board moving from signal connection into power transmission, which places correspondingly higher demands on the manufacturing system.

Etching behavior is the clearest example of why heavy copper requires process-specific design. Thick copper etches laterally as well as vertically, so a trace loses more absolute width during etching than a thin copper trace of the same nominal dimension. Artwork must be compensated for the specific copper weight and chemistry, and the compensation changes when either changes. A design produced for a 1 ounce process and transferred to 4 ounce copper without recalculation will consistently miss its target geometry.

Control Boards Move in the Opposite Direction

Battery management and energy management systems are upgrading along a different path from the power stage.

As battery clusters grow in size, sampling points multiply and system control complexity increases. That drives sustained demand for multilayer control boards, higher density interconnect and high speed communication between modules. Some highly integrated designs are beginning to introduce HDI and any-layer structures, and the higher control density is pulling mSAP and sub-0.075 mm line capability further into industrial and energy electronics.

This matters because it means a single storage system may contain both heavy copper power boards and dense multilayer control boards, with different manufacturing requirements. Producing both under one quality system, or sourcing them from suppliers whose processes are comparable, simplifies qualification and reduces the risk that a failure in one board category is misattributed to the other.

Data transmission between modules in large storage installations is also increasing, which raises the importance of high speed differential signal integrity. Material dielectric properties, via structure and differential impedance control in the plus or minus five percent range are gradually becoming relevant manufacturing metrics for some high speed communication boards in storage systems, requirements that would have seemed excessive for the sector a few years ago.

Storage Is Following the Automotive and AI Playbook

The PCB upgrade path in storage is not an isolated phenomenon.

From AI servers to intelligent vehicles to robots and storage equipment, new generation electronic systems are developing along a similar trajectory: compute density rises, power density rises, data transmission speed rises, and the available volume is constrained at the same time.

Intelligent vehicle battery management, electric drive and domain controllers have already converged on a combination of heavy copper power boards, high layer count control boards and high speed communication boards. Storage power conversion, battery management and energy management are showing the same pattern, and the parallel is not coincidental. Both industries must convert and manage large amounts of electrical energy while processing increasing quantities of control data in a confined enclosure.

Robots and low altitude aircraft add a further constraint through weight reduction, which pushes flexible and rigid-flex circuits into interconnections between sensors, joints and control modules. Different end products take different physical forms, but all of them are moving the PCB toward higher power, higher density, higher speed and higher reliability simultaneously.

For programs in this space, capability spanning heavy copper, high layer count multilayer, HDI and rigid-flex construction, combined with energy PCBA assembly and testing, allows the power, control and communication boards in one system to be developed against a consistent process baseline rather than three separate ones.

Component Shortages Push Competition Toward Integrated Delivery

When chip lead times extend from four weeks to three months, the central problem for a storage manufacturer shifts from procuring a single board to coordinating an entire bill of materials.

If the PCB, power devices, control chips, assembly and testing are each handled by a different supplier, delay in any one link multiplies into the overall system delivery schedule. A three month chip lead time is manageable if the board and assembly are ready when the chip arrives. It becomes unmanageable if the board is also delayed, because the two timelines no longer overlap.

This raises the value of an integrated model covering PCB fabrication, component procurement and BOM coordination, assembly and testing. Instead of a manufacturer offering only board fabrication, the supplier consolidates dispersed supply chain nodes into a relatively complete delivery loop. For storage customers, that reduces the coordination cost of managing multiple vendors and shortens the critical path when a single component is late.

For PCB companies, the competitive dimension is changing accordingly. In the past, comparison focused on the fabricator’s process capability and price. Going forward, it will also include material integration, engineering coordination, alternate part validation and response efficiency from prototype to volume production. A supplier that can propose an alternative device and verify it without restarting the qualification cycle saves more schedule than one that simply quotes a lower board price. Where a storage program requires turnkey PCB assembly including procurement and PCBA testing, that integration is what determines whether the delivery schedule holds.

Material and Capacity Conflict Is the Real Constraint

There is a more fundamental issue underneath the lead time numbers. Both AI servers and storage systems are competing for the same constrained inputs: high precision drilling capacity, plating capacity, low loss laminate supply and inspection equipment time.

Low loss laminate is a clear example. Material suppliers allocate capacity across customers, and when a single industry sector increases its consumption sharply, allocation for other sectors tightens regardless of the price those sectors are willing to pay. A storage manufacturer that has qualified a specific laminate may find availability constrained by demand originating in an unrelated industry, and requalifying an alternative material takes time that the project schedule may not allow.

Managing that risk requires visibility earlier in the design cycle than most programs currently allow. Selecting materials with multiple qualified sources, confirming that a chosen stackup can be produced with more than one laminate, and validating alternate constructions during prototype rather than during production are all measures that reduce exposure. They cost engineering time early and save schedule risk later.

Where This Leads

The influence of AI computing expansion is extending well beyond the server industry. When chips, PCBs, materials and manufacturing equipment become resources that several high growth industries compete for simultaneously, the competitive position of an electronics manufacturer depends less on owning capacity and more on the ability to organize materials, boards, components and assembly into a stable delivery system.

For storage companies, the practical response is to treat the board and assembly supply chain as part of system design rather than as a procurement afterthought. For manufacturers, the response is to build quality management and integration capability that can absorb variability in adjacent links. As AI and storage expand together, the redistribution of supply chain value is likely to favor organizations that can hold a schedule stable while the components around them move.