PCB Equipment Demand: Why Drilling Revenue Doubled First

In its 2026 interim results, Chinese PCB equipment maker Han’s CNC reported net profit attributable to shareholders of roughly 960 million yuan, an increase of about 263 percent year on year. Within that figure, drilling equipment revenue reached 3.62 billion yuan, up approximately 114 percent and accounting for more than seventy percent of company revenue. Inspection equipment contributed around 420 million yuan, forming equipment around 400 million, and exposure equipment around 300 million. The company also stated that its newer laser drilling solutions can now process dense microvias at 50 micrometres in diameter with stable output.

Equipment revenue is a leading indicator for board manufacturing, so the composition of that growth is worth examining. The company attributes the demand to AI server high-layer-count HDI, N+N structure high-layer boards, and mSAP-class carrier boards for 800G and 1.6T optical modules. In other words, the spending is not a general expansion of PCB capacity. It is a specific upgrade of the most difficult process steps.

Why Equipment Orders Move Before Capacity

PCB manufacturers buy equipment when they can see the products they will need to build in the coming two to three years. A new line has to be ordered, installed, qualified, and brought up to yield long before it contributes revenue, so capital expenditure decisions reveal the product mix a factory expects rather than the mix it currently ships.

That is why equipment order patterns are more informative than capacity figures. Adding laminators and etchers to a factory that already builds conventional multilayer boards expands the same product. Adding laser drills, back drilling capability, and high-precision exposure equipment changes what the factory is able to build. The composition of the spending, not the total, tells you what is coming.

The pattern in this cycle is unusually concentrated. Drilling revenue more than doubled, and it dominates the total. That single fact says that the binding constraint in advanced board manufacturing has moved to the hole.

Drilling First: Aspect Ratio, Back Drilling, Microvias

Three separate trends converge on the drilling step, which is why it responded first.

Higher aspect ratio through holes. AI server mainboards and switch boards have both more layers and greater finished thickness. A through hole that must penetrate a thicker stack has a higher ratio of depth to diameter, which makes uniform copper plating progressively harder. Plating solution must reach the centre of a deep, narrow hole and deposit evenly along its length; if it does not, the barrel thins in the middle, and that thin section is where a failure will eventually occur, often only after thermal cycling.

Back drilling. High-speed channels are degraded by via stubs, the unused portion of a plated through hole that extends past the layer where the signal connects. At 112G and 224G channel rates, the stub behaves as a resonant element that distorts the signal. Back drilling removes the stub by re-drilling the hole to a controlled depth after plating. It adds a process step with its own accuracy requirement, because the remaining stub length must be controlled, and drilling too deep damages the layer being connected.

Laser microvias. As high-order HDI and optical module carrier boards enter production, the requirement shifts from mechanical holes to laser-formed microvias. At 50 micrometres, the via is smaller than the diameter of a human hair, and the demands on laser energy control, hole position accuracy, and via wall quality all tighten at once.

What 50 Micrometre Microvias Actually Require

Shrinking a via is not an isolated change. It tightens tolerances across several processes simultaneously.

Laser parameters. The energy delivered per pulse must be controlled tightly enough to ablate the dielectric without damaging the copper pad beneath it. Too much energy damages the target; too little leaves residue. Polyimide and other high-performance dielectrics also absorb and ablate differently, so parameters are material-specific.

Hole position accuracy. A small via is unforgiving of positional error. Drift between the drill and the pad means the via lands partly off-target, reducing the contact area and creating a connection with higher resistance and lower reliability.

Via wall quality and desmear. Laser ablation leaves debris and a heat-affected zone. If it is not removed completely, the subsequent plating adhesion suffers and a void can form at the interface, which is a latent defect rather than an immediate one.

Plating and filling. Increasing microvia density means more holes to plate in the same area, and higher-order HDI frequently requires filled vias to maintain a flat surface for the next lamination. Copper filling must be void-free and uniform, or the layer above will show a dimple that disrupts the following lamination and the traces printed on it.

Registration and dimensional stability. With more layers and denser vias, layer-to-layer alignment has to hold through repeated lamination cycles. Every pressing operation offers another opportunity for material movement, and the cumulative error determines whether the outer layers line up with the inner ones.

Exposure, etching and inspection. If routing moves toward mSAP-class fine lines at the same time, imaging accuracy, etch control, and inspection resolution all have to improve together. This is why the equipment growth is not confined to drilling; the accompanying increases in inspection, forming, and exposure revenue reflect the same underlying requirement. Forming equipment at plus or minus 25 micrometre accuracy is a useful illustration of the tolerance level the whole process is converging on.

Pressure Travels to Copper, Registration and Yield

Once a board carries thick panels, high-aspect-ratio holes, back drilling, blind and buried vias, and high-speed differential pairs, manufacturing difficulty propagates along the process chain rather than staying in one step. Drilling is the first operation, not the last. What follows is plating uniformity, layer registration, and inspection coverage.

The economics of high-value boards reinforce this. When the finished board is expensive, the cost of a mid-process excursion is large, so process stability rather than peak capability determines profitability. A line that can achieve a difficult feature occasionally is worth less than a line that achieves a slightly less difficult feature consistently. That is why plating capability and registration control, expressed in measurable terms such as achievable aspect ratio for vertical continuous plating and demonstrated layer-to-layer tolerance, matter as much as the drilling specification itself. Those parameters belong in a PCB capability statement that can be verified rather than assumed.

Where boards serve AI infrastructure and high-speed optical modules, the finished cost of a defect also extends beyond the board. A high-layer-count panel that fails late in the process consumes capacity that was already scarce, and the delay propagates to the customer’s build schedule. Under those conditions, the value of quality control is measured in recovered capacity as well as in yield.

Equipment Is Not Capacity

The most important caveat in this cycle is that buying machines does not create advanced PCB supply.

A 50 micrometre microvia, a high-aspect-ratio through hole, an accurately back-drilled channel, and a very fine line each require a matched combination of process parameters, material system, engineering experience, and quality control. A new machine arrives with a specification, not with a process window. It must be installed, tuned, correlated against measurement, ramped to yield, and then validated by the customer before the output can be shipped.

That sequence takes time, and it explains why nominal capacity and deliverable capacity diverge. It also explains why the shortage of advanced board capacity persists even as capital spending rises sharply.

The strategic implication is straightforward. As AI computing continues to raise layer counts, density, and channel rates, the industry may not be short of general PCB capacity at all. What will remain scarce is the ability to combine precision equipment, complex processes, and stable yield into repeatable volume delivery. The capital expenditure figures for PCB manufacturing equipment describe intent; the competitive outcome depends on who can convert that intent into shipped panels, then verify them with a qualified test process after assembly.

For buyers, the practical lesson is to look past the equipment list. A supplier that can describe how a 50 micrometre microvia recipe was developed, how plating uniformity is measured, and how back-drill depth is controlled has a process. A supplier that can only quote a machine model has a purchase order. The difference will determine who can deliver the next generation of boards, and who is left with capacity that no longer matches what the market needs.

Frequently Asked Questions

Why did drilling equipment revenue grow faster than other PCB equipment? Because AI servers and high-speed networking raised layer counts and board thickness while simultaneously requiring smaller laser microvias and more back drilling, concentrating new demand on the hole-forming step.

What is back drilling and why is it needed? Back drilling removes the unused portion of a plated through hole that extends past the signal layer. That stub degrades high-speed channels, so removing it improves signal integrity at 112G and 224G rates.

What makes 50 micrometre microvias difficult? Laser energy control, hole position accuracy, via wall cleanliness, void-free copper filling, and layer registration must all hold simultaneously, and the requirements are material-specific.

Does more equipment mean more advanced PCB capacity? Not by itself. New machines require installation, process development, yield ramp, and customer qualification before they produce deliverable output, so effective capacity lags the capital spending.

What should a customer ask about drilling and plating capability? The achievable aspect ratio, how plating uniformity is verified, how back-drill depth is controlled, and what demonstrated registration tolerance the process holds in production rather than in a capability table.