Fab Equipment Engineering Directors in China: Who Can Own Uptime

China's factories made 54.3 billion integrated circuits in September, 18.4% more than a year earlier. For the first nine months of 2026, output reached 442.8 billion units, up 23.3% year on year, according to the National Bureau of Statistics release of 19 October.

Now put the profit line beside it. In the same nine months, profits in the computer, communications and other electronic equipment manufacturing sector rose 103.8%. Manufacturing overall was up 17.0%.

Output up a quarter. Profit up double. Any operations leader reading those two numbers together will arrive at the same conclusion: this was not a volume story, it was a utilisation and mix story. And utilisation is decided by machines that keep running.

The margin sits in the tool, not the design

Most of China's new capacity is mature node — 28nm and above, power devices, MCUs, driver chips, image sensors, analog. That is where the volume in those 442.8 billion units comes from. It is also where the competition is purely on cost per wafer.

At mature node, the difference between a profitable fab and a marginal one is not architecture. It is uptime. A 300-tool fab running at 88% availability instead of 92% loses roughly four percentage points of output before anyone touches yield. On a line producing tens of thousands of wafers a month, that is the whole gross margin argument.

Here is the part that has changed in the past two years. Chinese fabs are running imported tools longer than their original maintenance schedules assumed, because replacing them is slow and expensive. Spare parts that used to arrive in weeks now arrive in months, or arrive through secondary channels with uncertain provenance. The engineering response — parts cannibalisation, in-house refurbishment, second-source qualification, extended preventive maintenance intervals — is a discipline in itself, and very few people have practised it at scale.

That is why the search in Wuxi, Shanghai, Hefei and Shenzhen has shifted. Companies are not asking for another process integration engineer. They are asking for someone who has kept a mature-node fab at high availability while the supply chain underneath it was reconfigured.

What the job now contains

An equipment engineering director in a Chinese fab in 2026 owns four things that used to sit in four different departments.

First, availability. Not the reported number — the real one, measured per tool family, with the chronic offenders named. The candidates worth interviewing can recite their worst three tools and what they did about them.

Second, the spare parts strategy. Which parts to stock, which to refurbish, which to redesign for local supply. This is a supply chain problem wearing an engineering hat, and it is where most of the cost savings in the next two years will come from.

Third, vendor management. Equipment suppliers' field service teams are stretched across a growing install base. A fab that cannot get priority service pays for it in downtime. Directors who have negotiated service-level agreements tool-family by tool-family are worth more than their title suggests.

Fourth, the qualification pipeline. Every second-source part or refurbished module has to be qualified without disturbing production. That is a scheduling problem of real complexity, and it is usually where a new director's first ninety days are decided.

The candidate pool is smaller than the output number suggests

China now runs fabs in Wuxi, Shanghai, Hefei, Shenzhen, Shaoxing, Xiamen, Beijing and Dalian, plus a long tail of specialty lines. Output is growing at a pace that would normally imply a deep bench of equipment engineers.

It doesn't. The pool of people who have actually owned availability across a few hundred tools is a few hundred people nationally, concentrated in maybe eight companies. Everyone in the industry knows who they are, and hiring committees in Wuxi, Hefei and Shanghai are fishing in the same pond.

That creates two problems for a hiring committee. The first is time. Director-level searches in this market run three to four months, and candidates from competing fabs often carry non-compete agreements that are enforceable in practice, especially in the same province. Budget the notice period before you budget the salary.

The second is fit. An equipment director from a 300mm advanced line does not automatically transfer to a mature-node specialty fab, where tool mixes are older, documentation is thinner and improvisation is part of the job. The reverse transfer is easier. Companies that hire purely on brand-name fabs often discover this in month four.

On pay, no official survey isolates this title, and quoted figures are usually hearsay. The one hard anchor is national: per-capita wage income reached RMB 19,653 in the first nine months of 2026, up 5.3% year on year. A director's package is negotiated as a multiple of that base — the multiple being a function of how many tools the person can credibly own. In Wuxi, where living costs sit well below Shanghai, the multiple often closes faster; the constraint is rarely money and almost always the non-compete calendar.

Where Henderson Executive gets involved

When a fab or a specialty foundry calls Henderson Executive (China Headhunting Service), the brief is usually written as a job description. The useful version is a list of failure modes: the tool family that has been down too often, the parts that keep arriving late, the qualification that has slipped twice.

Turn those into interview questions and the search gets shorter. Ask what availability they inherited, what it was when they left, and which tool family they would refuse to own again. Ask how many second-source qualifications they personally pushed through, and what broke. Ask who they would bring with them — not as a poaching exercise, but because the good ones always know the two engineers who make their schedule work.

The counter-argument deserves airtime

Output growth of 23.3% is not automatically good news for margins. Some of it is capacity that came online this year, some is inventory build ahead of demand, and mature-node pricing has been soft for several quarters. Profit growth of 103.8% is measured against a weak base, and it is an industry aggregate that includes memory and components, not just foundry work. A single quarter of tool orders can reverse the picture.

That is fair. But it does not change the hiring logic. If pricing stays soft, cost per wafer becomes the only lever, and uptime is the cheapest cost lever available. If pricing recovers, output becomes the lever, and uptime is again the constraint. Either way, the person who owns availability is the person who decides whether the next twelve months are profitable.

Three decisions for the next planning cycle

Staff the function before the expansion. Fabs that add capacity first and hire equipment leadership second spend the first year of a new line at availability levels that erase the business case. The director should be in seat before the tools are.

Write the mandate in uptime and parts, not in headcount. A brief that says "manage a team of forty engineers" will produce managers. A brief that says "lift availability on these three tool families by four points and cut parts lead time by half" will produce operators.

And price the non-compete, not the salary. The salary negotiation is the easy part and rarely the reason a hire fails. The calendar is the reason. Start twelve weeks earlier than feels necessary, and treat the legal timeline as a project deliverable with an owner.

China is making a quarter more chips than it did a year ago, and doing it with a tool base that is ageing faster than it is being replaced. The output number will keep making headlines. The number that decides whether the money is made is availability, and it belongs to one job title.

Sources: National Bureau of Statistics of China (19 October 2026, industrial output and industrial production volumes, September and January–September 2026); National Bureau of Statistics of China (27 October 2026, industrial profits January–September 2026); National Bureau of Statistics of China (19 October 2026, household income January–September 2026).