Herrenknecht

Nearly every big tunnel on earth — metro lines, Alpine rail links, water mains — is cut by a machine from a village on the Rhine plain. Herrenknecht builds them one at a time, for ground that is never the same twice, and calls itself the last tunnel-borer maker left in the Western world. That claim is doing more work than it looks.

Fifty metres under the Pearl River Delta, in the spring of 2015, the largest machine of its kind ever built was working its way toward Hong Kong’s airport. It was 17.63 metres across the face — a six-storey building laid on its side — 120 metres long, and it weighed 4,850 tonnes [3]. Guinness World Records certified it that March as the largest tunnel boring machine on earth, a title it still holds [3].

Then it did something stranger than being big. After 646 metres it broke through into an intermediate shaft, and rather than being hauled out, it was rebuilt in place into a smaller machine — 13.95 metres — reusing the same drive and the same trailing gear, and carried on [4].

The machine was called Qin Liangyu, after a Ming dynasty general. It was built by a company from Schwanau, a village of about 7,000 people on the Rhine plain in Germany’s south-west. And it was assembled at that company’s factory in Guangzhou, China — a detail worth holding on to, because the firm that built it now describes itself as the last manufacturer of tunnel boring machines left in the Western world [5].

The Niche

A tunnel boring machine is not really a product. It is a one-off, project-specific factory that happens to move.

At the front, a rotating cutterhead presses against rock or soil with thousands of tonnes of force. Behind it run more than a hundred metres of trailing machinery that erects the concrete lining segments, hauls away the excavated ground, holds the tunnel face under precisely controlled pressure so the ground above does not collapse — sometimes under a river, sometimes under a live city — and navigates by laser toward a target kilometres away. Herrenknecht builds them from 0.1 metres in diameter up to 19.

The precision is the part that does not sound plausible. Four of its gripper machines bored the Gotthard Base Tunnel through the roots of the Alps; at the first breakthrough in September 2006, after four years and 13.5 kilometres of excavation, the machine met its target 5 centimetres off horizontally and 2 centimetres off vertically [10].

What makes the niche a niche is that almost none of this transfers. The company’s own chief engineer, Werner Burger, puts the constraint plainly: it comes down to the ground conditions, and unless you are boring parallel tubes, no two projects meet the same ground twice [13]. Alpine granite, Shanghai mud and London clay each demand a different machine, so nearly every machine is bespoke, and a machine costs on the order of €10 million — of which the steel is the smallest part. The price carries the engineering, the workshop assembly, the transport, the first filling, the cutting tools, the seals and the software [12].

Even the consumables run at a scale that is hard to picture. Schwanau refurbished its twenty-thousandth cutter roller in December 2025. Each one weighs about 300 kilograms, and the cutting rings last a few hundred metres of advance — in the hardest rock, only a few metres [19].

And failure is unaffordable. A machine stuck under a river can sink a billion-euro project, which means buyers do not shop on price alone. They shop on the confidence that the machine will work, and that an engineer will be standing on the jobsite when it does not. Too small, too custom, too unforgiving: exactly the shape of market that conglomerates decline and one obsessive company can own.

The Origin

Martin Herrenknecht was born in 1942 in Lahr, the son of an upholsterer; his elder brother inherited the upholstery shop. He took his engineering exam at Konstanz in 1964 and spent the next seven years as a design engineer and project manager — road-building machines in Switzerland, then Toronto, then John Deere in Mannheim, then a design department in Karlsruhe [9].

The decisive years were 1971 to 1975, when he ran the mechanical engineering service on the Huttegg section of the Seelisberg road tunnel in Switzerland [9]. The American shield machine on that site — reportedly the largest in the world at the time — kept breaking down and jamming in the rock. Where others saw a machine that did not work, he saw an industry that did not exist yet.

In 1975, aged 33, he set up a one-man engineering office in Lahr [9]. When he wanted to build machines of his own, the banks turned him down and his mother Elsa lent him the 25,000 Deutsche Mark of start-up capital; Herrenknecht GmbH was founded in 1977 [8]. The first order came in 1978, a sewage canal in Luxembourg. By 1979 there were six employees and a first million in sales [8].

He has never dressed up how slow it was. Asked years later about the early period, he said simply: “Niemand wollte unsere Maschine.” Nobody wanted our machine. Asked whether there were moments of despair, he answered: “Kann man sagen.” You could say that [8].

The company moved to Schwanau in 1980 and became a — deliberately unlisted — public company in 1998 [9].

The Climb

1985 — the technology bet. Working with the contractor Wayss & Freytag, Herrenknecht helped develop the first large Mixshield, a slurry machine that uses a compressed-air cushion to hold pressure at the tunnel face with unprecedented precision [23]. That was the key to tunnelling through soft, waterlogged ground under high pressure — which is to say, under rivers and under cities, exactly where the next four decades of megaprojects would be.

1997 — the giant leap. For the fourth tube of Hamburg’s Elbe tunnel came “Trude,” a Mixshield of 14.2 metres, at the time the largest tunnel boring machine ever built. Its cutting wheel has been on display at Hamburg’s Museum der Arbeit since 2001 — and the cutting wheel is the part the museum credits specifically to Schwanau [24]. After Hamburg, when a project was record-breaking, the phone in Schwanau rang first.

2000 — going where the digging is. The first machine went to the Beijing metro; by 2002 the company had opened its own plant in Guangzhou. Today Herrenknecht runs three manufacturing and assembly sites plus two sales and service sites in China [1]. While much of German industry was still debating globalisation, a village firm from Baden had built a Chinese operation.

2017 — the day it went wrong. On 12 August, during a drive near Rastatt on the Karlsruhe–Basel line, the ground above the tunnel subsided and the railway track collapsed. To stabilise it, roughly 10,500 cubic metres of concrete were pumped into 160 metres of tunnel, deliberately entombing the machine inside — a €18 million, 90-metre boring machine, buried on purpose to save a main European freight corridor [11]. The line was shut for seven weeks.

Burying a machine is not itself the disaster it sounds like; it is routine. On Crossrail, two 560-tonne Herrenknecht shields were deliberately left under London, filled with foam concrete, their main drives and screw conveyors abandoned in place because recovering them was not worth the cost [25]. What made Rastatt different was that the ground moved first and the railway went with it. That is the risk the customer is really buying insurance against, and it is why they do not shop on price.

2024 — the high-water mark. Order intake hit an all-time record of €1,473 million and revenue reached €1,288 million [1]. Even then the company missed its own revenue target, for a revealing reason: finished machines could not be shipped because customers’ construction projects had slipped [1].

2026 — the crowning order. In July, National Highways selected Herrenknecht to build a 16.4-metre Variable Density machine for the Lower Thames Crossing — the largest tunnel boring machine ever used in Europe, over 5,000 tonnes, and the largest the company has ever built at Schwanau. One machine will bore the first 2.6-mile tube, be turned around underground, and drive the second [14].

The Market Nobody Can Measure

Here is where a profile of this company usually reaches for a number. Herrenknecht holds 73 per cent of the world market, say the encyclopedias; 55 per cent, says an older profile.

Both numbers are worthless, and the best evidence for that comes from Herrenknecht itself.

In its audited management report — filed, signed off by Deloitte, and repeated in two consecutive years — the company states that because no neutral competitive or global market analyses exist for the mechanised tunnelling sector, Herrenknecht declines to make statements about its own market share. It says the same thing again at segment level: manufacturer-neutral market share data is not available [1]. The 73 per cent figure traces back to a 2014 newspaper piece in which both the numerator and the denominator were supplied by Herrenknecht. The company has since, in a statutory document, disowned the exercise.

That refusal is more interesting than any number would have been, because the thing being measured has quietly split in two.

On units, China is not close to losing. Chinese state media put China’s share of global tunnel boring machine output at around 70 per cent, ranked first in the world for eight consecutive years [16]. The two Chinese majors are enormous: CRCHI sold 198 machines in 2025, CREG’s parent reports its subsidiary at ¥8.2 billion of revenue. On value per machine, the picture inverts. Herrenknecht’s tunnelling revenue is roughly level with CREG’s on a fraction of the units — which is a polite way of saying its machines cost several times as much, because they are the ones sent into the ground nobody else wants.

So “market leader” is two different claims, and the honest version is narrow: in Europe, the Americas, the Middle East and Asia-Pacific outside China, Herrenknecht technology is what predominantly gets used on traffic tunnelling projects — which is how the company itself phrases it, carefully, in the same filing [1].

The business also has a second half that almost nobody writes about, and it is measurable even though the market is not.

Ten kilometres from headquarters, in Kehl on the Rhine, sits a remanufacturing plant of more than 100,000 square metres where around a hundred people process over 10,000 components a year — 4,900 tonnes of machine in 2024 alone [26]. Olivier Kraft, the board member for traffic tunnelling, gives the number that matters: outside China, two-thirds of Herrenknecht’s machines use remanufactured components [12]. Used machines are bought back at 10 to 15 per cent of their original price [12], and the scale of that commitment is one of the few hard numbers in the accounts: at the end of 2024 the parent company carried €227.9 million of contractual buy-back obligations on machines still in the ground, up from €204.4 million a year earlier [1].

It is not a charity operation — Kraft is blunt that contractors expecting a 50 per cent discount on remanufactured parts get something closer to 10 [12]. What it buys is position. Service was 13.6 per cent of the parent company’s revenue in 2024, up from 12.3 per cent the year before, and higher on the big traffic-tunnelling machines than on the small ones [1]. A competitor selling a cheaper new machine is bidding against a supplier who will buy the thing back, rebuild it, and still be there in fifteen years.

Why They Win

Hermann Simon’s hidden champion traits fit this company almost too neatly — ambition to be number one, narrow focus, globalisation from a small base, closeness to customers, independence, leadership continuity. Three carry the story, and the fourth is where it gets interesting.

The niche is narrow; the depth is bottomless. Herrenknecht makes machines that dig holes. That is it — and within that, everything: hard rock and soft ground, 0.1 metres to 19, plus the navigation, conveyors, segment moulds and erectors around them. Even the diversifications are the same competence pointed elsewhere. Deep drilling for geothermal energy went from €1 million of revenue in 2022 to €48 million in 2024 after the company bought the drilling specialist H. Anger’s Söhne [1]. Mining, offshore wind foundations and power-cable tunnels are all, underneath, the same problem: making holes in the ground, precisely, under conditions that punish error.

Independence is structural, not sentimental. The company is a public company that never went public. In 2015 Martin Herrenknecht transferred his shares into a family foundation, with family members waiving their statutory share of the inheritance, so that the firm cannot be broken up when he dies [15]. The balance sheet is built to match: equity ratio 49.1 per cent, bank debt negligible, and dividends deliberately small — €11.4 million paid out from the 2023 result while €30 million went to reserves [1]. That is a structure designed to make the company un-buyable and un-pressurable, and it has been tested twice with a real number attached. In 2018 he turned down €400 million for 25 per cent plus one share: “We’re not selling” [6]. In December 2025 he raised the stakes himself, telling Handelsblatt that the Chinese had already offered him two billion euros for the company [7].

Closeness to customers is the actual product. A machine without support is scrap metal in a hole. Herrenknecht fields more than 600 operative service specialists and runs close to 80 service and sales locations worldwide; engineers assemble the machine on site and stay through the drive [27]. On a metro drive in Mecca the company went further and duplicated the machine’s control stand in Schwanau, synchronised with a jobsite some 4,000 kilometres away and watched around the clock [27]. That is the part a lower price does not reach.

And then the deviation. Simon’s champions are proverbially quiet — publicity-shy engineers in small towns. This one is the opposite. Its founder gives combative political interviews, campaigns publicly for tariffs on Chinese imports, and in December 2025 called for cutting German sick pay from six weeks to three [7] — a proposal aimed squarely at his own workforce. In October 2025 he had roughly twenty concrete planters that his village council had installed as traffic calming removed from the road outside his factory and dumped on company land, telling reporters the planters were “ein rotes Tuch” — a red rag [22]. The template says these companies win by being invisible. This one won while being the loudest man in Baden.

The Cracks

Succession is the sharpest risk, and it is unresolved. Martin Herrenknecht was born in June 1942 and was still chairman of the management board in the accounts filed in January 2026, still giving interviews as chief executive at 83. There is no dated handover statement anywhere in the public record; German Wikipedia records that he planned to retire in 2019, then 2020, and did neither. His son Martin-Devid, born 1986, joined the board in August 2022 and now runs production, purchasing and the mining division [1][21]. The foundation settled who owns the company. It did not settle who runs it.

China is a risk from both directions at once. Every Western rival of consequence was sold, and mostly to China: Aker sold the Wirth tunnelling technology and brand to China Railway Tunnelling Equipment in 2013 [17]; NFM in France went to a Chinese state group and was liquidated in 2020; Lovat was sold on to a Chinese buyer and its production moved to China. But the “last Western manufacturer” line needs an asterisk the company never adds. Robbins came back — Lok Home’s Global TBM Company bought the assets out of receivership in February 2021, returning the brand to American ownership [18] — and Bessac in France, owned by VINCI’s Soletanche Bachy, still builds tunnelling machines. Herrenknecht is the last Western manufacturer at the top of the market, which is a real and defensible position, but it is a claim about scale, not a headcount of survivors.

Meanwhile the hedge is also the exposure. Herrenknecht is itself a large Chinese manufacturer: three plants, and a workforce there in the hundreds [1]. Its record-breaking Hong Kong machine was built in Guangzhou. That is a cost hedge against Chinese competitors and, simultaneously, a hostage — and a company lobbying Berlin and Brussels for tariffs against China is lobbying against a country in which it manufactures.

The business is cyclical and the cushion is thin. In 2019 delayed European projects cut order intake by 27 per cent in a single year [20]. In 2024 European order intake fell from €695 million to €428 million even as Asia-Pacific surged 44 per cent [1]. And the most recent full year went backwards on both counts: FY2025 revenue fell 6.7 per cent to €1,202 million and order intake fell 8.8 per cent to €1,343 million [2] — against guidance, issued a year earlier, of a slightly higher revenue level [1]. Profit before tax runs a little under 7 per cent of revenue [1]. That is a solid business and a thin cushion to fight a price war from.

Takeaways

A moat made of accumulated experience is real, and it is unmeasurable — which means it erodes without a warning light. There is no metric on Herrenknecht’s dashboard that would tell it the day its lead stopped being decisive, because the company itself says the market cannot be measured.

“Market leader” is two claims, and you should always ask which axis. On units Herrenknecht is nowhere near first. On revenue per machine it is far ahead. Both statements are true, and a competitor who understands the difference is more dangerous than one who does not.

The most credible number in a company’s file is often the one it refuses to give. A firm that declines to claim a market share in an audited document, while its own marketing history is full of them, has told you something about its judgement.

Ownership structures solve inheritance, not succession. A foundation can make a company impossible to sell or split. It cannot make an 84-year-old hand over the sales function.

Being the last Western maker of something is a market position and a political position, and they pull in opposite directions — particularly when your own factories are in the country you want tariffs against.

---

One question, if you have an answer: Herrenknecht’s engineers say no two tunnels ever meet the same ground twice, so nearly every machine is built once and never repeated. Where else does a business survive on work that can’t be standardised — and what stops a cheaper rival from simply learning it?