Ten Ways We Read Industry.

Inconel 718 is a nickel-based superalloy used in jet engine discs, hot-section turbine parts, and rocket nozzles. It was invented by Herbert Eiselstein at Huntington Alloys (now Special Metals) and patented in 1962. The patent expired around 1980. Forty-five years after expiry, anyone in the world can read the chemistry and make their own version. Many do.

Special Metals (now owned by Precision Castcomponents Corp) still holds the majority share of the aerospace-grade market. The reason is not the formula. The reason is the cumulative know-how of producing it to aerospace certification standards lot after lot for over sixty years. Vacuum induction melting, vacuum arc remelting, homogenisation cycles, forging routes, heat treatment windows, inclusion control, lot-to-lot consistency. Every one of those is a learned capability. None are written in the patent. All matter to whether a jet engine disc made from your alloy is allowed to fly.

In the late 2000s, GE Aviation was developing a new generation of jet engine called the LEAP — a joint venture with the French engine maker Safran through CFM International. The LEAP was destined to power the Airbus A320neo and the Boeing 737 MAX, the two most commercially important narrowbody aircraft in the world. The fuel nozzle had previously been made by brazing together twenty individual machined parts. GE Aviation decided to redesign it as a single component, made by direct metal laser melting — fusing cobalt-chrome alloy powder one twenty-micron layer at a time with a laser.

The result was a single part instead of twenty. About 25 percent lighter than the conventional version. Approximately five times more durable in operation. Brazed joints, which had been the highest-risk failure mode of the old design, were eliminated entirely. By 2024 GE had produced approximately one hundred thousand of these nozzles, every one of them flying in commercial aircraft over millions of flight hours.

What changed at once: the design moved from twenty parts to one, the manufacturing process moved from machining-and-brazing to laser-powder fusion, the materials specification was rewritten for additive construction rather than wrought, the inspection methods shifted from optical and dye-penetrant testing of welds to high-resolution computed tomography of the entire part, and the qualification pathway through the Federal Aviation Administration required first-of-kind safety approvals that took several years and set precedent for everything that has followed. A new entrant cannot match GE on any single one of these layers without matching the others as well.

In 1962, Bristol Siddeley Engines invented a programme called Power-by-the-Hour for the Viper jet engine on the de Havilland Hawker Siddeley business jets. Instead of selling the engine and then selling spares and overhauls separately, they sold a fixed-price service per hour of flying time. If the engine ran for a thousand hours a year, the customer paid a thousand units. If it ran for two thousand, they paid two thousand. Rolls-Royce inherited the model when they absorbed Bristol Siddeley in 1966 and turned it into TotalCare in the 1980s.

The numbers are now significant. Civil Aerospace services revenue is roughly equal to its product revenue. About half of all Rolls-Royce Civil Aerospace turnover comes from TotalCare contracts rather than engine sales. The economic shape of the business changed from selling capital equipment once every twenty years to selling availability continuously. Capital expenditure on the customer side became operating expenditure. Risk of unplanned downtime moved from airline to engine maker. The engine maker had every incentive to design for reliability because they were now paying for failure.

This was not substitution. Nobody used to buy thrust-by-the-hour and now buys it from Rolls-Royce instead of someone else. Power-by-the-Hour created a category that did not exist. Customers willingly paid for it because the value to them — predictable cost, reduced downtime, no spares inventory — was higher than the price. Both sides gained.

Beckhoff Automation is a family-owned industrial controls business based in Verl, in north-west Germany. It was started in 1953 by Arnold Beckhoff in the back of an electrical goods shop. His son Hans took it over in 1980, built it into one of the largest industrial automation businesses in the world, and is still managing it. Revenue in 2024 was approximately €2.1 billion. The business employs about 5,300 people across 75 countries. Roughly 4 percent of revenue goes into research and development every year — about €80 million annually.

Hans Beckhoff is a modern industrialist. He has a doctorate in physics. He spent decades on the shop floor. He understands controls, software, networking, mechanical engineering, and commercial strategy at depth. In 2003 his company released a communications standard called EtherCAT, which is now the dominant way industrial machines talk to each other on a factory floor — over 105 million devices deployed globally as of 2025. The semiconductor factories that make the chips inside the GPUs used to train AI models run on EtherCAT.

The point is not the product. The point is that one person, over four decades, with patient family ownership and a regional industrial ecosystem, built a quietly dominant business in a market with massive global stakes. That is what a modern industrialist looks like when given the conditions to operate.

The Siemens factory in Amberg, Bavaria makes Simatic programmable controllers — the brains of industrial machines. It has been running since 1989. About 1,200 people work there. They produce roughly 17 million products per year. The defect rate is approximately 12 defects per million products — meaning if you ordered a million of their products, you would expect about 12 of them to have a problem.

The plant is roughly 75 percent automated. The other 25 percent is human. The humans do the things humans are still better at — exception handling, judgement calls, interaction with customers and suppliers. The machines do the things machines are better at — repetitive precision tasks, around-the-clock operation, perfect record-keeping. The integration between digital and physical happens at every level. Every machine reports its state continuously. Every product has a digital twin from the moment its components arrive. Every defect is traced backwards to its root cause within minutes, and the production line adjusts before the next batch.

The plant did not start this way. In 1989 it was a normal Siemens factory. It has taken thirty-five years of continuous improvement, software upgrades, automation investment, and workforce training to reach the current state. The defect rate is now hundreds of times better than industry average. The output per employee is many times higher than industry average. Neither of those was the result of buying a piece of software. Both were the result of integrating digital and physical capability over decades.

Caterpillar makes the yellow earth-moving machines, generators, and industrial engines that show up on every construction site and mine in the world. In 1973 they started a business called Caterpillar Reman, which takes worn-out engines and transmissions from old machines, disassembles them, inspects each part, remanufactures the parts that can be saved, replaces the parts that cannot, and sells the result as a Reman product at roughly 60 to 70 percent of the price of a new equivalent. The Reman product carries the same warranty as a new product.

The numbers are real. Caterpillar Reman is a roughly $4 billion-a-year business. It operates around 70 facilities worldwide. Every remanufactured engine uses approximately 85 percent less raw material than a new one and produces about 86 percent less carbon dioxide over its lifecycle. The economics work because the original engine was designed with remanufacture in mind. The crankcase castings are heavy and durable. The bearing surfaces are designed to be resurfaced multiple times. The connecting rods are designed to be inspected and refurbished. Every design choice trades some original-build cost for higher remanufacture yield decades later.

Caterpillar customers love it because they get warranty-grade equipment at lower cost. Caterpillar loves it because the lifecycle revenue per machine is much higher than it would be if every replacement was a new build. The environment benefits because raw material consumption falls dramatically. Three-way win that took fifty years of design discipline to build.

The Global Positioning System is a constellation of around 31 satellites operated by the US Space Force. Every smartphone, every car satnav, every shipping container, every aircraft, every drone uses it. The civilian use is so universal that most people never think about the military origin. About six billion GPS receivers are currently in use globally. The economic value of civilian GPS to the world economy is estimated at over $300 billion per year.

What is interesting for our purposes is how the architecture handles the dual-use problem. The same satellites broadcast two signals at once. One is the Coarse/Acquisition code, which is open to anyone in the world with a $5 receiver chip, and gives accuracy of about three metres. The other is the Precise code, which is encrypted, only readable by US military and approved allies, and gives accuracy of better than one metre with anti-jamming and anti-spoofing protections. Same satellites. Same hardware. Two coding regimes. Two access tiers. Total economic value distributed across both. National security maintained on the upper tier. Civilian utility maximised on the lower tier.

The architecture took thirty years to evolve. The system was conceived in the 1960s, became operational in the mid-1990s, had civilian access deliberately degraded by something called Selective Availability until 2000, and has since added more signal bands and more accuracy. The point is not that any of this is simple. The point is that it has been done before in another domain. There is no first-principles obstacle to applying the same architecture to civilian industrial capability.

In 1856 a Swedish banker called André Oscar Wallenberg founded the bank that would become Stockholms Enskilda Bank. His descendants spent the next 170 years compounding capital into a portfolio of long-held industrial assets. Today their main vehicle is Investor AB, a publicly listed holding company with major positions in ABB, AstraZeneca, Atlas Copco, Ericsson, Saab, Electrolux, and others. Net asset value is approximately $130 billion as of 2024. The family motto, attributed to one of the early Wallenbergs, translates roughly as "to be, not to be seen."

The Wallenberg model is patient. They typically hold positions for decades. They take board seats. They influence strategy across generations of executives. When one of their companies needs to make a multi-decade bet — a new factory, a new technology platform, a new market entry — the Wallenberg presence on the board makes it easier to fund. Their explicit purpose is to ensure that these companies exist for the next hundred years. Compare that to a venture fund whose explicit purpose is to return capital within ten years.

The difference shows up in the kind of bets the companies can make. ABB can fund a twenty-year industrial automation roadmap. Atlas Copco can absorb a multi-decade integration of an acquisition like Edwards. AstraZeneca can fund a fifteen-year drug development pipeline. None of these would be possible if the controlling shareholder needed exit liquidity in five years. Patient ownership is itself a competitive advantage at the industrial scale.

EtherCAT is the industrial communications protocol mentioned in Chapter 04. It is the way machines on a modern factory floor talk to each other in real time. The protocol was released to the public in 2003. Beckhoff retained no proprietary rights — anyone could implement it, free of charge — and set up an industry organisation, the EtherCAT Technology Group, to govern it. The strategy was to win on technical merit first and standardise the result.

The timeline is worth memorising as a benchmark for industrial standards adoption.

From release to ~30 percent market share took seven years. From release to today's dominance — over a hundred million devices, the standard on which the world's semiconductor fabrication runs — took twenty-two years. And EtherCAT is considered one of the faster successful industrial standards. By comparison, the Robot Operating System, which is now the de facto open standard for academic and research robotics, was released in 2007 and is still gaining ground in commercial robotics two decades later. The combustion engine took roughly thirty years from invention to ubiquitous adoption. Industrial change runs on decadal cycles, not annual ones.

Aerones is a Latvian company that builds large drones for inspecting and repairing wind turbine blades. A modern wind turbine blade is sixty to ninety metres long, exposed to weather, and prone to leading-edge erosion that reduces aerodynamic efficiency by 1 to 5 percent over a few years of operation. Repair traditionally requires a team of rope-access technicians who hang from the nacelle and grind, fill, and repaint the blade by hand. It is slow, dangerous, expensive, and weather-dependent.

Aerones built a drone-based system that does the same job. The drone hovers next to the blade, grinds the eroded area, applies filler, and recoats — all controlled from the ground. By 2024 they had serviced approximately 20,000 wind turbines globally. Their customers include Vestas, Siemens Gamesa, ENGIE, and Ørsted — the largest wind turbine makers and operators in the world. They have raised approximately $80 million in capital across multiple rounds. The business has clear customers today, clear growth into adjacent maintenance applications tomorrow, and clear potential to become the standard infrastructure for renewable energy maintenance as the installed base of wind turbines triples over the next ten years.

What is interesting is how each time horizon has a different customer profile. Today's customer is the wind turbine operator doing reactive maintenance. Tomorrow's customer — three years out — is the integrated services contract for entire wind farms across multiple operators. The ten-year customer is the standard inspection and maintenance infrastructure for offshore wind, which is the fastest-growing renewable category globally. The same capability serves three different commercial relationships, each one larger than the last.