A massive 1% fuel saving sat on a shelf for 30 years. Here is why

This is article 8 of 8 in The Third Number series. Here is article 7.

Thirty years on a shelf.

A Boeing 777-300ER goes into a hangar for a week, and technicians press roughly three thousand patches of a fifty-micrometre film onto the fuselage and the engine nacelles.

Each patch is turned by hand so its ridges line up with the direction the air will travel across that piece of skin at cruise. Nine hundred and fifty square metres, about forty percent of the aircraft’s surface. The ridges stand fifty micrometres high, millions of them, and if you ran your hand across the finished skin you would not reliably feel them.

Then the aircraft goes back into revenue service and burns about one percent less fuel. And keeps burning about one percent less fuel, every flight, for decades.

The world waited thirty years for that hangar, and the waiting had nothing to do with the science.

The science was finished in 1978

Not the film. Not the certification. Not the marketing. The physics.

NASA Langley had been working since 1972 on the thin, churning skin of air that clings to anything moving through it, which is where most of the friction that slows an aircraft actually lives. The reason is written into the primary source in language that could have been published this year: the rising energy costs of the early 1970s made the public aware of the need for new concepts to reduce energy usage.

That is the oil shock talking, not the ocean.

Walsh and Weinstein published the first riblet paper in 1978, and through the early 1980s Walsh optimised the geometry and reported roughly eight percent drag reduction from a surface of small longitudinal grooves. Twenty years of wind tunnel time, public money, and the full apparatus of twentieth-century fluid dynamics, and at the end of it, a number.

Then, in 1982, a palaeontologist at Tübingen named Reif, working with Dinkelacker, published on the hydrodynamics of fast-swimming sharks and noticed something.

Shark skin is not smooth. It is armoured in millions of tiny tooth-like scales, and every one of them is grooved. The spacing of those grooves fell almost exactly inside the band NASA had spent a decade computing.

Two searches, run independently, with nothing in common. One had four billion years, a planet’s worth of parallel trials, and no theory at all. The other had twenty years, a budget, and the equations.

They arrived at the same answer.

Two independent searches landing in the same place. Shark denticles are grooved, and the spacing of those grooves falls almost exactly inside the band NASA spent a decade computing. One search had four hundred million years and no theory. The other had twenty years and the equations.

I want to be careful here, because this is nearly always told backwards, and backwards it is a weaker story. The sharks did not give us riblets. The sharks confirmed them.

And a confirmation from a source that has been running the experiment for four hundred million years is not a charming footnote to an engineering result. It is the strongest thing an engineering result can have. When an evolutionary optimum and a computed optimum land in the same place, we have stopped talking about inspiration and started talking about biomimetic solutions.

Moving air does not care whether the grooved surface beneath it grew on a shark or came off a roll in a German chemical plant.

And then nothing happened for thirty years

By the mid-1980s they had everything. Bechert’s group at DLR and TU Berlin built an oil channel and worked the geometry against shark morphology, cutting friction at the skin itself by roughly ten percent, and Lufthansa Technik still calibrates its turbulence model in that same channel today. The answer was confirmed, published, optimised. 3M made a product. The film went onto Stars and Stripes, which won the America’s Cup in 1987.

Put yourself in that laboratory in 1987 and the conclusion is obvious, and I want to be fair to the people who drew it, because I have drawn it myself more than once. They believed that proof was the hard part. Twenty years of public money and wind tunnel time had produced a number, an independent four-hundred-million-year trial had landed on the same number, a manufacturer had turned it into a product, and the product had just won the most visible sailing race in the world. Everything a rational engineer understands as the finish line had been crossed.

They were wrong, and the way they were wrong is the whole subject of this series. Proof was never the hard part. Proof is the cheap part.

NASA’s own database carries the epitaph on its own technology, in the flattest language available: product is no longer commercially available.

The physics did not change. The sharks did not change. What was missing was not knowledge and not proof. It was that nobody could put a number (the Third Number) on what the thing was worth over the life of the asset it was stuck to. So nobody bought it, and so it died.

Thirty years is a long time to hold a solved problem, and every one of those years had aircraft flying with fuel going out the back that did not need to. Nobody has ever added that quantity up, because there was no line to add it on. Even though right outside the window the ledger existed to write it on.

What it looks like when somebody counts

Lufthansa Technik and BASF brought the film back as AeroSHARK, certified it with EASA in November 2019, and then had the singular misfortune of launching an aviation product into a year when a COVID world parked its fleet.

By April 2026 it was flying on more than thirty aircraft across seven airlines and four Boeing variants, with more than 350,000 flight hours behind it. SWISS finished its twelve 777-300ERs in May 2024, the first complete long-haul fleet in the world. And in February 2026, LATAM, the first customer outside the Lufthansa group, having flown one modified aircraft for most of a year and checked the promise against its own operation, put the film across its entire 777-300ER fleet.

Here is what got counted, please read it slowly.

The aerodynamic effect is nine tenths of one percent, measured against the same aircraft’s own drag before modification, at more than thirty thousand data points per flight, corrected for the way the wings flex in flight, for weight shifting as fuel burns off, for changing wind, for the sideways push the earth’s rotation puts on anything travelling across it, and for the fact that gravity is not quite the same over the equator as it is over Zurich, to a stated precision of one tenth of one percent.

When the airline then reduces the fuel it loads by the amount it no longer needs to carry, the saving compounds into itself and the total reaches one point one percent.

One aircraft, one year: roughly 325 tons of jet fuel not burned, more than a thousand tons of carbon dioxide not emitted, and more than a quarter of a million euros not spent. Across the whole fleet flying today, cumulatively, more than 22,000 tons of fuel and more than 70,000 tons of carbon dioxide.

Nine tenths of one percent, measured to a tenth of a percent, on a surface you cannot feel with your hand.

That is what a counted Third Number credit looks like. It is not enormous. It is real, it is physical, it is verifiable by anyone with access to the flight data, and it accrues every single day the aircraft flies.

So ask when the counting could have started

The measurement did not require anything invented after 1985. Fuel burn was metered. Flight hours were logged. Carbon per tonne of kerosene is arithmetic.

Everything needed to post this credit existed decades before anyone posted it. The reason it went unposted is that no one had agreed what would be counted, over what boundary, across what horizon. So there was no place to put the answer, and no way for a buyer to compare one supplier’s claim against another’s. Which to me sounds like a familiar problem solved by the chip industry decades ago.

You can still see the wound today. Lufthansa Technik prices the film against each customer’s own saving, aircraft by aircraft, negotiation by negotiation, because there is no comparable unit in the market to price it against. Its own sales lead reaches for the only object in the buyer’s world that is the right size, and tells the trade press the return is comparable to an engine upgrade.

Every one of those conversations is a private rediscovery of a number that could have been public in 1985.

That is the cost of the missing count, and it is not a philosophical cost. It is denominated in tons of kerosene, in euros, and in years.

And this field cannot even size itself

Which brings me to the most uncomfortable thing I know about the industry I am trying to help build.

An earlier draft of this series opened with a market figure. The biomimetics market is worth roughly thirty-four billion dollars. The edits removed it because it isn’t a fact. It is one firm’s answer to a question nobody has defined.

For 2025, one research house sizes the biomimetic materials market at about 2.6 billion dollars. Another sizes the bioinspired materials market at about 50.2 billion. Same year, nearly the same category name, a ratio of nineteen and a half to one.

I found a version that removes every excuse. One firm titles a report “Biomimetics Market” and puts 2024 at 33.59 billion dollars. Another titles a report “Biomimetics Market” and puts 2024 at 8.32 billion. Same title, same base year, same listed segments. Four times apart, because the thing they claim to count has no definition either is obliged to honour. No standard for a Third Number that can also be generational.

Now search the North American Industry Classification System, the official Census tool, for biomimetic. Zero records. Biomimicry. Zero. Bionic. Zero. Bio-inspired. Zero. The legacy Standard Industrial Classification. The United Nations ISIC. The European Union’s NACE. The Harmonized Tariff Schedule governing goods crossing borders. Nothing.

And a zero result means nothing without a control, so here is one we can use. Search the same system for nanotechnology and you get twenty-two records, including a code for nanotechnology research and development laboratories.

The contrast is almost cruel. Semiconductors have had their own code since the system was adopted in 1997, and one before that. The Bureau of Labor Statistics publishes their employment, and a monthly producer price index for the industry, continuously, since January 1967.

You cannot measure an industry the statistical machinery of the world does not recognise as one. The nineteen-fold spread is not sloppiness by the analysts. It is the arithmetic that necessarily results when a dozen firms count a thing nobody has defined.

That immeasurability was my initial diagnosis, and it is the strongest evidence for everything this series has argued.

Recognition precedes valuation. You cannot price what you have not first agreed to count. We want to create that shared count.

A man who gave up a patent so a standard could exist

There is a precedent for what has to happen next, and it is not from my industry of origin.

Before the middle 1960s, freight moved break-bulk, piece by piece, and the shipping containers that did exist belonged to whoever built them. McLean’s Sea-Land ran thirty-five-foot containers. Matson ran twenty-four. Every carrier was internally consistent and no two could hand cargo to each other, so a box often had to be emptied and repacked at the very point in the journey the box was supposed to make unnecessary.

Cost, delay and loss reappeared exactly where the innovation was meant to remove them.

The International Organization for Standardization formed a technical committee in 1961. ISO 668 was published in January 1968, defining terminology, dimensions and ratings. Between that and the identification standard that followed, those documents turned a set of incompatible carrier networks into one system. A box built to the standard can be lifted by any compliant crane, stacked on any container ship, and carried on any rail wagon or truck chassis anywhere in the world.

The emerging decision: Malcolm McLean released his patent on the container corner post so that the standard could exist.

The corner casting is the piece that lets containers be lifted and stacked and survive being stacked. It was structurally the whole invention, and it was his. He gave up proprietary control of it because a standard that belonged to Sea-Land would have been a Sea-Land sales document, and every competitor would have read it that way and built something else.

Levinson’s history estimates the cost of moving a ton of cargo fell by something like ninety percent between the early 1960s and the 1980s. Levinson himself wrote that how much the container mattered to the world economy is impossible to quantify. Econometric estimates since have found effects ranging from economically large to delayed to modest. Direction uncontested, magnitude fought over, which is where we always end up when something matters.

Around ninety percent of non-bulk world trade now moves in standardised steel boxes.

What a standard has to be able to do

I want to narrow the claim, because there is a counterexample sitting inside my own field.

Sustainability reporting has never suffered from a lack of standards. It has had the Global Reporting Initiative since 1997, the Carbon Disclosure Project since 2000, and a dozen national and sectoral schemes besides. Thirty years of standard-setting, by serious people, with substantial institutional backing.

And the label collapsed. Among S&P 100 companies, the share publishing standalone reports with ESG in the title peaked at around 40 percent in 2023, fell to roughly 25 percent by 2024, and was down to about 6 percent in the reports published in 2025.

Standards were not the missing ingredient there. There were too many, and that is only half the problem. The deeper failure is that almost all of them were self-reported, weighted into a composite, and averaged, so a strong score in one category absorbed a failure in another and the output came out respectable no matter what went in.

The fundamental problem is that a standard that cannot fail a project disciplines nothing. It certifies.

So the claim is narrower than standards make industries, and it should be. A standard makes an industry when it is specific enough to fail a project, held by a body with nothing to sell, and wired to a consequence somebody actually pays.

Those are three conditions we can meet, and meeting them is institutional work rather than scientific work, that an international association can do.

What we are actually asking

The queue behind the riblet film is long and getting longer.

An enzyme in Clermont-Ferrand that takes PET plastic apart at ordinary temperature and pressure and returns it to feedstock as good as virgin. A demonstration plant running since 2021, a commercial plant broken ground in 2024, and construction stopped. Not because the enzyme failed. Because a lender was being asked to underwrite the difference between a tonne of PET that goes back into the economy and a tonne that does not, over decades, in a market where nobody has ever agreed how to count that difference.

A pipe in Quebec carrying enough warm water to grow eighty-two thousand tons of tomatoes, with nothing financed at the far end.

Two point two million predicted crystal structures from one paper in 2023, of which 736 have been made in a laboratory by anyone, and as of early 2026 none has reached commercial production.

The rate at which we can propose has gone up by two or three orders of magnitude. The rate at which we can make, test, qualify, permit, finance and deploy has gone up by very little. Anyone selling you the first number without the second is selling you a marketing slide. We are building an industry association and ecosystem to deliver the second.

And the gap between those two rates is not a science problem. Longlaville is not waiting on an enzyme. The Quebec pipe is not waiting on a heat pump. The 2.2 million crystals are not waiting on more crystals.

Every one of them is stopped at the same joint, the point where a physical claim has to become a financeable one. It is the driving need for the Third Numbers and recognition of the ledger that contains them.

Which is a strange kind of good news, and I do not want to oversell it. It means the binding constraint on this entire field is a piece of institutional plumbing rather than a scientific breakthrough, and institutional plumbing is the sort of thing a determined group of people with the advent of AI can install in a few months if they decide to.

We are not waiting on a genius. We are waiting on an agreement.

Where this started

Eight instalments ago this began at a town meeting that ended at ten fifteen with nobody satisfied, because two sides were measuring the same generational decision with instruments that could not read each other.

Everything since has been about building the missing instrument. Not to make anyone virtuous. To let capital recognise and price what the living world does for us and what we take from it, so that the design which actually fits gets financed rather than merely admired. That is the entire purpose of Biomimetics International.

The BiosVerse-Compatibility Index (BCI) works today and so does the Generational Cost of Ownership (GCOO) model. What is not settled is how they are governed. Who holds the standard. What counts as evidence, in what form, checked by whom. How a design gets failed, and who has the standing to fail it.

Those are not questions a laboratory answers. They get answered by the people who will have to live inside the answer, which means the people who make the drawings, sign the checks, run the towns, and farm the section next door.

The account always clears the Third Numbers on nature’s ledger and now ends up on our balance sheets sooner. The only thing still in our hands is which side of it we are posting the Third Numbers to, and whether we can agree, soon enough to matter, on how to measure them, count them, and make the entries.

What we are asking for

This is the last instalment, so here is the ask in full. The BiosVerse-Compatibility Index (BCI) scores whether a design fits. The Generational Cost of Ownership (GCOO) model prices what the fit or the misfit costs across generations. Both exist and both run today. As we have said throughout, what is not settled is how they are governed: who holds the standard, what counts as evidence, and who has the standing to fail a project. That is the work of Biomimetics26, on November 9 and 10 at the Minnesota Landscape Arboretum, and it is open. Bring what you know is buildable, what you have measured, and what you think the standard will get wrong. The third of those is the most useful.

This concludes the series. Look for the forthcoming book: The Third Number: Nothing We Build Goes Uncounted

illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.

The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.