Where is industrial spider silk? (Part 1)

Last updated: 2 November 2026 · 64 min read

Part 1 of 2 · Next: Where is industrial spider silk? (Part 2)

The Dream Fiber

The spider hangs at the center of its web, and if you shine a flashlight just right, the silk catches the light like fiber optic cable. Which, in a way, it is—each strand thinner than a human hair, yet capable of stopping a bee traveling at full speed without breaking. The web flexes, swallows the impact, and holds the bee instead of flinging it back out. The spider feels the vibration and comes to collect.

This tiny physics demonstration has obsessed materials scientists for thirty years.

The obsession is rational. Spider dragline silk has a tensile strength of roughly 1.0-1.5 GPa, comparable to high-grade steel, and it is about one-sixth as dense, which means a strand of spider silk is five times stronger than the same weight of steel. It’s tougher than Kevlar, the stuff in bulletproof vests, absorbing more energy before breaking. It can stretch by roughly a quarter to a half of its length before breaking, turning most of the absorbed energy into heat instead of springing back.

The spider makes it in her abdomen, at room temperature, out of digested bugs and water, with no factory, no petroleum and no vats of hot concentrated acid.

Governments, investors and lenders have poured hundreds of millions of dollars, by some counts more than a billion, into companies trying to copy it over the past three decades. None of them has yet turned it into an industrial fiber.

The Holy Grail That Refuses to Be Found

In 1990, Randy Lewis at the University of Wyoming published the sequence of a spider dragline silk gene. A decade later, in January 2000, the Montreal company Nexia Biotechnologies unveiled “BioSteel” goats engineered to secrete spider silk protein in their milk. The media went wild, with talk of soft body armor and a “milk of steel.”

Defense agencies took an interest and investors began writing checks. That was more than a quarter of a century ago.

You still can’t buy a spider silk bulletproof vest. You can’t buy spider silk rope, or spider silk parachute cord, or spider silk surgical sutures at scale. A handful of companies will sell you something made with recombinant spider silk or silk-like proteins, such as a limited-edition necktie, a luxury parka, a designer shirt, a watch strap or a skin cream, but in small volumes and at premium prices, often blended with conventional fibers. Nobody sells it by the thousands of tons, the way the fibers it was meant to replace are sold.

The puzzle is that we understand, in remarkable detail, what makes spider silk work. We’ve decoded its genes, mapped its molecular structure, and published thousands of peer-reviewed papers analyzing its architecture. We’ve successfully produced the protein in bacteria, yeast, goats, silkworms, and even genetically modified alfalfa.

And yet, after three decades of effort, hundreds of millions in investment, and a great deal of sophisticated biotechnology, spider silk remains a niche specialty material rather than an industrial fiber.

Something so thoroughly understood still can’t be made at industrial scale outside the spider.

Why Everyone Wanted This So Badly

To understand the obsession, you need to understand the gap in the materials world that spider silk seemed designed to fill.

Modern civilization runs on a surprisingly small number of high-performance materials. If you need something light and stiff, you use carbon fiber, which is brilliant for bicycles and aircraft but brittle. If you need something that absorbs impacts without failing, you use Kevlar; it saves lives in body armor, but it degrades in sunlight and slowly loses strength when kept hot. If you need something very strong by weight, you use ultra-high-molecular-weight polyethylene, excellent for ropes and cut-resistant gloves, but it creeps under constant load and melts at around 150 °C.

Every material trades off properties: high strength usually comes with brittleness, toughness with weight, flexibility with weakness. Spider silk appeared to break these rules.

Dragline silk sits in an unusual spot on the strength-toughness curve that engineered fibers have struggled to reach.

A strand of it, the kind the spider uses for its safety line and the radial threads of its web, has a strength comparable to high-grade steel (several times higher by weight) and, at the same time, a toughness that exceeds Kevlar’s.

This convergence created a rare moment of agreement across wildly different industries. The Pentagon wanted lighter body armor that could absorb more bullet energy. Textile manufacturers wanted biodegradable performance fabrics that didn’t require petroleum. Medical device companies wanted biocompatible sutures that the body wouldn’t reject.

Aerospace engineers wanted ultra-lightweight tethers and composites. Every one of these groups wanted spider silk.

The material seemed purpose-built for the 21st century: stronger than what we could synthesize, produced sustainably, and compatible with living tissue. In the flush days of the biotechnology revolution, when scientists were just learning to edit genes like software code, spider silk looked like proof that nature had already solved our hardest materials problems. All we had to do was copy the recipe.

The logic was seductive: evolution has been refining silk for close to 400 million years. We just needed to borrow the blueprint.

The “Perfect Material” That Wasn’t

The hype starts with a phrase you always hear, “stronger than steel,” which is technically true but misleading.

Spider silk is stronger than steel by weight, which is what engineers call specific strength. This matters enormously if you’re building aircraft or spacecraft, where every gram counts. It matters much less if you’re building a bridge or a building, where absolute strength and rigidity are what you need.

Materials scientists think about performance in three key dimensions: strength (how much load it bears, for its thickness, before it breaks), stiffness (how much it resists stretching or bending), and toughness (how much energy it can absorb before failing). You can visualize this as a three-way tradeoff. Carbon fiber owns the high-strength, high-stiffness corner but shatters under impact. Kevlar dominates the high-toughness zone but isn’t particularly stiff. Rubber is elastic but weak.

Spider silk combines good strength with exceptional toughness: it absorbs large amounts of energy without breaking, which makes it ideal for stopping flying insects or, in theory, dissipating impact forces.

But it’s nowhere near as stiff as carbon fiber or even high-grade steel. For applications requiring rigid structures, such as aerospace frames, automotive components and construction materials, spider silk doesn’t compete. It would flex and deform where you need something that holds its shape under load.

Then there’s the thermal and chemical stability problem. Kevlar does not melt; it only decomposes above about 425 °C, although it is rated for long-term use at around 150–180 °C. Carbon fiber tolerates even more heat in the absence of oxygen. Spider silk is a protein. Dry dragline keeps its strength from well below freezing to at least 150 °C, but water makes it shrink and soften, still a serious handicap next to aramids in demanding environments. Expose it to UV light for extended periods and it degrades. Hit it with certain solvents and it dissolves.

The early marketing never mentioned any of this. The “miracle material” narrative implied universal superiority: that spider silk was better than synthetic alternatives across the board. It suggested that once we figured out how to make it, every high-performance application would naturally switch over.

This was an oversimplification, and it points to something deeper: the seduction of biomimicry.

There’s an almost romantic belief in materials science that nature has already solved our hardest problems, that evolution, with its hundreds of millions of years of R&D, has optimized solutions we can barely imagine. Sometimes this is true. Velcro came from burrs. Sharkskin-inspired surfaces reduce drag. Gecko feet inspired new adhesives.

But spider silk became the cautionary tale—the example where “copying nature” stopped being clever engineering and became a trap. Evolution optimized for something else: a solitary predator that needs to catch flying insects using a structure it can produce from its own body, recycle when damaged, and deploy without external energy or tools.

Evolution did not optimize for: factories, profit margins, industrial throughput, quality control, regulatory approval, or cost per kilogram.

The spider doesn’t care that its silk production is “inefficient” by industrial standards, that the process only works at tiny scales, or that each strand requires nanoscale precision.

It works to no deadline, pays no wages, and recycles its mistakes by eating them. A factory has none of those luxuries.

The Cycle That Won’t Break

And yet, every five to seven years, the same headline returns: “Scientists Create Super-Strong Spider Silk.” The press releases follow a template. A research team announces a breakthrough in producing the protein, or a marginal improvement in fiber properties, or a new spinning technique inspired by the spider’s spinneret. Reporters call it a “game-changer.”

Defense magazines run breathless features, and venture capital firms schedule pitch meetings. Then nothing changes.

The companies that raised millions pivot to “adjacent markets,” which usually means cosmetics or wound care rather than body armor. The researchers publish their findings, note that “industrial scale-up remains challenging,” and return to their labs.

The cycle has repeated enough times that it has become its own genre of science journalism—the miracle material that’s always five years away.

Part of the reason it keeps happening is structural.

Spider webs are visually stunning. They practically film themselves. The spider-versus-bee video is catnip for science documentaries. The phrase “stronger than steel, lighter than a feather” is marketing gold. Add the word “biomimicry” and you have a story that appeals to technologists, environmentalists, and futurists simultaneously.

Spider silk offered every beat a deep-tech investor looks for: a revolutionary biomaterial, a massive total addressable market in defense, medicine and textiles, sustainable production, and a clear path to commercialization.

Every few years, a team does achieve something new. It gets the protein to express at higher yields in yeast, keeps it from clumping in solution, or designs a synthetic spinneret that gets a little closer to the spider’s natural process.

These are real advances, and they move the field forward. A lab demonstration showing 10% better fiber strength is legitimate scientific progress. That same result gets packaged into a press release about “next-generation body armor,” and the cycle begins again.

The problem is that moving the science forward and moving the manufacturing forward are not the same thing. Scientific progress is measured in publications and citations. Industrial progress is measured in tons per year and dollars per kilogram. The gap between a proof-of-concept in a university lab and a profitable product shipping at scale is where spider silk has died, repeatedly, for three decades.

The Gap That Won’t Close

We know how to produce spider silk protein in industrial quantities using genetically modified organisms. Companies have demonstrated this. The protein exists. You can buy it, in limited quantities, from specialized suppliers.

What we don’t know is how to transform that protein into a fiber that maintains the properties that make spider silk special—at a cost that makes commercial sense, at a speed that industrial production requires, with the consistency that regulated markets demand.

That gap, between a vat of expensive protein solution and a spool of usable fiber, has swallowed hundreds of millions of dollars and thousands of researcher-years.

The spider does it in her abdomen in a fraction of a second, and in one sense we know how.

We can describe the process in extraordinary detail. The spider’s silk gland is a chemical and mechanical marvel: it adjusts pH, manages ion gradients, applies precise shear forces, and triggers molecular self-assembly, all simultaneously, inside a gland and duct small enough to fit in a spider’s abdomen. We’ve mapped much of it at molecular resolution.

What we can’t do is replicate that process in a factory, at the speeds and volumes required to compete with nylon, whose raw resin costs about $2 per kilogram and which we produce in quantities measured in millions of tons per year.

The spider’s spinneret works because it’s small, because it handles tiny volumes, because it’s integrated into a living system that provides precise biochemical control. Scale it up, making it larger, faster and compatible with industrial equipment, and the physics breaks down. The fluid dynamics change. The flow conditions that gently align proteins at spider scale, a precise mix of shear and stretching timed to chemical triggers, have proved extremely hard to reproduce at factory speeds. The ion gradients that work in a microscopic duct become far harder to maintain in a pipe.

Our understanding of the mechanisms has not translated into engineering. The spider’s solution is suited to being a spider and badly suited to being a factory.

The spider silk industry has spent three decades on this problem: the material is extraordinary, but the manufacturing process that transforms the liquid protein into a solid fiber requires a level of nanoscale control that our best industrial equipment cannot achieve at economically viable speeds.

So far you can have spider-quality fiber at spider-scale speeds, producing grams per day at costs measured in thousands of dollars per kilogram, or you can have industrial-scale speeds, producing tons per day of a fiber that has lost the properties that made spider silk special in the first place. Strength and toughness drop, and you end up with expensive, mediocre synthetic fiber that can’t compete with Kevlar or even regular nylon.

It has been a trade-off between quality and scale, and nobody has yet shown a process that escapes it.

Why This Matters Beyond Spider Silk

The science held up. Spider silk works, and spiders make it continuously and reliably, in virtually every terrestrial ecosystem on Earth.

This is a story about the gap between scientific achievement and commercial viability—between what’s possible in a lab and what’s possible in a market. It’s about why “copying nature” is a seductive but often misleading strategy for engineers. It’s about the structural mismatch between venture capital timelines and materials development cycles, which the US Materials Genome Initiative puts at 10 to 20 years from discovery to market.

Most of all, it’s about the compounding difficulty of solving five hard problems at once: producing the protein cheaply, maintaining its structure, spinning it into fiber at industrial speeds, ensuring batch-to-batch consistency, and doing all of this at a cost that can compete with materials that have had half a century or more of manufacturing optimization.

The intense focus on replicating nature distracted the industry from the actual goal: creating a high-performance fiber that people would buy. Buyers would judge that fiber on performance and cost, not on whether it came from a spider gene or an entirely synthetic approach.

The surviving companies have learned this lesson. They’ve abandoned the pure biomimicry approach of trying to perfectly recreate the spider’s process in favor of bio-inspiration: borrowing principles while using entirely different manufacturing methods. Some have pivoted away from bulk fiber entirely, focusing instead on high-margin medical applications where a few grams of material in a surgical implant can sell for thousands of dollars, making the production cost irrelevant.

Others have moved beyond spider proteins, designing new proteins (and, in academic labs, synthetic polymers) that borrow silk’s molecular architecture: the block structure, the crystalline-amorphous balance. These materials will never be “true” spider silk, but they might make it to market.

The spider still hangs in its web, wrapping prey in a material we can admire but cannot replicate at scale. After thirty years, hundreds of millions of dollars in investment, and thousands of research papers, it is clear that the most elegant solution in nature can make a poor template for industry.

The industry wanted a fiber it could make by the ton, cheaply and identically every time, and nothing in the spider’s biology was built for that.

That mismatch, more than any single technical problem, explains why thirty years of work have not produced an industrial fiber.

Nature’s Masterpiece: What Makes Spider Silk So Special

If you took a strand of spider silk and looked at it under an electron microscope, you’d see something that looks unremarkable: a smooth, uniform cylinder, about five microns across. Zoom in closer, to the molecular level, and you’d find a precisely ordered structure.

What you’re seeing is nature’s solution to a problem that industrial chemists still can’t fully replicate: how to build a material that’s simultaneously strong, tough, and elastic, using nothing but protein and water, at room temperature, in a fraction of a second.

The answer lies in how the protein’s building blocks are arranged, from single molecules up to the finished thread.

The Molecular Blueprint

Start with the basics. Spider silk is a protein, or more precisely a family of proteins called spidroins. If you’ve taken high school biology, you might remember proteins as long chains of amino acids that fold into specific shapes to do jobs like carrying oxygen or regulating blood sugar.

Spider silk proteins do something different: they form structures.

Other structural proteins, such as collagen in your tendons and keratin in your hair, are built from repetitive chains too, and they get their strength from the way those chains twist together and cross-link.

Spidroins take a different route. They’re modular, almost like LEGO blocks, with distinct sections that serve radically different functions. Picture a long chain made of alternating segments: some sections are rich in the amino acid alanine, arranged in sequences that naturally want to form tight, crystalline sheets. Other sections are rich in glycine, creating loose, amorphous regions that remain flexible.

The alanine-rich blocks fold into what chemists call beta-sheets—flat, layered structures where the protein chains stack on top of each other like paper in a ream, held together by hydrogen bonds. These crystalline regions are strong and stiff. They’re the skeleton of the fiber, providing tensile strength.

The glycine-rich blocks do the opposite. They stay loose and disordered, forming amorphous regions that can stretch and deform. These are the fiber’s shock absorbers, providing elasticity and energy absorption.

On their own, neither structure is particularly special. Crystalline proteins are strong but brittle; they snap under stress. Amorphous proteins are flexible but weak; they deform permanently. But combine them in precise ratios, at precise intervals, along the same molecular chain, and you get a material that is both strong and able to stretch.

The crystalline blocks carry the load while the amorphous ones let the fiber stretch and absorb energy.

The Hierarchy That Makes It Work

The organization doesn’t stop at the molecular level. Spider silk is built hierarchically, with structures nested inside larger structures, each level adding new capabilities.

At the nanometer scale, individual spidroin molecules align parallel to each other, their crystalline regions forming tiny, rigid domains embedded in a softer amorphous matrix. Think of it like rebar in concrete, except the rebar and the concrete are made from the same molecule, just folded differently.

These aligned molecules bundle together into nanofibrils—protein cables just tens of nanometers across. The nanofibrils twist together into fibrils. The fibrils align into the final fiber.

At every level, the alignment is critical. If the molecules are jumbled randomly, the fiber loses most of its strength—the crystalline regions can’t share the load, and the whole structure falls apart under stress. The spider achieves a high degree of alignment by controlling how the liquid protein flows through its spinning duct, using shear forces and chemical triggers to coax the molecules into position before they solidify.

Human manufacturing hits its first major wall here. We can make the protein. We can even make it fold correctly. What we can’t do reliably, quickly and at scale is get millions of protein molecules to align perfectly as they transition from liquid to solid.

The spider does this in a duct that narrows to a few thousandths of a millimeter, in a fraction of a second, at room temperature.

It does not do it perfectly, since natural silk varies a lot from thread to thread, but it does it well enough to stake its life on. We have been trying to match it for thirty years.

Why Different Silks Do Different Jobs

A single spider produces up to seven different types of silk, each suited to a specific function. The orb-weaver sitting in your garden is running a small materials factory.

The structural frame of the web, the non-sticky radial threads and the outer support lines, is made from major ampullate silk, also called dragline silk. This is the one everyone studies, the “miracle fiber.” It’s strong, tough, and relatively stiff. The spider uses it as a safety line when it drops from a surface, trusting its life to a single strand.

The sticky capture spiral that catches insects is viscid silk, made from different glands. It’s about half as strong as dragline and thousands of times less stiff, as soft as a rubber band, but it can stretch to several times its length and is coated with sticky glycoprotein droplets. Its job isn’t to hold the insect but to trap it long enough for the spider to arrive.

The egg sac gets wrapped in cylindrical silk, which is tough but flexible, optimized for protecting the eggs without crushing them. When the spider wraps prey, it uses aciniform silk, which is produced in large quantities and bonds to itself easily.

Each silk has a different protein composition, different crystalline-to-amorphous ratios, different mechanical properties. Instead of one super-material, the spider makes a toolkit of specialized materials, each matched to its task.

The industry chose to focus on dragline silk for a simple reason: it has the best all-around properties, combining strength, toughness and stretch. It’s the closest natural analog to what you’d want for body armor, high-performance textiles, or aerospace components.

But this focus on dragline also reveals an industrial bias. We wanted the one material that could do everything, a universal replacement for Kevlar, nylon, and carbon fiber.

Nature makes specialized materials for specialized tasks, on demand and in tiny quantities.

What “Toughness” Means

The word “strong” gets thrown around carelessly, so it helps to be specific about what makes spider silk extraordinary.

In materials science, there are three critical but distinct properties:

Strength is how much force a material can withstand before breaking. Pull on a steel cable until it snaps; the force at the break, divided by the cable’s cross-section, is its tensile strength.

Stiffness is how much a material resists deformation. Press down on a wooden board versus a foam cushion—the wood is stiffer because it barely bends.

Toughness is how much energy a material can absorb before failing. This is the property that matters for body armor, crash protection, and catching flying insects. It’s measured by the area under a stress-strain curve—essentially, how much work you have to do to break something.

Of the three, toughness is where spider silk stands out.

Kevlar has higher tensile strength than spider silk in absolute terms: around 3.0-3.6 GPa, compared with spider silk’s 1.0-1.5 GPa. Steel is stiffer. But neither can match spider silk’s ability to absorb energy. Kevlar’s toughness is 30 to 50 megajoules per cubic meter. The toughest spider silk, from Darwin’s bark spider, can reach 350-520 MJ/m³—over ten times tougher than Kevlar.

When force hits Kevlar, the fabric stops it by distributing the impact across the weave—but Kevlar fibers themselves fail by rupture. The fibers break through a combination of tensile overload and fiber pullout. Once broken, the vest is compromised, and the wearer still absorbs significant blunt force trauma.

Spider silk would theoretically do something different at moderate impact speeds. Because it combines strength with high elongation (it can stretch by a quarter or more of its length before breaking), it absorbs impact energy by deforming rather than shattering. The crystalline regions provide strength, preventing total failure. The amorphous regions unfold, stretching and dissipating energy like molecular springs.

At the molecular level, this happens through a mechanism called sacrificial bonding. The hydrogen bonds holding the protein structure together are relatively weak individually—they’ll break under stress. But there are millions of them, and they don’t all break at once. Instead, they break sequentially, each one absorbing a tiny amount of energy. The protein chain unfolds in a controlled manner, like a carefully deployed airbag rather than a balloon popping.

So a thread can stop a bee without breaking. The silk stretches, absorbing the bee’s kinetic energy over a longer time and distance, converting that energy into molecular deformation rather than structural failure. The web flexes and the silk holds.

Most of that absorbed energy, around two-thirds, is turned into heat rather than stored, so the web does not catapult the insect back out. Lightly loaded threads largely recover; badly overstretched ones are replaced by the spider.

Kevlar works differently. It stretches only a few percent before breaking, so it can absorb far less energy per unit volume, and fibers that fail stay broken.

This combination of high strength, high elongation and the ability to shed energy as heat gives spider silk a unique place on the performance envelope. Few materials, natural or synthetic, come close to its toughness.

The problem, of course, is that toughness doesn’t sell if you can’t manufacture the material. And manufacturing it at the quality the spider achieves, with that crystalline-amorphous architecture, that alignment and that precise ratio of structure to flexibility, remains the unsolved challenge.

We know what makes it work: we can see it under microscopes, measure it with X-ray diffraction and model it with computational chemistry.

Thousands of papers explain why spider silk works, and we still can’t make it at industrial scale.

Evolution spent hundreds of millions of years perfecting the spider’s manufacturing capability, and we, with all our biotechnology and materials science, still cannot match it.

Knowing what makes spider silk special has turned out to be very different from knowing how to make it yourself, especially when you need to do it profitably, at scale, in a factory that answers to investors and customers rather than natural selection.

The first serious attempt to close that gap began with a gene and a herd of goats.

The First Wave: Bold Promises and Failed Shortcuts (1990s–2000s)

In 1990, a molecular biologist named Randy Lewis at the University of Wyoming published something that seemed, at the time, like the first page of a science-fiction script: the genetic sequence of a spider’s dragline silk protein. Within a decade, a Canadian company would use genes like his to try to convince goats to make spider silk.

Not spin spider silk—that would come later, maybe. First came the raw material: the liquid protein that spiders produce in their abdomens before transforming it into fiber. The logic was impeccable. Spiders cannibalize each other, making them impossible to farm. Goats, on the other hand, are docile, productive, and already optimized by thousands of years of agricultural breeding to produce large quantities of protein in their milk.

All you needed to do was insert the spider silk gene into the goat’s genome, target it to the mammary glands, and let nature’s existing dairy infrastructure do the work.

When the goats did produce milk containing spider silk protein, the news exploded. It was biotechnology delivering on its most audacious promise: rewriting the genetic code of one species to give it the abilities of another.

The media treatment was predictably breathless. “Nexia’s transgenic ‘spider goat’ to produce milk of steel,” announced Canada’s Globe and Mail in January 2000. Defense contractors called. Textile manufacturers sent inquiries.

Venture capitalists began the math: if one goat produces X liters of milk per day, and the milk contains Y percent silk protein, then a herd of Z goats could produce… On paper, the numbers looked incredible.

The Gene Was Supposed to Be the Hard Part

To understand the optimism of the early 1990s, you need to understand where biotechnology was at that moment. The Human Genome Project was underway. Genetic engineering was transitioning from theoretical possibility to practical tool. Researchers had successfully expressed human insulin in bacteria, creating a renewable source of a life-saving drug that previously required harvesting cow and pig pancreases.

The paradigm was simple and seductive: DNA is the instruction manual, and if you can read it, you can copy it, paste it into a new organism and press “run.”

Spider silk seemed like a perfect test case. The silk genes were well-characterized—long, repetitive sequences encoding the modular protein structures described in the previous chapter. Getting those genes into bacteria, yeast, or mammals was established technology. The organisms would become living factories, churning out spider silk protein using nothing but their normal metabolism.

This was the promise that launched a wave of research programs and startups.

Its premise was that the genetic engineering was the hard part and everything else was just industrial scale-up. That assumption turned out to be badly wrong.

The Menagerie of Silk Factories

The goats were just the beginning. Over the next fifteen years, researchers threw the entire biotechnology toolkit at spider silk production, engineering a bizarre menagerie of organisms.

The transgenic goats, developed by Nexia Biotechnologies and later continued by Randy Lewis at Utah State University, were the flagship effort. The advantages were obvious: large animals producing liters of protein-rich fluid daily, using existing dairy infrastructure for collection and processing. The spider silk protein would be dissolved in the milk; you’d extract it, purify it, and spin it into fiber.

The problems were equally obvious, though they took years to fully appreciate. First, milk is a complex biological soup containing hundreds of proteins, fats, and sugars. Separating out one specific protein required expensive chromatography and filtration. Second, goats are expensive to maintain. They require land, feed, veterinary care, and many months before a new transgenic doe can be bred and produce milk. Third, each goat produced a slightly different concentration of silk protein depending on genetics, diet, and lactation cycle. Consistency, the industrial holy grail, was nearly impossible.

And fourth, perhaps most damningly: scale required herds of hundreds of goats, eventually thousands, to produce commercially relevant quantities. The romanticism of spider-goats evaporated quickly when confronted with the logistics of industrial dairy farming.

The bacteria were more practical but came with their own curse. E. coli has been the workhorse of biotechnology since the 1970s—cheap, fast-growing, easy to manipulate genetically. Getting bacteria to produce spider silk protein was straightforward. Getting them to produce useful spider silk protein was not.

The problem was inclusion bodies. When bacteria try to produce large quantities of foreign protein, especially large, complex proteins like spidroins, they often get overwhelmed. The proteins misfold and aggregate into dense, insoluble clumps inside the cell. These inclusion bodies are a headache: the protein is in the wrong shape and has to be dissolved and refolded before anyone can spin it.

Researchers could break open the cells and extract the inclusion bodies using harsh chemicals and high heat, then try to refold the protein into its correct structure. Sometimes this worked. Often it didn’t.

When it did work, the process was so energy-intensive and expensive that it negated any cost advantage from using bacteria in the first place.

Yeast offered a middle path. Pichia pastoris and other industrial yeast strains have more sophisticated protein-folding machinery than bacteria—they’re eukaryotes, with cellular compartments and chaperone proteins that help fold complex proteins correctly. They can be grown in massive bioreactors using well-established fermentation technology, the same basic process used to make beer or industrial enzymes.

Later, a second generation of companies bet heavily on microbial fermentation. Bolt Threads in California engineered yeast; AMSilk in Germany used bacteria; Spiber in Japan says only that it uses “microorganisms.” All developed proprietary strains capable of producing spidroin-like proteins at yields measured in grams per liter. This was real progress. The protein came out soluble, properly folded, and in concentrations high enough to be economically interesting.

But “economically interesting” turned out to be a dangerously low bar. Growing yeast or bacteria requires sugar feedstock—lots of it. Industrial fermentation requires temperature control, sterile conditions, and constant agitation. All of this requires energy. After fermentation, you still need to separate the protein from the cells and the growth medium, then concentrate it to the high densities needed for spinning.

When companies ran the full cost accounting, the numbers were sobering. Estimates varied wildly. One silkworm-based competitor, Kraig Labs, once estimated that small fermentation batches cost $35,000–50,000 per kilogram of protein, a figure it has since stopped standing behind. For bacterial production, modeling puts the minimum selling price of finished fiber at about $760 per kilogram for a first commercial plant, falling to roughly $23 per kilogram only in an optimized, nth-of-a-kind plant, under assumptions that no company has yet demonstrated.

For context, a kilogram of nylon resin costs about $2. Kevlar, one of the most expensive mass-produced performance fibers, costs many times more than nylon.

The transgenic silkworms seemed like they might solve everything. Silkworms already produce silk—lots of it, reliably, for thousands of years. The sericulture industry existed, with established infrastructure for growing the worms, harvesting cocoons, and extracting fiber. If you could just get silkworms to produce spider silk instead of their native silk, you’d have an instant industry.

Researchers at the University of Notre Dame, the University of Wyoming, and institutions in China and Japan all pursued this approach. They successfully created transgenic silkworms that produced silk containing spider silk proteins, either pure or blended with the worm’s native silk.

It worked. The worms spun cocoons containing the engineered protein, and composite fibers tested in 2012 were, on average, tougher than ordinary silkworm silk and about as tough as natural dragline. But the spider protein made up only a fraction of the fiber, which remained mostly silkworm silk, and the results varied from line to line.

And there was a more fundamental problem: silkworms spin their cocoons in one continuous fiber over several days, using a completely different spinning process than spiders, and those early composite fibers did not reproduce the spider’s chemical and mechanical choreography. Only in 2023 did a Chinese team get silkworms to spin whole spider-silk fibers, with a strength of about 1.3 GPa and a toughness around 320 MJ/m³, on a par with natural dragline, and even that remains a laboratory result rather than a product.

Plants and algae were the long shots. Some researchers engineered tobacco, alfalfa, and even potato plants to produce spider silk proteins. Others tried algae, thinking that photosynthetic organisms might offer a sustainable, low-cost production platform.

These efforts produced papers and patents but no industry. Plants could make the protein (up to about 2% of the soluble protein in tobacco and potato leaves), and it was easy to purify thanks to its heat stability. But the plant-made proteins were smaller than the giant natural spidroins, and no plant or algal system has come close to commercial production.

What Worked, and What It Meant

By the mid-2000s, the first wave of spider silk companies could claim a real achievement: they had successfully produced spider silk protein in non-spider organisms at scales that could be measured in kilograms per year rather than milligrams per week.

This was not nothing. Fifteen years earlier, the only way to get spider silk protein was to dissect it from spiders. Now, you could grow it in a bioreactor.

But this achievement came with a hard realization: producing the protein was only the beginning. The real problem, which would consume another two decades and hundreds of millions more dollars, was what to do with the protein once you had it.

The protein existed as a concentrated solution, sometimes called “silk dope”—a viscous, water-based liquid containing roughly 30–50% protein by weight. In the spider, this dope sits in the major ampullate gland, waiting to be transformed into fiber by the spinning duct’s precise sequence of chemical and mechanical operations.

In the factory, the dope sat in tanks and containers, and researchers stared at it, trying to figure out how to turn it into fiber that worked.

Early attempts used conventional textile extrusion methods, forcing the protein solution through a small nozzle, sometimes into a coagulation bath of methanol or acetone, sometimes just into air.

These methods worked for nylon, polyester and even Kevlar, but fell well short with spider silk.

The resulting fibers were thicker than natural silk and much weaker. Even Nexia’s widely publicized 2002 fibers, which matched dragline in toughness and stiffness, could not match its strength. The finer and better-oriented the fibers were drawn, the stronger they became—a clue that the problem lay in how the proteins were spun and aligned, not in the proteins themselves.

Industrial extrusion was too fast and too harsh. The proteins didn’t have time to align before solidifying. The crucial beta-sheet crystals didn’t form properly. The fiber looked like spider silk under a microscope but fell far short of it under testing.

Several companies announced that they had produced “spider silk fiber.” It was fiber made from spider silk protein, but its mechanical properties fell far short of the real thing.

The Pivot, the Silence, and the Shutdown

By 2009, the first wave was ending. Nexia Biotechnologies, the Montreal company behind the spider goats, had sold its most promising goat-milk drug program to the US firm PharmAthene in 2005 and then quietly went under. Two of its goats ended up at the Canada Agriculture Museum. Randy Lewis’s team took others from Canada to Wyoming, and later to Utah State University, where his lab carried on the work as a publicly funded research program rather than a commercial venture.

Kraig Biocraft Laboratories, a later entrant focused on transgenic silkworms, spent years oscillating between military contracts and consumer textiles while its shares traded for pennies over the counter. Only in 2026, after scaling silkworm rearing in Vietnam, did it announce its first commercial yarn shipment—an order its CEO called “small by design.”

Other companies made softer exits. They stopped talking about bulletproof vests and started talking about wound dressings. They stopped promising disruption of the textile industry and started targeting niche medical applications where high costs could be justified by high margins and low volumes.

Some ran out of money and shut down without press releases or explanations. Their websites went dark, their patents expired or were sold, and the researchers moved on.

What’s striking in retrospect is how little drama accompanied these failures. There were no spectacular bankruptcies or exposés; the companies faded. Press releases became less frequent. Timelines quietly extended. Nexia, for example, had told reporters in January 2000 that large-scale production would begin before the year was out. It never did.

The infrastructure and the knowledge remained, and protein production technology continued to improve incrementally: yeast strains and purification methods got better and costs came down, just not fast enough or far enough.

But the original promise, the vision of spider silk as a revolutionary material that would displace Kevlar, reinvent body armor and launch a new bio-based materials industry, had faded, mourned mainly by the researchers and investors who’d bet their careers and capital on it.

The Lesson of the First Wave

The first wave failed because it operated on a fundamental misunderstanding of where the difficulty lay.

The genetic engineering was never the main bottleneck. Yes, it was technically challenging, but it was solvable with existing tools. Inserting genes into organisms, optimizing expression, scaling up fermentation—this was known territory, the subject of textbooks and commercial practice.

The bottleneck was always the transformation from liquid to solid: the spinning process that happens in a fraction of a second inside a spider’s abdomen and that we still, more than thirty years after the first spider silk gene was decoded, cannot replicate at industrial scale while maintaining the material’s exceptional properties.

The first wave assumed that the spider’s achievement was the protein—that evolution’s masterpiece was the molecular structure. Therefore, once you had the protein, the hard part was done.

The harder achievement was the spinneret, the biological machine that takes the protein and converts it into fiber with remarkable efficiency, using nothing but microfluidic flow control and carefully orchestrated chemistry.

That realization would shape the second wave of attempts. But first, the industry had to confront an even more basic question, one that should have been asked at the beginning: if making spider silk is so hard, why not just farm spiders?

The answer to that question explains why every approach, no matter how clever, eventually runs into the same wall.

Why You Can’t Farm Spiders

Anyone who talks about spider silk for long gets asked the same question.

If silkworms can be farmed to make regular silk, why can’t we just farm spiders?

If you could farm spiders, the industry would not need genetic engineering, biotechnology or hundred-million-dollar research programs; it would just build spider farms.

People have tried for centuries, and it has never worked. The reason it doesn’t work reveals something fundamental about the constraints that shaped every subsequent attempt to produce spider silk commercially.

The Experiment That Keeps Failing

In 1709, a French magistrate and naturalist named François Xavier Bon de Saint Hilaire attempted to launch the world’s first spider silk industry. He gathered spider egg sacs, boiled and carded their silk, and had it knitted into stockings and gloves, which he sent to the French Academy of Sciences and to the Royal Society in London.

The experiment was technically successful. The gloves existed. They were made from spider silk. The Academy took the idea seriously enough to have it studied.

The experiment was economically catastrophic. When the Academy asked the naturalist René-Antoine Ferchault de Réaumur to assess it, he found that caged spiders fought and ate one another, and that egg-sac silk was so sparse that a single pound would take, by his arithmetic, 55,000 of the best spiders or more than 660,000 of the common garden kind. The idea never became an industry.

Three hundred years later, two men tried again. In Madagascar, textile artist Simon Peers and designer Nicholas Godley spent about eight years collecting silk from more than a million golden orb-weaver spiders (Nephila).

They made two pieces: a golden brocade textile, shown at the American Museum of Natural History in 2009, and an embroidered cape displayed at London’s Victoria and Albert Museum in 2012. The textile was stunning, and the process behind both pieces was grueling.

Workers collected spiders from the wild each morning. Each spider was gently restrained while silk was drawn by hand from its spinnerets, a process called “silking,” and the animal was returned to the forest the same day.

The yield was tiny: about 23,000 spiders for 25 grams of silk—not even an ounce. The cape alone weighs about 1.9 kilograms.

The cape became a museum showpiece and a demonstration that spider farming is commercially impossible.

The Biology That Breaks the Model

The reason is written into spider biology at every level, starting with the most obvious: spiders are predators, and predators don’t farm well.

Silkworms are herbivores that eat mulberry leaves. You can pack thousands of silkworms onto trays stacked in warehouses, feed them cheap leaves, and they’ll peacefully coexist until they spin their cocoons. They’ve been domesticated for roughly 5,000 years. They’re now so specialized for silk production that Bombyx mori, the domestic silkworm, can barely survive in the wild. It’s the dairy cow of invertebrates: docile, productive and thoroughly optimized for human use.

Spiders are nothing like that. Most species of interest for silk production, orb-weavers like Nephila and Argiope, are solitary hunters.

They’re territorial, and their entire evolutionary strategy is built around defending a web-shaped piece of real estate and eating anything that comes near it.

Put two spiders in proximity and they don’t cooperate. They fight. The larger one usually eats the smaller one.

Steady feeding damps the aggression, and a few laboratories keep well-fed orb-weavers in shared rooms, but nobody has turned that into a farm; the behavior is fundamental, evolved over millions of years. Female spiders sometimes eat males even during mating, and sexual cannibalism is common enough in some species to be the default outcome. Keeping hundreds of spiders at the density a farm would need has never been done.

You could, theoretically, house each spider individually. But now you’re not farming—you’re running a zoo. The labor and infrastructure costs scale almost linearly with the number of spiders, and there is little economy of scale.

And unlike silkworms, which each spin one large cocoon that can be harvested in bulk, spiders produce silk continuously in small amounts. They spin webs, which you could collect, but web silk is sticky, and the dragline silk you want is only one of several silk types in it.

The only practical method is manual extraction—the “silking” process used in Madagascar, where humans physically restrain each spider and pull silk from its spinnerets. It’s slow, labor-intensive, and stressful for the spider, which reduces future silk production.

The Math That Doesn’t Work

Let’s run the numbers on what industrial spider farming would require.

A large Nephila spider can be “silked” by hand for perhaps tens of meters of dragline at a session. That sounds promising until you calculate the mass: dragline silk is roughly 5 microns in diameter, so one hundred meters of it weighs only about 2.5 milligrams.

Industrial textile fibers are sold by the ton. A single ton is one million grams. To produce one ton of spider silk per year via farming, you would need more than a million spiders each yielding 100 meters every single day, assuming perfect collection efficiency and no losses.

In practice, accounting for mortality, stress, seasonal variation, and the impossibility of harvesting every single day, you’d likely need several million spiders in active production at any given time.

Now add the infrastructure: individual enclosures (spiders can’t share), feeding (each spider needs live insects), waste management, climate control, and the labor cost of manually harvesting silk from millions of individual spiders daily.

Compare this to sericulture. Modern silkworm farms produce multiple tons of silk from a single warehouse using seasonal labor and bulk harvesting. The silkworms don’t need individual housing, don’t cannibalize each other, and produce their silk automatically in convenient, harvestable cocoons.

Or compare it to synthetic fiber production. A single nylon production facility produces thousands of tons per year using a fully automated process, with no feeding, no waste management and no individual animal care.

The Decision That Shaped Everything

This biological dead end is why the entire spider silk industry took the path it did. Since you can’t farm spiders, you need an alternative source of silk protein. That means biotechnology: engineering other organisms to produce the protein for you.

But accepting this necessity meant accepting a second, harder problem: if you’re not using spiders, you’re not using their spinnerets either. You don’t just need to produce the protein—you need to invent an entirely new process for converting that protein into fiber.

The spider’s silk production is an integrated biological system. The protein composition, the chemical environment of the gland, the mechanical shear forces in the spinning duct, the precise timing of pH changes and ion exchanges—all of these evolved together as a matched set. You can’t extract one part and expect it to work independently.

When researchers chose to abandon spider farming in favor of genetic engineering, they implicitly chose to solve two problems instead of one:

1. Produce the protein in a non-spider organism

2. Build an artificial spinneret that can replicate the spider’s process

The first wave of companies thought problem #1 was the hard one. They were wrong. Problem #1 turned out to be solvable with existing biotechnology, albeit at costs higher than hoped.

Problem #2, the spinning, turned out to be persistently difficult, so difficult that it remains unsolved at industrial scale three decades later.

Why the Farming Dead End Matters

The impossibility of spider farming is the original constraint that forced every subsequent decision in the field. It is why spider silk became a biotechnology story, and why hundreds of millions of dollars went into fermentation tanks rather than arachnid husbandry.

And it’s why the comparison to silkworms, which makes spider silk farming sound so plausible, is misleading. Silkworms are a different category of organism entirely: domesticated, docile, optimized over millennia for human use.

Spiders are wild predators, shaped by an evolution that never anticipated human agriculture.

That constraint shaped everything. The genetic engineering pathway was chosen because it was the only option. And once that choice was made, the industry found itself trying to replicate not only a material but the entire biological manufacturing process that evolution had spent hundreds of millions of years perfecting.

We couldn’t farm the animal, so we farmed the protein instead, only to discover that having the protein was half the problem, maybe less.

The spider sits in its web producing a material we want but can’t economically harvest. That impossibility launched an industry, and it explains much of why that industry has spent thirty years failing to deliver on its promise.

Unable to farm spiders, the industry set out to rebuild the spider’s spinning machinery itself, which turned out to be even harder.

The Core Technical Bottleneck: Spinning, Not Protein

The usual end point of months of painstaking, expensive work in a spider silk lab is a vial of concentrated silk protein solution, and no reliable way to turn it into fiber.

The fermentation worked, the purification succeeded and the molecular structure is correct: beta-sheets, amorphous regions, everything in the sequence that nature designed.

You’ve got perhaps 100 milliliters of solution containing 30–40 grams of silk protein, more dragline protein than a single spider could make in its lifetime, and it might as well be expensive soup.

The next step is converting that liquid into a fiber with the properties that make spider silk special. After three decades of research and hundreds of millions in funding, it remains the unsolved problem that has killed nearly every commercial spider silk venture.

Understanding why spinning is so hard explains why the industry has been stuck for thirty years despite continuous progress in every other dimension.

Why the Protein Was Not the Hard Part

By 2010, multiple research groups and companies could produce spider silk protein at scales measured in kilograms. Bolt Threads had proprietary yeast strains. Spiber in Japan had their own fermentation technology. Academic labs in Wyoming, Germany, Sweden and elsewhere had demonstrated gram-scale production.

The protein problem wasn’t solved in the sense of being cheap—modeled selling prices for finished fiber ran from hundreds of dollars per kilogram for a first plant down to a few tens of dollars only under optimistic, fully optimized assumptions. But it was solved in the sense that the technology existed, was reproducible, and was steadily improving. Every year brought higher yields, better folding, more efficient purification.

If protein production were the only challenge, spider silk would by now be expensive but available, like certain specialty polymers or pharmaceutical ingredients.

But having the protein only gets you to the starting line. The hard part begins when you try to make fiber.

That liquid silk protein is a highly concentrated aqueous solution of massive, repetitive proteins suspended in a delicate chemical balance. The proteins are folded but not yet assembled into the final fiber structure. They’re soluble, which means they’re surrounded by water molecules and maintaining enough separation that they don’t aggregate and crash out of solution.

In the spider’s major ampullate gland, this “silk dope” sits at concentrations of 30-50% protein—about as thick as you can get while maintaining fluidity. It’s stored in a carefully controlled chemical environment, with set pH, ion concentrations and temperature. Change any of these and the protein starts to aggregate prematurely. Get it wrong and your expensive solution turns into expensive cottage cheese.

The spider keeps the dope stable until she’s ready to spin. Then, in a fraction of a second, she transforms that liquid into a solid fiber with highly aligned molecules and exceptional mechanical properties.

We’ve been trying to figure out how to do that since the 1990s, and we’re still trying.

The Molecular Choreography We Can’t Replicate

The spider’s spinning process is a feat of chemical and mechanical engineering compressed into a long, S-shaped duct that tapers to just a few thousandths of a millimeter. Inside it, the dope goes through five stages.

Stage one: concentration. The silk dope enters the spinning duct at high concentration but still with enough water to keep it liquid. As it flows through the initial section of the duct, water is actively reabsorbed through the duct walls. The protein concentration increases further, forcing the proteins closer together.

Stage two: acidification. The pH drops sharply, from about 7.6 in the gland to 5.7 or lower in the duct. This isn’t random. The silk proteins have specific amino acids that respond to pH changes. At higher pH, they repel each other electrostatically. As the pH drops, that repulsion weakens. The proteins begin to associate.

This pH transition is precise. Too fast or too slow and the assembly goes wrong. The spider controls it with specialized cells lining the duct that actively pump protons, creating a smooth pH gradient.

Stage three: ion exchange. Simultaneously with acidification, the ionic environment changes. Sodium and chloride ions, which stabilize the liquid state, are removed. Potassium and phosphate ions are introduced. These ion swaps further destabilize the dissolved state and promote protein aggregation.

The changes follow a spatial pattern along the duct, a sequence of chemical shifts that guides the protein assembly.

Stage four: mechanical shear. At this stage physics takes over from chemistry. The spinning duct is tapered, getting narrower along its length. As the thickening protein solution is pulled through this narrowing channel, it experiences increasing shear forces.

Shear is what happens when fluid flows past a surface or through a constraint. Imagine honey flowing off a spoon—the honey right at the spoon’s surface moves slower than the honey further away, creating layers sliding past each other. That’s shear.

In the spider’s duct, the shear forces act on the silk proteins, physically stretching and aligning them in the direction of flow. This is critical. The crystalline beta-sheet regions need to form parallel to the fiber axis. The amorphous regions need to be properly distributed between them. Random alignment gives you weak fiber. The shear forces from the tapered duct create directional alignment.

But the shear has to be strong enough to align the proteins but gentle enough not to disrupt their folding. Too little shear and you get poor alignment. Too much and you denature the proteins, destroying their structure.

The spider achieves this through laminar flow—smooth, layered flow without turbulence. The proteins slide past each other in orderly sheets, gradually aligning, gradually assembling into the final fiber structure as the chemical triggers (pH, ions) tell them when to lock into place.

Stage five: solidification. By the time the dope reaches the end of the spinning duct, it’s no longer a liquid. The proteins have assembled into aligned bundles. Most of the water has been drawn out. The fiber emerges solid but still somewhat elastic, completing its final hardening as it’s pulled away from the spinneret.

The entire process, from liquid entering the duct to solid fiber emerging, happens within a fraction of a second.

Why Industrial Extrusion Destroys Everything

Now consider what happens when you try to replicate this process using industrial fiber production equipment.

Conventional fiber spinning comes in two main varieties: melt spinning, used for nylon and polyester, and wet spinning, used for rayon and some aramids. Both involve forcing a polymer through a small hole, called a spinneret, to form a continuous fiber.

Melt spinning uses heat. You melt the polymer and extrude it through tiny holes. As it emerges and cools, it solidifies. This works great for simple synthetic polymers that are thermally stable.

It’s useless for spider silk protein. Silk proteins do not melt; they degrade before they flow, so they cannot be melt-spun like nylon. Melt spinning typically operates at 200-300°C. You’d end up with protein-flavored char.

Wet spinning avoids heat by using chemical solvents. You dissolve the polymer in a solvent, extrude it into a coagulation bath (usually a different chemical that causes the polymer to precipitate), and pull the resulting fiber out.

This is closer to what might work for spider silk. Several research groups have tried variations: extruding the silk dope into methanol, or acetone, or various salt solutions that cause the protein to aggregate and solidify.

And it works, sort of: you get fiber made of spider silk protein, and under a microscope it looks like a fiber.

But the mechanical properties disappoint. The tensile strength is often only a fraction of natural spider silk’s, and the toughness, the critical property that makes spider silk special, usually falls short too.

The trouble comes from three directions: alignment, kinetics and shear.

Problem one: alignment. Industrial extrusion is fast. You need high throughput to be economically viable—meters of fiber per second, not millimeters. At these speeds, the dope can be stretched and sheared faster than the proteins organize, and a harsh coagulation bath can freeze them in place before they have lined up.

Often the silk proteins don’t get the chance to align, and when they solidify, they’re jumbled.

The crystalline regions point in random directions, and the load-bearing structure that depends on parallel alignment doesn’t form properly. The result is weak fiber that fails at a fraction of the stress natural silk can handle.

Problem two: kinetics. The spider’s transformation, fast as it is, is carefully sequenced. The pH changes gradually along the duct. The ions exchange in a specific order. The proteins have time to fold, associate, and align before they’re locked into the final structure.

In a typical coagulation bath, the protein solution can crash out of solution almost instantly. The proteins aggregate wherever they happen to be, however they happen to be oriented. There’s no time for careful assembly.

The result is fast precipitation instead of controlled self-assembly, the difference between carefully stacking bricks to build a wall and dumping a truckload of bricks in a pile.

Problem three: shear. At industrial flow rates, the shear forces in the spinneret are enormous—orders of magnitude higher than what the spider applies. These forces can break chemical bonds, disrupt protein folding, and stretch the molecules faster than they can settle into an orderly structure.

But you can’t just slow down. Slow flow means low throughput means uneconomical production. The spider can get away with it because she only needs a few meters of silk at a time. A factory needs kilometers per hour to compete with nylon production.

The physics doesn’t scale. The gentle, controlled flow that works in a duct a few micrometers wide cannot simply be replicated in a larger system operating at higher speeds. The flow conditions fundamentally change, and the precise control the spider relies on becomes much harder to achieve.

The Economic Trap of Slow Spinning

Some research groups have achieved impressive results by mimicking the spider more closely: slow extrusion through microfluidic channels, careful pH gradients, controlled ion exchange, gentle pulling forces.

In 2017, a team at the Swedish University of Agricultural Sciences and Karolinska Institutet built a spinning device that mimicked the spider’s pH drop and shear using only water-based buffers. From one liter of bacterial culture they spun a kilometer of fiber—the toughest as-spun artificial spider silk reported at the time.

It was a real breakthrough, but the fiber was still far weaker than natural dragline, and the process ran at laboratory flow rates, not factory ones.

Industrial melt spinning of polyester and nylon runs at thousands of meters per minute. Even aramids like Kevlar, spun from acid solutions, can be wound up at hundreds of meters per minute. Specialized high-performance fiber lines assume speeds measured in meters per minute, not meters per hour.

So far, the processes that come closest to the spider’s, and to its fiber, have run at laboratory rates, and slow production is expensive production.

You can have spider-quality fiber at spider speeds and spider scale, producing grams per day at costs in the thousands of dollars per kilogram. Or you can have industrial throughput producing tons per day—but the fiber loses the properties that made spider silk worth pursuing in the first place.

Nobody has yet shown dragline-grade fiber coming off a line at factory speed. The companies that announced “spider silk fiber” production typically chose the industrial-speed option, accepting dramatically reduced mechanical properties in exchange for achievable production rates. Their fiber was “spider silk” in molecular composition but not in performance.

Why This Problem Has Swallowed Hundreds of Millions

Understanding the spinning bottleneck explains why the spider silk industry has evolved the way it has—and why it’s failed to deliver on its promises.

Protein production was worked out, more or less, by the mid-2010s: fermentation works, yields have risen and costs have come down, though not yet to the level of a commodity. If protein were enough, we’d have a spider silk industry.

Instead, protein is expensive raw material sitting in a tank, waiting for a manufacturing process that doesn’t exist at industrial scale.

The spinning process requires simultaneously controlling chemistry (pH, ions), fluid dynamics (laminar flow, specific shear forces), and kinetics (timing of assembly), all in a continuous process running fast enough to be economical. Nature does this in a duct a few micrometers wide at its tip, refined over hundreds of millions of years of evolution. We’re trying to do it in industrial equipment optimized for different polymers with different assembly mechanisms.

Scaling up tends to break something: a larger duct makes the chemical gradients harder to control, a faster process leaves less time for alignment, and stronger coagulation, used to speed solidification, can disrupt the protein structure.

The spider’s process fits spider scale, spider speeds and spider control mechanisms, and both the physics and the economics resist moving it into a factory.

After thirty years, then, we know what spider silk is and we can make the protein. What remains beyond our ability to replicate economically at scale is the transformation from liquid to solid, the split second of molecular choreography that happens in a spider’s abdomen, and that is why you still can’t buy a spider silk bulletproof vest.

The Biomaterial Scale-Up Problem

By the early 2000s, Nexia Biotechnologies had a problem that, on paper, sounded like success. Its goats produced spider silk protein in their milk. In 2002 its scientists reported fibers wet-spun from recombinant spider silk protein made in mammalian cell cultures.

The problem was that nobody knew how to turn that protein into anything profitable.

The company had proof of concept, publications, patents and a deal with Canada’s defense department. What it didn’t have was a path from “we can do this in the lab” to “we can sell this for more than it costs to make.”

In 2005 it sold its lead drug program, and by 2009 it was gone.

Nexia had entered what investors call the valley of death.

The valley lies between demonstrated technology and viable manufacturing. Companies enter it once the science works in the lab and leave it only when production is scaled and customers are buying. Spider silk has been stuck in it for thirty years.

The Illusion of Progress

There’s a peculiar dynamic in materials science research that makes failure look like forward momentum. Every year, someone publishes a paper showing improved protein yield, or better fiber properties, or a novel spinning approach. Every few years, a startup announces it has achieved “breakthrough” production capacity.

The numbers sound impressive—claims along the lines of a “10x improvement in fermentation efficiency,” “fiber strength reaching 800 MPa,” or “production capacity of 50 kilograms per year.”

To someone outside the field, such as an investor, a journalist or a defense contractor, these sound like major milestones. They sound like an industry getting closer to commercial viability.

To someone who understands industrial manufacturing, they are rounding errors.

The gap between laboratory success and industrial viability is a series of compound problems that multiply each other, creating a barrier that gets steeper as you approach it.

What “Industrial Scale” Means

When a startup announces production of 50 kilograms per year, the press release will often include projections: “This capacity could scale to 500 kilograms, then 5 tons, enabling commercial applications in high-performance textiles.”

That projection misses the scale at which industrial materials are consumed: thousands of tons, not kilograms.

Global nylon (polyamide) fiber production: about 7 million tons in 2024. Para-aramid fiber (the family that includes Kevlar): on the order of 100,000 tons of annual capacity, by industry estimates. Even specialty aramid fibers occupy market niches measured in thousands of tons annually.

To be relevant in the performance fiber market (not dominant, just relevant), you need to be able to produce at minimum hundreds of tons per year. Otherwise, you can’t supply contracts, guarantee consistency or achieve the economies of scale that make your price competitive.

Fifty kilograms per year sounds like a lot to a researcher who previously produced 50 grams. But it is about 140 grams per day; a day’s output would fit in your coat pocket, and getting from there to 100 tons per year is a 2,000-fold scale-up, with new problems at every step.

The Contamination Catastrophe

One of the hardest aspects of biological manufacturing is the contamination risk. It’s a problem that pharmaceutical companies have spent decades learning to manage, at enormous expense. Biomaterials companies are learning the same lessons, with far less funding and far less margin for error.

Suppose you’re running a 10,000-liter bioreactor growing yeast that produces spider silk protein. The fermentation takes 3-5 days. At the end, if everything goes perfectly, you’ve got 10,000 liters of fermentation broth containing roughly 30 kilograms of protein.

Those 30 kilograms are still just protein in a broth, not yet purified, not yet spun, and the batch has already consumed days of reactor time and thousands of dollars of sugar, nutrients, energy and labor.

Now imagine a contamination event. Bacteria get into the reactor. It might come from the air handling system, an improperly sterilized valve or the water supply. The contamination doesn’t just slow the yeast’s growth—it actively consumes the nutrients meant for your engineered strain. It produces waste products that can denature your protein. It turns your expensive batch into unsalvageable waste.

In a small laboratory setup, with 1-liter flasks, careful sterile technique and researchers monitoring constantly, contamination is rare. In a 10,000-liter industrial bioreactor running continuously for days, with multiple feed lines, sampling ports, and temperature control systems, contamination is a persistent threat.

Pharmaceutical manufacturing deals with this through extreme measures: cleanrooms, redundant sterilization, single-use bioreactor components, extensive quality testing at every stage. These measures work. They also cost millions of dollars to implement and maintain.

Biomaterials companies trying to compete with nylon can’t afford pharmaceutical-grade contamination control. But they also can’t afford to lose batches. A contamination rate of even 5%, or one failed batch in twenty, can destroy your economics entirely when your margins are already thin.

The Purification Cost Nobody Talks About

After fermentation, you have a complex biological soup: yeast cells, spent growth medium, metabolic byproducts, and somewhere in that mess, your spider silk protein. Now you need to extract it.

This process, called downstream processing, is often the most expensive part of biological manufacturing. For many fermentation products it is one of the largest slices of production cost, though the share varies widely from product to product.

The protein needs to be separated from the cell mass. This requires breaking open the cells (if the protein is intracellular) or separating it from the cells (if it’s secreted into the medium). Then you need to remove all the other proteins, nucleic acids, lipids, and cellular debris.

This typically involves multiple steps: centrifugation to remove cells, filtration to remove large contaminants, chromatography to separate your protein from everything else, and finally concentration to get the protein to the high density needed for spinning.

Each step costs money. Centrifuges consume energy. Filters clog and need replacement. Chromatography resins are expensive and have limited reuse cycles. Concentration requires either expensive ultrafiltration membranes or energy-intensive evaporation.

Worse, these costs don’t scale down proportionally. Running a small purification process costs almost as much per kilogram as running a large one, because you need the same equipment, the same quality control, the same skilled labor.

This creates a vicious cycle. You can’t afford industrial-scale equipment until you’re producing at industrial volumes. But you can’t reach industrial volumes profitably until you have industrial-scale equipment that brings your per-kilogram purification costs down.

Multiple spider silk companies have discovered, after years of development, that the cost of purification alone, before spinning or any value-add, made their product uncompetitive with existing materials. They’d optimized fermentation, achieved high yields, and still couldn’t make the economics work.

Consistency: The Invisible Killer

In the lab, variability is expected. Batch A produces 27 grams of protein per liter. Batch B produces 31 grams per liter. You note the difference in your lab notebook, investigate what changed, and move on.

In industrial production, the same variability is a catastrophe. Industrial customers, such as textile manufacturers, defense contractors and medical device companies, require materials with specific, guaranteed properties.

When they place an order for 1,000 kilograms of fiber with a tensile strength of 1.0 GPa and an elongation at break of 15%, they need every kilogram to meet that specification.

That means every kilogram in every batch, not just the average.

This is extraordinarily difficult with biological manufacturing. Fermentation performance varies with subtle changes in temperature, mixing rate, feed timing, and even the age of the cell culture. Protein quality varies with fermentation conditions—the same genetic strain can produce protein with slightly different folding, different post-translational modifications, different purity.

These variations cascade. Slightly different protein going into the spinning process produces fiber with slightly different mechanical properties. A batch that falls outside the tolerance range on any of those numbers, stiffer, stretchier or weaker, is a problem; the customer has designed a product around the spec.

Achieving batch-to-batch consistency requires obsessive process control: every parameter held within tight windows, inputs such as feedstock, water and air kept consistent in quality, and equipment that performs identically every time.

Pharmaceutical companies achieve this through what’s called process validation: extensive documentation, statistical process control, and exhaustive testing. They can afford this because pharmaceuticals have enormous margins. A protein therapeutic might sell for $10,000 per kilogram or more.

Spider silk protein, to be competitive as a material, needs to sell for under $100 per kilogram—ideally under $50. There’s no margin for extensive quality control overhead. But there’s also no market without it.

Picture the tension this creates: fiber with excellent average properties but too much variation between batches, one lot at 90% of natural dragline’s strength and the next at 60%.

To an industrial customer, this inconsistency makes the material unusable. You can’t design a product around a material whose properties you can’t guarantee. It doesn’t matter if the average is good if the range is too wide.

The Capex Trap

The harshest economic reality of materials scale-up is that the capital expenditure hits before the revenue arrives.

To produce spider silk fiber at commercially relevant volumes, say 100 tons per year, you need:

– Industrial-scale fermentation capacity: multiple 50,000+ liter bioreactors

– Downstream processing equipment: industrial centrifuges, filtration systems, chromatography columns

– Fiber spinning equipment: custom-designed systems (because commercial spinning equipment doesn’t work for spider silk)

– Quality control laboratories: analytical equipment, testing rigs, trained personnel

– Facility infrastructure: cleanrooms, utilities, waste handling, storage

Nobody knows for sure what a facility capable of producing 100 tons per year of spider silk fiber would cost, because no company has publicly demonstrated one. But the warning signs are clear: Spiber, the best-funded company in the field, recorded a ¥28 billion (roughly $180 million) special loss in 2024 after construction of a large US plant fell badly behind plan, and had to restructure its debts in 2026.

This money needs to be raised and spent before you produce your first commercial ton, before you have customers or any revenue, and before you know for certain that your process will work at full scale.

This is what venture capitalists call a “capital-intensive” business model, and they hate it. The ideal venture-backed business is asset-light: software, services, things that scale with minimal additional capital. Materials manufacturing ties up heavy capital and is slow to reach profitability.

The returns are also lower. Even if everything goes right, a materials company looks more like a specialty chemicals business than a software company. US specialty chemicals firms run gross margins around 35% and operating margins around 12%; in software, gross margins above 70% are typical. For the same amount of invested capital and risk, VCs would much rather fund software.

This explains why so many spider silk companies have run out of money just as they were approaching scale. They raised money to develop the technology and more to build a pilot plant. Now they need $100 million to build commercial production, but investors are exhausted, the timeline keeps stretching, and nobody wants to write the next check.

The Deadly Dance of Scale

The cruelest aspect of the biomaterials scale-up problem is that you can’t validate your process until you build at scale, but you can’t justify building at scale until you’ve validated your process.

Small-scale production, even at 10,000 liters, doesn’t predict how the process will perform at 100,000 liters. Mixing dynamics change. Heat transfer becomes harder. Contamination risks increase. Equipment behaviors shift.

Pharmaceutical companies handle this through a methodical scale-up process: extensive pilot studies, careful characterization at each scale, conservative projections. They can afford this because they’re working toward a product that might sell for $100,000 per kilogram.

Materials companies are working toward a product that needs to sell for $50 per kilogram. They can’t afford years of careful pilot studies. They’re pressured by investors to move fast, to get to commercial scale quickly, to start generating revenue before the money runs out.

So they make larger jumps. They scale from 1,000 liters to 50,000 liters based on limited data, and sometimes the plant behaves differently than expected, with more contamination, lower protein yield or poorer purification.

Now you’ve built a facility that doesn’t perform as projected, your cost per kilogram is well above what your model predicted, and you can’t raise more money because you’ve already failed at scale.

A company that was “just a few years from commercial production” can find itself months from bankruptcy.

Why “Kilograms Per Year” Is a Trap

When spider silk companies announce production milestones, such as “We’ve achieved 100 kilograms of production capacity,” they’re often technically correct but economically meaningless.

A hundred kilograms a year is enough to supply research labs, to make prototype materials, to demonstrate proof of concept. It’s nowhere near enough to supply a single industrial customer with a single product line.

As a rough illustration, an automotive manufacturer using high-performance fiber in a composite component might need tens of tons per year for just that one application. A body-armor program could need hundreds of tons per year. A textile manufacturer could need thousands.

The gap between “we can produce this” and “we can produce enough of this to matter” is where most biomaterials companies get stuck. They’ve solved the scientific problem, demonstrated the technology, and now they’re trapped in a scale-up phase that requires capital they can’t raise, expertise they don’t have, and time their investors won’t give them.

They celebrate kilogram milestones because that’s real progress from where they started. But industrial buyers measure supply in tons, and they also care about consistency and price.

And that’s why, after thirty years of progress, after thousands of research papers and hundreds of millions in investment, you still can’t buy industrial quantities of spider silk fiber at prices that make commercial sense.

The valley of death has claimed nearly everyone who’s tried to cross it. The few survivors who’ve made it partway across are still years from the other side and still burning cash.

The Biomimicry Trap: Why “Copying Nature” Keeps Failing

In 1941, a Swiss engineer named George de Mestral returned from a walk in the Jura mountains with burrs stuck to his trousers and his dog’s fur. Instead of cursing and picking them off, he examined them under a microscope. The tiny hooks on the burr’s surface had caught in the loops of the fabric. It took him about a decade of tinkering, but by the mid-1950s he had patented Velcro.

This is biomimicry’s origin story, repeated in business school case studies and innovation keynotes: look at nature, copy the mechanism, profit. It’s a seductive framework. Nature has had billions of years to optimize solutions, and we only need to work out how they function and copy them.

Spider silk became the poster child for this approach. Evolution had spent hundreds of millions of years perfecting a super-material. All we needed to do was copy it.

Thirty years later, we’re still trying. And the consistent failure reveals something uncomfortable about biomimicry as an innovation strategy: sometimes copying nature is a trap that leads you systematically in the wrong direction.

What Evolution Optimizes For

The misunderstanding at the root of the spider silk industry’s troubles is this: evolution doesn’t optimize for efficiency, cost, or scalability. It optimizes for reproductive success within a specific ecological context.

The spider’s silk production system is optimized for a solitary predator that needs to produce a few meters of fiber per day to catch insects and avoid being eaten. That is what evolution was optimizing for.

The system needs to work reliably enough: not perfectly, just well enough to keep the spider alive long enough to reproduce. It needs to use resources available to the spider, the proteins from digested prey and the metabolic energy from those same meals. It does not need to be fast, or cheap (in economic terms), or consistent in ways that matter to industrial manufacturing.

The spider recycles its web every day, eating the old silk to recover the protein. If a strand breaks, the spider just makes another one. If silk production is slower on a cold morning, that’s fine—the spider will catch fewer insects that day, but it’s not going to starve. The biological system has built-in flexibility, redundancy, and error tolerance.

Industrial manufacturing can’t tolerate any of this. A factory that produces 20% less fiber on cold days is a failed factory. A process that requires recycling and reprocessing mistakes is an uneconomical process. A system that works “reliably enough” instead of “perfectly every time” gets shut down.

Evolution optimized the spider for survival in nature; a factory has to optimize for profit in a market, which is a different problem.

The Scaling Laws That Nature Ignores

There’s a deeper issue: natural systems don’t scale linearly, and often don’t scale at all.

The spider’s spinning duct tapers to only a few micrometers across. The silk dope moves through it in tiny volumes, and the finished thread comes out at a few centimeters per second as the spider works, and at more than a meter per second when the line is pulled out hard. These dimensions create specific fluid dynamics—laminar flow, controlled shear forces, predictable diffusion of ions and pH gradients.

Now imagine scaling this up. You want to process vastly more silk, so you build a duct ten times longer and ten times wider—which makes it about a thousand times larger in volume.

The physics doesn’t scale along with it, because the relationship between a system’s surface area and its volume changes with scale.

If you double the linear dimensions of a tube, you quadruple its surface area but increase its volume eightfold. This affects heat transfer, diffusion rates, and mixing dynamics in ways that are mathematically unavoidable.

More critically, the flow regime changes. The spider’s tiny duct operates in a range where viscous forces dominate—the flow is smooth and predictable. Scale it up and increase the flow rate to maintain economic throughput, and the balance of shear, stretching, diffusion and chemistry that the spider relies on is hard to keep: ions and protons take longer to diffuse across a wider channel, and the proteins may be deformed faster than they can assemble.

There is no simple way to engineer around this. The fluid dynamics equations are non-linear, and fluids behave differently at different scales.

You can’t just build a bigger spinneret. The bigger spinneret operates in a different physical regime where the spider’s solution doesn’t work.

The Integration Problem

The spider’s silk production system is integrated into the spider’s entire physiology.

The silk gland is supplied with nutrients from the spider’s digestive system, which has already broken down and processed the raw materials. The pH gradients in the spinning duct are maintained by cells that are powered by the spider’s metabolism and controlled by its nervous system. The mechanical pulling force comes from the spider’s legs, with proprioceptive feedback telling the spider exactly how fast to pull and how much tension to apply.

The spider’s body temperature provides temperature control, its hemolymph (blood) supplies ions, and its excretory system removes waste. If the silk isn’t working properly, the spider compensates behaviorally: she pulls harder, adjusts her web architecture, or rebuilds entirely.

The entire system works because it’s embedded in a living organism that provides context, control, and correction automatically.

Now try to extract just the spinning duct and replicate it in a factory. You need to provide all those supporting systems artificially. You need pumps to circulate ions, control systems to manage pH, temperature regulation, force sensors and feedback loops, and analytical equipment to detect when something goes wrong.

In effect you’re trying to copy the entire spider, minus the parts you don’t want, and it turns out you can’t cleanly separate them.

The elegant solution you’re trying to copy only works inside the organism that runs it.

The Cost Structure Evolution Doesn’t Care About

Consider what it “costs” a spider to produce silk.

In accounting terms, the question has no answer. The spider doesn’t buy feedstock. It catches prey, digests it, and uses the resulting amino acids. There’s no invoice, no price per kilogram, no cost of goods sold.

The spider’s metabolism provides the energy, powered by the same prey, so there’s no electricity bill. The capital equipment, its silk glands, grew naturally as part of the spider’s development, so there’s no depreciation schedule.

The spider’s “manufacturing facility” is self-replicating, self-maintaining and paid for in prey rather than cash, and so are its raw materials and energy. The quality control is built-in neural feedback, and the labor is the spider itself.

Now consider what it costs a factory to produce silk:

– Feedstock: sugar at well under $1 per kilogram, but several kilograms of it for every kilogram of protein

– Energy: electricity for bioreactors, pumps, temperature control, purification

– Capital: bioreactors, spinning equipment, quality control labs—depreciated over time

– Labor: skilled operators, engineers, quality control technicians

– Overhead: facility maintenance, regulatory compliance, insurance

– Waste disposal: spent fermentation broth, failed batches, purification solvents

Every cost the spider pays in metabolism, the factory pays in cash, and often a lot of it.

Evolution optimized a system where all these costs are folded into the spider’s normal metabolism and biological functions. We’re trying to replicate the output while paying for every input explicitly.

This is why copying nature, taken literally, has not worked here. We were trying to copy the end result of a manufacturing process while using completely different economics and constraints.

It’s like watching someone cook a meal in their home kitchen and thinking, “I’ll copy that and start a restaurant.” The home cook doesn’t worry about food cost percentages, labor efficiency, or health department regulations. The restaurant has to worry about all of it. The same recipe produces completely different economics in different contexts.

When Biomimicry Works

To be fair, biomimicry isn’t always a trap. Velcro worked. Sharkskin-inspired surfaces that reduce drag have been successfully commercialized. Gecko-inspired adhesives are real products.

These successes have one thing in common: they copied a principle rather than a process.

Velcro doesn’t try to grow burrs. It uses nylon hooks and loops woven on industrial looms. The mechanism of hooks catching in loops is biomimetic, but the implementation is industrial.

Sharkskin-inspired surfaces don’t try to replicate shark skin’s biological growth process. They use microfabrication techniques to create similar surface patterns on different materials. The pattern is biomimetic; the production is conventional manufacturing.

The failures, with spider silk as the premier example, happen when you try to copy the biological process itself and make the factory behave like the organism.

The spider produces silk through a biological process shaped by biological constraints and biological economics. Trying to replicate that process in an industrial context, with industrial constraints and industrial economics, is a category error.

The Sunk Cost of Commitment

By the mid-2000s, many spider silk researchers understood this problem. The pure biomimicry approach of replicating the spinneret and mimicking the natural process as closely as possible wasn’t working. The closer a process came to copying the spider, the slower and costlier it ran.

But by then, hundreds of millions of dollars had been spent on this approach. Companies had built their technology stacks around biomimetic spinning. They’d hired biologists who specialized in spider physiology. They’d filed patents describing bio-inspired manufacturing processes.

Pivoting away from biomimicry meant admitting that the fundamental approach had been wrong, writing off years of research and explaining to investors why the core strategy needed to change.

So some companies were slow to pivot, and kept trying to make the biomimetic approach work, adjusting parameters and conditions in pursuit of marginal improvements within a flawed framework.

Once a company has committed to a direction intellectually, financially and organizationally, changing course is hard.

The survivors, the companies still working on spider silk today, have mostly abandoned pure biomimicry. They’ve shifted to what might be called bio-inspiration: using principles from spider silk (the protein structure, the crystalline-amorphous architecture) while completely redesigning the manufacturing process for industrial reality.

Some have moved beyond spider proteins, designing new proteins (and, in academic labs, synthetic polymers) that borrow silk’s molecular architecture. They make no attempt to rebuild the spider’s gland, relying instead on careful molecular design that borrows concepts from nature without trying to copy nature’s implementation.

These approaches might work. They are no longer biomimicry in the strict sense.

What Spider Silk Taught Us

What failed in spider silk was the strategy. The story is a case study in how following nature too literally can lead you systematically away from viable innovation.

If an industrial spider silk ever arrives, it will come from understanding what makes the fiber work at the molecular level and then designing a different process that reaches similar results with industrial methods, industrial economics and industrial constraints.

That is roughly the path the surviving companies have taken: borrowing silk’s molecular architecture while redesigning the manufacturing process for industrial reality.

The spider sits in its web, a beautiful solution to a problem we don’t have, and the industry spent years trying to copy it because it looked elegant.

Related articles

• The rare earth supply chain: the hidden dependency behind every modern device
• The science of aging: why we age, why some species barely do, and whether the process is inevitable
• The Great Stagnation: why productivity growth slowed and whether it can recover

Further reading

As an Amazon Associate, Eikleaf earns from qualifying purchases. The book links below are affiliate links: if you buy through them, we may earn a small commission at no extra cost to you.

Recommended videos

Gen AI Disclaimer

Some contents of this page were generated and/or edited with the help of a Generative AI.

Media

Vũ Bụi – Pexels

Andrii Lobur – Pexels

Chokniti Khongchum – Pexels

Soumyojit Sinha – Pexels

Engin Akyurt – Pexels

Sinan – Pexels

Erick Horta Reyes – Pexels

Rafael Minguet Delgado – Pexels

Jason Reid – Pexels

References

Gosline, Guerette, Ortlepp & Savage (1999). The mechanical design of spider silks: from fibroin sequence to mechanical function. Journal of Experimental Biology 202:3295 https://journals.biologists.com/jeb/article/202/23/3295/8324/The-mechanical-design-of-spider-silks-from-fibroin

Agnarsson, Kuntner & Blackledge (2010). Bioprospecting finds the toughest biological material: extraordinary silk from a giant riverine orb spider. PLoS ONE https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0011234

Spider silk module (extensibility and strength of radial threads) – L. Gross, University of Tennessee https://lgross.utk.edu/LGrossTIEMwebsite/home/gross/public_html/bioed/bealsmodules/spider.html

Spider silk – School of Chemistry, University of Bristol https://www.chm.bris.ac.uk/motm/spider/page2.htm

Selden, Shear & Sutton (2008). Fossil evidence for the origin of spider spinnerets. PNAS https://doi.org/10.1073/pnas.0809174106

Xu & Lewis (1990). Structure of a protein superfiber: spider dragline silk. PNAS 87:7120 https://doi.org/10.1073/pnas.87.18.7120

Nexia’s transgenic “spider goat” to produce milk of steel – The Globe and Mail (13 January 2000) https://www.theglobeandmail.com/report-on-business/nexias-transgenic-spider-goat-to-produce-milk-of-steel/article1035969/

Lazaris et al. (2002). Spider silk fibers spun from soluble recombinant silk produced in mammalian cells. Science https://doi.org/10.1126/science.1065780

Healthcare Acquisition Corp. definitive proxy statement, DEF 14A (12 July 2007; describes PharmAthene’s March 2005 purchase of Nexia’s Protexia assets) – SEC EDGAR https://www.sec.gov/Archives/edgar/data/1326190/000114420407036639/v081056_def14a.htm

A web of possibilities: Utah researcher uses goats to make one of the strongest known substances – Deseret News (5 July 2011) https://www.deseret.com/2011/7/5/20374017/a-web-of-possibilities-utah-researcher-uses-goats-to-make-one-of-the-strongest-known-substances

USU spider silk research – Utah State University Today (2018) https://www.usu.edu/today/index.cfm?id=57744

100 Years of Insulin – US Food and Drug Administration https://www.fda.gov/about-fda/fda-history-exhibits/100-years-insulin

Scheller et al. (2001). Production of spider silk proteins in tobacco and potato. Nature Biotechnology https://doi.org/10.1038/89335

Teulé et al. (2012). Silkworms transformed with chimeric silkworm/spider silk genes spin composite silk fibers with improved mechanical properties. PNAS https://doi.org/10.1073/pnas.1109420109

Xu et al. (2018). Mass spider silk production through targeted gene replacement in Bombyx mori. PNAS https://doi.org/10.1073/pnas.1806805115

Mi et al. (2023). High-strength and ultra-tough whole spider silk fibers spun from transgenic silkworms. Matter https://doi.org/10.1016/j.matt.2023.08.013

Edlund, Jones, Lewis & Quinn (2018). Economic feasibility and environmental impact of synthetic spider silk production from Escherichia coli. New Biotechnology https://doi.org/10.1016/j.nbt.2017.12.006

Comparing spider silk production technologies – Kraig Biocraft Laboratories (company marketing page) https://www.kraiglabs.com/comparison/

Knight & Vollrath (2001). Changes in element composition along the spinning duct in a Nephila spider. Naturwissenschaften https://doi.org/10.1007/s001140100220

Andersson et al. (2014). Carbonic anhydrase generates CO2 and H+ that drive spider silk formation via opposite effects on the terminal domains. PLoS Biology https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.1001921

Young, Holland, Shao & Vollrath (2021). Spinning conditions affect structure and properties of Nephila spider silk. MRS Bulletin 46:915 https://doi.org/10.1557/s43577-021-00194-1

Rammensee, Slotta, Scheibel & Bausch (2008). Assembly mechanism of recombinant spider silk proteins. PNAS https://doi.org/10.1073/pnas.0709246105

Holland, Terry, Porter & Vollrath (2006). Comparing the rheology of native spider and silkworm spinning dope. Nature Materials https://doi.org/10.1038/nmat1762

Andersson et al. (2017). Biomimetic spinning of artificial spider silk from a chimeric minispidroin. Nature Chemical Biology https://www.nature.com/articles/nchembio.2269

Kevlar Aramid Fiber Technical Guide – DuPont (archived copy) https://web.archive.org/web/20250503122604/https://www.dupont.com/content/dam/dupont/amer/us/en/safety/public/documents/en/Kevlar_Technical_Guide_0319.pdf

US3767756A – Dry jet wet spinning process (DuPont) – Google Patents https://patents.google.com/patent/US3767756A/en

US4134882A – Poly(ethylene terephthalate) filaments (DuPont) – Google Patents https://patents.google.com/patent/US4134882A/en

Spider silk cape goes on show at V&A – The Guardian (24 January 2012) https://www.theguardian.com/artanddesign/2012/jan/24/spider-silk-cape-show

Golden spider silk – Victoria and Albert Museum https://www.vam.ac.uk/articles/golden-spider-silk

Largest item of clothing woven from spider silk (1,896 g cape, V&A, 2012) – Guinness World Records https://www.guinnessworldrecords.com/world-records/101277-largest-item-of-clothing-woven-from-spider-silk

Réaumur’s assessment of Bon’s spider silk (55,296 and 663,552 spiders per pound), in Barrow (1758), vol. 2 – Heidelberg University Library digital copy https://digi.ub.uni-heidelberg.de/diglit/barrow1758bd2/0307

Spider silk and Réaumur’s 55,296-spider estimate – Household Words, vol. II (1850), Dickens Journals Online https://djo.org.uk/household-words/volume-ii/page-67.html

Liebsch, Fliess, Kuhbier, Vogt & Strauss (2020). Nephila edulis: breeding and care under laboratory conditions. Development Genes and Evolution 230:203 https://pmc.ncbi.nlm.nih.gov/articles/PMC7127998/

Bolt Threads brings a glimpse of the future of advanced materials to SXSW (Microsilk tie) – PR Newswire (10 March 2017) https://www.prnewswire.com/news-releases/bolt-threads-brings-a-glimpse-of-the-future-of-advanced-materials-to-sxsw-300421756.html

Givaudan acquires AMSilk’s cosmetics business – Labiotech.eu (2019) https://www.labiotech.eu/trends-news/givaudan-amsilk-spider-silk-cosmetics/

AMSilk’s silk protein yarns debut in Balenciaga collection (Spring 2026) – AMSilk https://www.amsilk.com/amsilks-silk-protein-yarns-debut-in-balenciaga-collection/

Spiber officially commences operations under new management – Spiber (2026) https://spiber.inc/en/news/new-spiber

Spiber out-of-court workout – Business Insider Japan (March 2026) https://www.businessinsider.jp/article/2603-spiber-out-of-court-workout/

Kraig Biocraft Laboratories ships its first commercial spider silk yarn order – Kraig Biocraft Laboratories (29 September 2026) https://www.kraiglabs.com/kraig-biocraft-laboratories-ships-its-first-commercial-spider-silk-yarn-order-for-performance-sports-brand/

History – Kraig Biocraft Laboratories https://www.kraiglabs.com/history/

From Human Genome to Materials “Genome” – US DOE Office of Science (2014) https://science.osti.gov/Science-Features/News-Archive/Featured-Articles/2014/127050

Materials Market Report 2025 – Textile Exchange https://textileexchange.org/app/uploads/2025/09/Materials-Market-Report-2025.pdf

Outlook of the global aramid fiber industry – Doshine (vendor page) https://www.doshinematerial.com/news/outlook-of-the-global-aramid-fiber-industry-84086734.html

Nylon 6 and nylon 66 price guide – Zhejiang Derun (vendor page) https://www.derunnylon.com/News/nylon-6-and-nylon-66-price-guide-costprice-per-kg-from-china

Margins by sector (US), January 2026 – A. Damodaran, NYU Stern https://pages.stern.nyu.edu/~adamodar/New_Home_Page/datafile/margin.html

World Bank Commodity Price Data (Pink Sheet), October 2026 https://thedocs.worldbank.org/en/doc/74e8be41ceb20fa0da750cda2f6b9e4e-0050012026/related/CMO-Pink-Sheet-October-2026.pdf

The Velcro IP story – WIPO https://www.wipo.int/en/web/ip-advantage/w/stories/the-velcro-ip-story

Wang, Q. & Schniepp, H. C. (2018). Strength of Recluse Spider’s Silk Originates from Nanofibrils. ACS Macro Letters, 7(11), 1364–1370 (author PDF) https://nanomat.as.wm.edu/wiki/_media/public/news/blog/2018-paper-nanofibrils_in_recluse_silk.pdf

Dyneema comprehensive fact sheet – Van Walt https://www.vanwalt.com/pdf/fact-sheets/Dyneema-Comprehensive-factsheet-UHMWPE.pdf

Nyffeler & Birkhofer (2017). An estimated 400-800 million tons of prey are annually killed by the global spider community. The Science of Nature https://doi.org/10.1007/s00114-017-1440-1