Spider Silk — Stronger Than Steel, Tougher Than Kevlar

A spider, falling on its dragline, hangs from a fiber thinner than a human hair that will not break. The same fiber, scaled up, would carry the weight of a small car. We have known this for at least fifty years. We still cannot make it at scale. The gap between what a spider does in a backyard and what DuPont can do in a billion-dollar plant is one of the cleanest unsolved problems in materials science.

The numbers that make this strange

Dragline silk from Nephila and similar orb-weavers measures around 1.1 GPa in tensile strength — the range of high-grade steel, and several times steel's strength per unit weight (silk density ~1.3 g/cm³ vs steel's 7.8). Toughness — the area under the stress-strain curve, the energy a fiber absorbs before snapping — runs around 180 MJ/m³. Kevlar sits near 50. Caerostris darwini, the Darwin's bark spider described in 2010, exceeds 350 MJ/m³, the toughest biological material on record.

Toughness is the trick. Kevlar is stronger in pure tensile pull. Silk is tougher because it stretches — dragline silk reaches 40–50% strain before failure, and the flagelliform capture spiral stretches over 500%. A bullet, a wind gust, an insect hitting a web at speed — these are toughness problems, not strength problems. Silk wins the toughness fight by an order of magnitude.

A single spider produces these properties from a gland the size of a sesame seed, using ambient temperature, water as the solvent, and roughly one second of processing time per centimeter of fiber.

Why we can't farm them

The economics fail before the chemistry does. Spiders are territorial and cannibalistic — house them together and the silk farm becomes a feeding pen. Each animal yields milligrams of silk per day. Bombyx mori, the domesticated silkworm, has been bred for docility since at least 3000 BCE, but its fibroin has a different mechanical profile: strong, not tough.

So the field went around the spider. Since the early 1990s, spidroin genes have been expressed in:

Bacteria (E. coli) — fails because dragline spidroins are very large, highly repetitive proteins that bacterial translation chokes on. Truncated proteins lose toughness.
Yeast — handles larger proteins; Bolt Threads' Microsilk and others used this route.
Goats — Nexia's BioSteel goats (Canada, 2002) secreted spidroin in milk. Fiber quality was passable; the company folded in 2009; AMSilk acquired the IP.
Transgenic silkworms — Kraig Biocraft has engineered silkworms producing spider-silkworm hybrid fibers; their 2026 BAM-1 line measures 1.79 GPa, near the natural dragline ceiling.
Designed-from-scratch proteins — Spiber (Japan) skips spider biology entirely, designing sequences for production fitness, not biological fidelity. Their Thailand plant runs at ~100 tons/year as of 2025.

The proteins, by now, are not the bottleneck. The spinning is.

The spinning secret

Inside a spider's major ampullate gland, spidroin sits as a liquid crystalline dope at ~50% protein concentration. As the spider draws it through a narrowing duct toward the spinneret, four things happen in sequence over millimeters:

pH drops from ~7.2 to ~5.7, flipping protein conformation.
Sodium and chloride leave the duct; potassium and phosphate enter, driving hydrophobic collapse.
The spider mechanically pulls the emerging fiber, aligning the chains.
Disordered regions snap into crystalline beta-sheets, embedded in an amorphous elastic matrix.

The result is a hierarchical nanocomposite — rigid crystalline "bricks" suspended in elastic "mortar." Industrial wet-spinning from solvent reproduces around 60–70% of natural properties. The protein is right; the spin is wrong. A 2025 paper in Advanced Functional Materials (Sulekha et al.) reported a disulfide-locked spidroin that lets manufacturers control the liquid-to-solid transition externally — the most direct attack on the spinning problem to date.

What's contested

Three open questions sit at the field's core.

First, the property ceiling. Whether recombinant silk can ever fully match wild dragline silk — or whether some part of the spider's spinning duct geometry is genuinely irreproducible at industrial scale — is unresolved. Kraig's 1.79 GPa results approach the natural ceiling for tensile strength, but toughness numbers in independent labs remain below wild silk.

Second, scale economics. The synthetic spider silk market sat near $74M in 2025, projected to roughly double by 2035. Compared to nylon (millions of tons annually), this is a specialty material. Whether silk ever becomes a commodity textile or stays in luxury and biomedical niches is a business question, not a biology one.

Third, the design philosophy. Spiber's bet is that nature's sequences are evolved for survival, not manufacturability — that a top-down designed spidroin will beat a copied one. Kraig's bet is that the wild sequence, expressed in a domesticable host, is the shorter path. Both are shipping product. Neither has won.

Why this has to do with other realms

The spider silk problem is the inverse of the tech jacquard loom problem. The Jacquard loom (1804) took an existing biological fiber — silkworm silk — and made the pattern programmable, turning textile into early computing. Spider silk biotech does the opposite: it tries to make the fiber itself programmable, designing protein sequences to spec. One end of the textile chain became digital two centuries ago. The other end is becoming digital now, via concept synthetic biology and engineered organisms. The full chain — from genome to garment — has never been programmable end to end. It is about to be.

An open question

If a spinning process is more important than the protein it spins, what is the minimum non-biological apparatus — pH gradient, ion exchange, draw geometry — that could spin any protein dope into a fiber matching wild dragline silk? Nobody has built that machine yet.

Key sources

Spider Silk: Evolution and 400 Million Years of Spinning, Waiting, Snagging, and Mating — Leslie Brunetta and Catherine Craig (2010). Best lay-accessible book on the evolutionary biology behind the material.
Sulekha et al., Advanced Functional Materials (2025) — disulfide-locked spidroin spinning control. The current state-of-the-art spinning paper.
Breslauer (2024), Advanced Functional Materials — review of microbial-production scaling and the spinning-vs-protein gap. To verify: exact title and issue.
Agnarsson, Kuntner, Blackledge (2010), PLoS ONE — original Caerostris darwini toughness measurements.
Spiber Inc. annual disclosures and Kraig Biocraft investor filings — only sources for current production volumes; commercial, treat skeptically.

Abhishek's take

I watch this every season when the synthetic-fiber vendors pitch "the next big thing" for performance wear. The lab numbers on spider-silk blends are always impressive—until you run the lead-time math. A 100-day production cycle for recombinant protein fibers means you’re committing to a trend call before the previous season’s sell-through data even lands. The only floor I’ve seen bet seriously on this was for a limited-edition collaboration where the marketing story (not the margin) justified the 3x unit cost. Even then, we had to lock the buy nine months out and eat the overstock when the hype faded faster than the fiber’s tensile strength. The real constraint isn’t the material—it’s that fashion moves at the speed of memes, and biology still moves at the speed of silkworms.

Abhishek