AI-Designed Enzymes Just Brought Manufacturing Outside the Cell

Two hundred liters of hyaluronic acid, made without a single living cell. That pilot run, completed in late 2025 by Israeli biotech Enzymit with Dutch cooperative Cosun, is the first scaled proof that the cell itself is no longer the bottleneck in biomanufacturing. Enzymes designed by AI, running outside any organism, produced multi-kilogram batches of the molecule used in dermal fillers, wound care and drug delivery. The maker chooses the polymer size by dial, not by rewriting a genome.

The field had promised this before. Enzyme engineering has spent two decades as a story about tomorrow: bio-based chemicals cheaper than petroleum, made in gentle reactors instead of cracking towers. The promise was real and the delivery kept slipping, because designing a useful enzyme took years of directed evolution in the lab. AI changed the design loop, and cell-free production changed the economics that follow.

The cell was the constraint all along

Industrial fermentation works. It also carries limits baked into biology. A microbe spends carbon on staying alive before it spends any on your product. Toxic compounds accumulate and kill the batch. A strain that behaves in a 10-liter flask needs years of re-engineering to survive a 100,000-liter tank. Cell-free systems remove those constraints by construction: enzymes extracted from their host and run in solution, with the pathway engineered around the product instead of around the organism's survival.

The pilot run matters less for its size than for what it controls. The same enzymatic process produces hyaluronic acid from roughly 10 kilodaltons to 4 megadaltons with a narrow size distribution, adjusted through standard production settings rather than by engineering a new organism. In cell-based manufacturing, molecular weight is a property you inherit from the strain. In a cell-free system it becomes a process parameter, which is a different kind of industrial asset.

The players with proprietary enzymes

Four companies carry the near-term commercial thesis, and they divide the field in instructive ways. Enzymit (backed by Khosla Ventures and Grove Ventures) owns the scaled cell-free process proof. Scindo in London sells cell-free routes to ingredients for flavors, fragrances and cosmetics, and counts a food major's venture arm among its investors. eXoZymes, listed on Nasdaq, is the market's liquid bet on cell-free natural products. Arzeda in Seattle leads a $7.8 million US National Science Foundation consortium on electricity-driven enzyme cascades.

This isn't just a breakthrough in hyaluronic acid production — it's a glimpse into the future of how we make molecules. We've proven that precision manufacturing of biomolecules is no longer limited by what cells can or can't do.— Gideon Lapidoth, CEO, Enzymit

Scindo, founded in 2020 by Gustaf Hemberg, Ben Davis and Juliet Sword, raised £4 million in seed funding co-led by Kadmos Capital and Clay Capital, with participation from PINC, the venture arm of Paulig. The company maps enzyme functions that traditional chemistry struggles with, carbon-hydrogen activations and carbon-carbon bond cleavages, and converts renewable feedstocks directly into ingredients. Two products, one for flavor molecules and one for cosmetics, sit near pilot scale, and both rely on the same cell-free logic as Enzymit's process.

The investment logic differs across the four. Enzymit and Scindo are private, venture-backed bets on process economics. eXoZymes gives an investor direct exposure through public equity: its Q2 2026 report in August confirmed pilot-scale runs with Cayman Chemical and a collaboration with Lawrence Berkeley National Laboratory on AI-driven digital twins of enzyme manufacturing, funded at roughly $747,000. Management reported taking a molecule from concept to pilot validation within two years.

Arzeda's angle is the furthest from market and the closest to a step change. Its NSF consortium, run with Missouri University of Science and Technology and Novonesis, targets electricity-driven enzyme cascades that convert carbon dioxide into methanol, aiming above 70% energy efficiency against losses of more than 50% typical of chemical electrocatalysis. If cell-free manufacturing removes the organism, electricity removes the feedstock. CO₂ becomes the input.

With AI-designed enzymes that run directly on electrons, we can bypass many limitations of living cells and build cleaner, more efficient routes for difficult-to-synthesize critical chemicals like methanol.— Alexandre Zanghellini, CEO, Arzeda

Why the economics shifted

Two costs fell simultaneously. AI compressed the enzyme design cycle from years to weeks, and gene synthesis plus high-throughput screening got an order of magnitude cheaper. That combination is what makes cell-free production viable as a business rather than a laboratory curiosity: the enzyme is now cheap enough to design, make and recycle at industrial volume.

Scindo's investor framing makes the market logic explicit. The specialty chemicals industry still leans on petroleum derivatives for flavors, fragrances and cosmetic ingredients. The alternative is an enzyme that performs the same transformation on renewable or waste carbon. Clay Capital partner Ali Morrow described the approach as designing molecular craftsmen, enzymes built for specific industrial jobs that unlock feedstocks the petrochemical route cannot use.

Scindo's approach creates molecular craftsmen — enzymes designed for specific industrial jobs that offer cost-competitive natural alternatives and unlock previously inaccessible feedstocks, creating significant opportunities globally to end the industry's reliance on crude oil.— Ali Morrow, Partner, Clay Capital

The addressable prize is large. The specialty chemicals and ingredients industry still sources flavors, fragrances and cosmetic components from petroleum. Enzymes touch only a sliver of that value chain today. As we wrote in August, when LanzaTech pushed carbon recycling toward a commercial ethanol barge, the pattern repeated: a process that works at pilot scale only becomes an investment thesis once a plant runs continuous shifts. Industry market research points to 50–70% energy savings and 40–60% water savings from enzyme-based processes against conventional chemistry. Those are the kind of numbers that eventually move procurement decisions, not just laboratory budgets.

The history of enzyme promises argues for caution

None of this is guaranteed to scale. Cell-free production has carried a cost problem since its academic origins: purified enzymes and cofactors are expensive, and running a pathway for long enough to be economic demands stability that natural enzymes do not always deliver. Twenty years of bio-based chemical announcements that did not reach commercial volume are a standing reminder that pilot success and plant economics are different things.

The competing startups illustrate the technical risk too. Manchester's Imperagen combines quantum-physics simulation with AI and robotic labs to report 500-fold enzyme improvements across five cycles. Biomatter in Lithuania built its generative platform on a protein GAN and lists Thermo Fisher Scientific, BASF and Kirin as partners. Cradle works with Novonesis, Bayer and Lundbeck. A field with this many credible players usually means the technology works and the differentiation is still to be proven.

What happens to cell-free biomanufacturing a year from now?

Development scenarios

Sources