What's Behind the SynBio Bust?

And what lessons can we learn?

Midjourney, “synthetic biology, biomanufacturing, bioengineered microbes”

Not too long ago, synthetic biology was fashionable.

Then, in quick succession, we saw massive stock price collapses of three unicorn synbio startups in 2021-2023: Ginkgo Bioworks, Zymergen, and Amyris.

What went wrong? And can synthetic-biology startups be successful at all?

Biomanufacturing Basics

“Synbio”, as far as I can tell, is the 21st-century buzzword term for “well-funded new biomanufacturing company.”

Biomanufacturing generally refers to the industrial-scale production of materials from simple living things. In other words, extracts produced from cultures of yeast, bacteria, fungi, algae, or other types of cells, typically grown in bioreactor vats.¹

Fermented foods and beverages like beer and koji are the oldest biomanufactured goods, and they played a pioneering role in the development of industrial biomanufacturing techniques. William Sealy Gossett, the statistician behind the “Student’s t distribution” and some of the first concepts of experimental design based on “statistical significance”, developed his techniques as the Head Brewer of the Guinness beer company.

Many other substances are routinely biomanufactured, from industrial enzymes (used in food and beverage processing, detergents, and cosmetics) to chemical compounds used in biodegradable plastics and cleaning agents, to food additives like citric acid and xanthan gum.

Proteins and peptides are almost always biomanufactured, including protein and peptide drugs (e.g. insulin, growth hormone, and antibiotics like bacitracin) and the antibodies used in many diagnostic tests². Viral vectors for vaccines and gene therapies are also biomanufactured in mammalian cell culture.

Most large, complex biological molecules must be synthesized in living things, the way they are in nature. It can be very difficult to find a chemical synthesis method that can do what a microbe can.

This is the source of the famous organic chemist’s prayer:

“Lord, I fall upon my knees

And pray that all my syntheses

May no longer be inferior

To those conducted by bacteria.”³

However, while microbial syntheses may be very “efficient” locally (thanks to billions of years of evolution designing enzymes to catalyze them), it’s usually more expensive to biomanufacture a compound than to synthesize it artificially. An artificial synthesis only needs enough materials and energy to run one chemical reaction; a biological route involves all the “overhead” of keeping the organism alive! Then, you have to extract the compound from the organism, which can itself be a messy business.

Traditionally, it’s almost impossible for a biomanufacturing process to compete on price with an artificial synthesis of the same compound. This is, for instance, why biofuel companies have such a shaky track record. It’s just hard for any biofuel to compete with fossil fuels or solar on cost per joule.

The same is true for biomanufactured alternatives to plastics and other organic compounds, which are usually products of the petrochemical industry.⁴

Simple molecules, produced thermochemically (i.e. from petroleum products, without living things, with the chemical reaction energy coming from heat), are invariably cheaper than biomanufactured molecules. Biomanufacturing in practice is only used for highly complex (and expensive) compounds that can’t be made thermochemically at all.⁵ It’s unlikely that biomanufactured alternatives to the entire petrochemical industry can be economically viable without government intervention.

So you’d think the sweet spot for biomanufacturing startups would be producing novel, high-value, low-production-volume complex biologics. Maybe their comparative advantage would be coming up with a new protein-based drug or catalyst, or more efficient manufacturing techniques. Unfortunately, as we’ll see…they kinda didn’t do this.

Amyris: (Energy) Math is Hard, Let’s Go Shopping!

Amyris was founded in 2002 around a genuine scientific advance — the synthesis of artemisinin, an antimalarial drug, from genetically engineered yeast. But malaria is a disease of poor countries and they were never going to make serious money on that.

Amyris pivoted to biofuel around 2006, on the advice of their investors. Their goal was using fermentation on Brazilian sugarcane to make a “green” form of diesel. The biodiesel performed well in tests, but Amyris struggled to meet the 40-50 million-liter production quotas that are standard in the fuel industry. Their yields were poor; the yeast kept dying and bacterial contamination was a recurring problem. And the production costs were far too high. In their first large contract, with Rio de Janeiro’s and Sao Paolo’s bus transit systems, Amyris was selling the fuel at a loss even with Brazil’s hefty public subsidy. It’s simply difficult to develop a new biofuel production process that can scale to compete with conventional diesel before you run out of money.

So in 2012, Amyris pivoted again, away from biofuels to personal care and beauty ingredients, which call for smaller production volumes and command higher price-per-liter.

These were still relatively simple molecules which have alternative production methods (either extraction from plants/animals or total chemical synthesis.) Amyris didn’t have much of a moat. Nevertheless, they generated substantial sales from these compounds and launched their own skincare line in 2016. But it still wasn’t enough to pay for their “highly trained personnel and sprawling infrastructure of industrial fermentation vats.” Eager for more revenue, they went on a buying spree, acquiring many celebrity beauty companies. They got up to $153M in revenue in 2019 — but they never managed to reach profitability. By 2023, they were filing for bankruptcy.

Biomanufacturing math is unforgiving. Developing new processes costs time and skilled labor; and it’s hard to get production costs down and sales up fast enough to achieve profitability in a few years.

Zymergen: Yeast Ain’t Everything

After raising over $100M from investors to biomanufacture new industrial materials and valued at nearly $5B in 2021, Zymergen experienced a precipitous stock price plunge later that year, finally filing for bankruptcy in 2023. Its first product, Hyaline, a film to be used in foldable phone screens, had been a commercial failure, and Zymergen had no other sources of revenue.

What went wrong?

Zymergen, like Amyris, created real innovations: high-throughput lab automation systems, machine-learning methods for genetically optimizing microbes, and so on. But all of this was focused on the first part of the biomanufacturing process: generating a new yeast strain that produced a chemical of interest and had good performance characteristics in the lab.

Zymergen was a yeast strain engineering company, but not a manufacturing company. They had little in-house manufacturing capacity, and outsourced production to contract manufacturing organizations, who struggled to meet production goals. All Zymergen did was design the yeast strains. Growing the yeast at scale, extracting the key molecule, and processing it into a polymer film — all of these steps were left to the contractors, with little hands-on attention from Zymergen.

According to their 2021 prospectus, Hyaline was not even yet being produced by fermentation, but was “a non-fermentation produced biomolecule sourced from a third party” which they planned to “convert to a fermentation-produced molecule” in 2022. They had limited access to CMO manufacturing capacity, leading to the warning that “we may not have the manufacturing capacity required to meet our commercial needs after the end of this year.”

Not having comparative advantage at the manufacturing process itself, Zymergen had few options when samples of their flagship product Hyaline failed quality tests and prospective customers hesitated to make orders.

This is not how you develop a manufacturing process. Anywhere.

“Process dev” (the department that designs and optimizes the process) and “manufacturing” (the department that executes it) need to work closely together and maintain tight feedback loops. Manufacturing departments are archetypally conservative; their job is to carefully execute a given procedure and not fuck it up. They are not (stereotypically) innovators and tinkerers; that’s process dev’s job.

The situation is analogous if your entire company is “process dev” and your “manufacturing department” is an external contract organization. You need to be onsite doing continual testing and adjustment of the process to get it working well. Vendor relationships in manufacturing need to be very tight and interdependent; your vendors are your coworkers.

This is bog-standard advice. It’s repeated ad infinitum in Deming’s Out of the Crisis, which has defined how manufacturing works since the 1980s. If you are sending a new, untested manufacturing process off to a third-party facility and washing your hands of it, you will have devastating quality problems.

Zymergen, it seems, wasn’t thinking like a manufacturing company.

Ginkgo Bioworks: You May Have To, Like, Actually Sell Things

Ginkgo is technically still in business, but their stock price plummeted almost to nothing in 2021 and hasn’t recovered, following a devastating short seller’s report that revealed that virtually all their apparent revenue came from the accounting shenanigans of “selling” to internally funded and operated “customers.”

In Ginkgo’s 15 years of operation, they’ve reported 219 “programs”, only 6 of which are “commercialized.” Despite its heavy investment into R&D, with highly automated labs designed for testing many genetic modifications to yeast in parallel (they claim to be able to “create 50,000 different genetically modified cells in a single day”) they have failed to come up with a blockbuster product.

They do seem to have built cool lab automation and computational biology workflows — I’m a big fan of their blog — but, one way or another, it never translated to profitability.

Even their most successful products, aromachemicals sold to fragrance firm Robertet, are uneconomical use cases for biomanufacturing. I happen to know that the peach scent gamma-decalactone has been cheaply chemically synthesizable since the turn of the 20th century. (Lactones are responsible for the rich, long-lasting peach scent of Guerlain’s Belle Epoque masterpiece Mitsouko.)

Likewise, other Ginkgo products, like their nitrogen-fixing bacterial product for agriculture, may struggle to find customers when chemically manufactured fertilizers are so cheap. (Again, the Haber-Bosch process is a hundred years old!)

Did Ginkgo’s highly automated biomanufacturing process give them a speed or price advantage over competitors? Again, no! Customers complained of high costs and delayed deliveries, and internal documents reveal policies of padding invoices.

So, unsurprisingly, their revenues were low, and even some of the “customers” that were actually Ginkgo-owned sockpuppets were unhappy and switched to competitors.

Takeaways

None of these criticisms needs to reflect badly on the vast majority of current or past employees of these three companies. By all accounts, they hired serious scientists and engineers who did a great deal of quality work.

The problems were at the strategic and management level, with making biomanufacturing economically viable.

It simply doesn’t make sense to biomanufacture just anything. You only want to biomanufacture products that can’t be made more cheaply some other way.

And, it doesn’t make sense to specialize exclusively in microbial strain engineering and not the rest of the manufacturing process, when you’re developing new products that will necessarily require manufacturing process development.

Notably, Amyris, which did the best of the three at achieving good manufacturing performance and revenue, reports having had the most difficulty achieving satisfactory production results when they were working with a CMO that didn’t allow the participation of Amyris staff. It really pays to be hands-on.

Biomanufacturing isn’t doomed as an industry, of course — it’s been around for centuries.

The real question is whether new, high-tech startups can succeed in an industry dominated by incumbents. I don’t think it’s impossible, but I think it would be important to be manufacturing-oriented from the start and to choose products very carefully to maximize comparative advantage.

1. It’s not quite exclusively unicellular organisms — many biomanufactured products are produced by molds, like the Penicillium genus from which penicillin was first extracted. Molds, also known as filamentous fungi, are technically multicellular organisms.

2. An immunoassay detects the presence of a compound by its reaction with an antibody that binds to it. Pregnancy tests and rapid COVID tests are immunoassays; so are many blood tests.

3. https://pubs.acs.org/doi/10.1021/acscentsci.0c01496

4. https://www.orcasciences.com/articles/checking-my-prejudices-on-materials-decarbonization

5. Complex molecules like proteins break down at high heat; they need finesse and mild temperatures. Simple molecules, by contrast, can be synthesized faster if you crank up the heat and pressure, and fossil fuels are great for that.