CIRCE.

Here is a thing about carbon dioxide that is easy to forget: it is a molecule, and molecules are useful. The whole climate industry has spent a decade trying to make CO2 disappear - pump it underground, mineralize it, pay someone to not emit it. Circe Bioscience, a Boston company that spun out of Harvard's Wyss Institute in 2021, looked at the same molecule and drew a different conclusion. If CO2 is carbon and oxygen, and carbon and oxygen are what most useful things are made of, then the emission is not a waste product. It is inventory.

The mechanism is gas fermentation, which sounds exotic but is roughly what plants have been doing since before there were people to burn things. Circe engineers microbes - the metabolic wiring inside them - so that instead of eating sugar, they eat gas. Feed them carbon dioxide, water and electricity, and they build triglycerides, the precursor molecules of fats and oils. They can also be tuned to make sugars, biodegradable plastics and fuels. The inputs are three items long. The outputs are, in principle, most of the chemicals industry.

The number that makes this interesting rather than merely clever is 3:1. For every kilogram of oil the process produces, it consumes about three kilograms of CO2. This is the part worth pausing on. Plenty of companies sell "sustainable" versions of things that are marginally less bad than the incumbent. Circe's arithmetic runs the other direction - the more product it makes, the more carbon it pulls out of the air. Carbon-negative is not a slogan bolted onto a marketing deck here; it is the stoichiometry of the reaction.

Circe was conceived by Shannon Nangle and Marika Ziesack in the lab of Wyss faculty member Pamela Silver, and it carries the intellectual DNA of an academic project that decided to become a business. That transition is where most deep-tech science quietly dies - the gap between a result in a paper and a molecule someone will pay for is wide and full of expensive equipment. Circe crossed a meaningful chunk of it in three years, taking a worldwide exclusive license to the technology in 2024 and, by late that year, standing up a pilot plant.

The proof-of-concept that got everyone's attention was, of all things, a chocolate bar. Cocoa butter is the fat that gives chocolate its snap and its melt, and it comes from cacao trees, which grow in a narrow tropical band under increasing climate stress and with a supply chain that has its own troubles. Circe's microbes make a cocoa butter that is molecularly identical to the tree-grown version - same molecule, no tree, no deforestation. The company used it to make what it calls the world's first chocolate produced with fermentation-derived cocoa butter, or, as the press cheerfully put it, chocolate made out of thin air.

This is a good demonstration for a reason that goes beyond novelty. If you can make cocoa butter, you have shown that the platform hits a specific, high-value, molecularly precise target - not a vague "bio-oil" but the exact fat a confectioner needs. That precision is the whole thesis. Circe is not selling one product; it is selling a process that is, in its own description, industry-agnostic, having produced prototypes spanning confectionery, dairy ingredients, sugars, proteins, polymers and biofuels.

Whether that breadth is a strength or a temptation is the open question every platform biotech faces. Doing one molecule extremely well is a business. Doing twelve molecules adequately is a research program. Circe's answer, sensibly, has been to start with fats and oils - high value, clear buyers in food and cosmetics - and let the platform's range be a promise for later rather than a distraction now.