Put a colony of bacteria on a cathode, feed it CO₂ and surplus electricity, and it will turn them into acetic acid — a platform molecule every chemical plant already knows how to use. The science is settled enough that the arguments have moved past chemistry. They now sit in engineering and economics: can a living cell take electrons directly from a wire, tolerate the interruptions of a renewable grid, and make something worth more than the power it consumes?
A Stanford team sustained acetate production at up to 2 mM/h from a pure thermophilic culture on intermittent power — a concrete step toward grid-coupled operation.
The bottleneck has shifted from biology to engineering: standardised bioreactor hardware and strain tolerance, with AI-driven design entering to compress the wet-lab cycle.
As an investment lens: still research-to-pilot, best read as a signal on where biology-based CO₂ conversion is heading rather than as today's returns.
Renewable generation rose from 2,279 TWh in 1990 to 7,504 TWh in 2020, and its share of global electricity is expected to reach roughly 36% by 2026. More electrons, cheaper, and increasingly surplus at odd hours. That surplus is the raw material this technology wants to buy at a discount.
The underlying idea is older than the term. Microbes have been reducing CO₂ for three billion years; what changed is delivery. An electrode wired into a bacterium's metabolism transfers electrons at the membrane — no sugar, no biomass feedstock, just current, gas, and a living catalyst.
The machine on the cathode
An electrosynthesis cell looks modest: an anode, a cathode, a membrane, and a biofilm. Electroactive microorganisms sit on the cathode and accept electrons the way other organisms accept glucose. Those electrons supply the reducing power that fixes CO₂ into multi-carbon organics. The main product today is acetate, the two-carbon molecule at the base of most fermentation value chains. Beyond it: volatile fatty acids, PHB (a bioplastic), methane, and with further engineering, single-cell protein (SCP).
Acetate production rate, sustained
A pure thermophilic culture powered by intermittent electricity held this rate with no crash. Environ. Sci. Technol., 2022
Concentration window sustained
Product concentration is the metric that determines whether recovery and purification make economic sense downstream. Environ. Sci. Technol., 2022
The division that matters for investors is between mixed and pure cultures. Mixed cultures are hardy — they self-select, resist contamination, and run for months — but they limit the product spectrum to a few low-value molecules. Pure cultures can be engineered, which is what opens access to high-value products; the cost is fragility and the need for controlled, sterile operation. A 2024 review in Trends in Biotechnology made the point bluntly: the field's ceiling is not biological, it is the absence of standardised, off-the-shelf electrobioreactors that would let engineered pure cultures be tested at scale.
Electrons on a schedule
Renewable power is intermittent, and most bioprocesses hate intermittency. A fermentation run assumes a steady feedstock; shut off the electrons and the culture starves, then recovers unevenly. The Stanford result is notable precisely because it refused to assume steadiness. A pure culture of Thermoanaerobacter kivui kept producing acetate at up to 2 mM/h across current interruptions, reaching an average concentration of 60–90 mM. Supplying a background current of just 1–5% of the maximum during off-times was enough to blunt the impact of the interruptions.
That changes the economic character of the technology. A process that can ride the shape of a renewable grid behaves less like a chemical plant and more like a battery — one that stores surplus electrons in chemical bonds rather than losing them to curtailment. For grid operators and asset owners paying for negative prices at midday, that is a property worth pricing. It is also the property most early research papers underplayed while lab experiments ran on steady bench power.
Where the money sits
Here is the uncomfortable part. The main product, acetate, is cheap — commodity cheap. A 2026 review in Advances in Industrial and Engineering Chemistry concluded that MES alone is unlikely to reach near-term commercial viability. The path that keeps appearing in the literature is two-stage: electrosynthesis makes the acetate, then a second bioprocess upgrades it into something worth selling. PHB as a bioplastic. Single-cell protein as a food ingredient. Isopropanol and longer-chain acids as fuels and solvents.
| Product | Pathway | Status in MES |
|---|---|---|
| Acetate | Direct reduction of CO₂ | Predominant today; commodity price |
| PHB | Single-step MES cell | Demonstrated; pilot-scale work under way |
| Single-cell protein | Power-to-protein, two-stage | Commercial siblings exist in gas fermentation |
| Longer-chain acids | Mixed cultures, C₃ and up | Now accessible; selectivity still a gap |
The two-stage framing matters because it changes what to watch. The value sits in the coupling rather than in the first reactor — a cheap, electrified first step feeding a high-margin second step. That is the same architecture that made solar power compelling for hydrogen and green steel: not the panels, but the system that prices cheap electrons into valuable molecules.
The bottleneck moved
The chemistry was never the whole problem. Acetate accumulates and poisons the culture — the Stanford paper names product inhibition as the key challenge to efficient, intermittently powered MES. Titers are low relative to conventional fermentation. Reactors are still adapted from microbial fuel cells rather than designed for electrosynthesis, which shows up as mass-transfer limits, pH drift, and biofilm stratification on the electrode.
A 2020 review asked whether microbial electrosynthesis from CO₂ was "forever a promise." Six years later the field is measurably better — pure cultures, intermittent-power tolerance, engineered strains — but not yet commercial. Treat milestone claims against that baseline.
The good news is that the failure modes are engineering problems, not biology problems. Strains with high acetate tolerance, reactors built for electron delivery rather than sugar mixing, and cheaper product recovery are all tractable targets. The literature now treats them as a checklist rather than a research wishlist.
What the AI layer changes
Every iteration in this field has run on the same expensive loop: engineer a strain, run a wet-lab experiment, wait, repeat. A 2026 perspective in Current Opinion in Chemical Engineering argued that the slow pace and cost of wet-lab research is exactly what has held bio-electrocatalytic CO₂ conversion back — and that AI and machine-learning tools are the proposed lever to compress the cycle. Strain design becomes a search problem. Electrode choice becomes a prediction problem. Process optimization becomes a recommendation problem.
Early research platforms point in the same direction. A Georgia Tech research platform couples carbon-felt electrodes with engineered Clostridium ljungdahlii and Sporomusa ovata strains, reporting carbon conversion efficiencies up to 90% and acetate production above 50 g/L/day — numbers that would have seemed optimistic a decade ago, and that were reached by choosing the biology and the electrode as one system.
Microorganisms act as living catalysts whose biological traits and energy requirements determine system performance.— Korth & Harnisch, Communications Sustainability (Nature), 2026
Why an investor should care about a lab-scale process
Two routes to the same molecule
Electrosynthesis is not the only way to turn CO₂ and electricity into chemicals. The abiotic route — pure electrochemistry, no biology — is racing for the same molecules, and a 2026 review in the Journal of Environmental Chemical Engineering covered exactly this range of options for turning carbon dioxide into value-added products. Copper-based electrodes reduce CO₂ to multi-carbon products; new catalytic materials are chasing selective multi-carbon output. On the surface, the two fields compete for the same feedstock and the same electrons.
The distinction that matters is what the catalyst is. An abiotic electrode is a fixed object: once built, its performance is set, and it degrades with use. A microbial catalyst is self-replicating and self-repairing — the cell rebuilds its own active machinery, and a reactor can run for months without replacing the catalyst. The trade-off is speed and precision. Abiotic systems generally report higher current densities and cleaner selectivity, which is why they dominate the funding and the headlines. Biology brings tolerance, longevity, and a tolerance for the messiness of real gas streams that poisons most metal catalysts.
Both routes share the same economic ceiling: they produce commodity molecules whose price is set by petrochemical producers. That is why neither route will stand or fall on the cathode alone. Both need a second stage — biology upgrading its acetate, electrochemistry upgrading its ethylene and formate — before the economics bite. The sensible reading is not a contest where one route wins and the other disappears. Both are testing the same hypothesis from opposite ends, and their cost curves will converge somewhere in the middle of the decade.
Power-to-product as a category
Step back far enough and electrosynthesis stops being a single technology and becomes a member of a broader family: power-to-product, the business of turning cheap electricity and captured carbon into molecules. The family already has commercial members. Single-cell protein from gas fermentation has been produced commercially since the 1990s, and companies such as Unibio and Calysta have demonstrated that the market accepts microbe-made feed and food ingredients. Their feedstock is different — methanol and natural gas rather than electrons — but the downstream logic is the same.
The economics of the whole family turn on one number: the price of the energy input. The 2026 review in Advances in Industrial and Engineering Chemistry framed the challenge precisely — hydrogen prices at several dollars per kilogram, and electrochemical processes that upgrade CO₂ into chemicals more valuable than hydrogen itself as the key step toward commercial viability. Electrosynthesis's bet is that its catalyst, the microbe, is cheaper per unit of product than the platinum-group and copper systems used in abiotic electrolysis, even if it is slower per square centimetre of electrode.
That is the lens a principal should use. The category is real; the valuations inside it are early and scattered. The companies worth tracking are not the ones claiming to have solved CO₂ conversion — everyone claims that — but the ones that can show a second-stage product with a real buyer, a reactor that ran continuously for months, and an electricity contract that prices surplus power correctly. Those three pieces together are the actual moat. Everything else is a lab slide.
Signals to watch
As we wrote in August, Solar Foods raised €77.8 million to scale protein made from air — the gas-fermentation neighbour to this technology. The two fields share the same bet: that surplus electricity and captured carbon can be priced below the feedstocks they replace. Electrosynthesis is the earlier, rougher version of that bet.
A pure-culture product beyond acetate reaching pilot scale — PHB or a specialty chemical
Standardised electrobioreactor hardware coming to market, the missing piece named in Trends in Biotechnology
An AI-designed strain-and-electrode combination reporting a step-change in titer
A corporate partnership or facility announcement that ties MES to a specific plant — the field's first commercial moment
The honest timeline looks like this: two to three years of reactor and strain engineering, a first generation of full-chain pilot economics built on two-stage coupling, and a commercial decision point after that. The technology does not need to beat petrochemical economics on cost at birth. It needs to beat curtailment — and on that field, it has the advantage of a three-billion-year-old catalyst.
Set aside the question of whether acetate will be made from electrons at scale in 2027. The narrower claim is that the coupling of cheap surplus power, engineered biology, and a molecule with a real buyer is a structure the market will eventually price. Every major energy transition in the last two decades was paid for by exactly that kind of coupling — cheap input, biological or chemical upgrade, marginal-cost product. Solar and wind did it with electrons. The question now is which platform learns to do it with carbon.
For the patient principal, the practical takeaway is simple and familiar from power markets. Do not underwrite the current claims, because they will almost all be wrong on timing. Underwrite the direction instead: the field is producing real, repeatable results — intermittent-power tolerance, pure-culture engineering, AI-assisted strain selection — and the cost of computing and of renewable electrons is falling while the cost of conventional feedstocks wobbles. Direction bets on declining input costs usually survive individual company disappointments. That is the same reason infrastructure funds kept buying into storage before the chemistry fully worked out.
Keep the mailbox open for three specific developments. A company that names a second-stage product with a contracted buyer. A reactor operator that publishes ninety solid continuous days of operation. And an AI-designed strain that outperforms the mixed-culture baseline in a peer-reviewed head-to-head. Any one of those three is the field's first real commercial signal.
