Industries / Food, farming and forestry / Food and beverage processing / Brewing and distilling
Fermentation
Fermentation is the stage where yeast turns the sugar in wort into alcohol and carbon dioxide, both for beer and for the wash that a distiller then puts through a still.
6
Milestones achieved, from 1873 to 2022.
2
To watch for: announced, not yet confirmed outside.
2
Open deliverables. Nobody has shipped them.
The road
AchievedWatch forOpen
Achieved
1873
Cold on demand for lager
Lager is fermented and stored cold, so brewers had depended on cellars and natural ice. In 1873 Carl von Linde, working with the Spaten brewery in Munich, built a mechanical refrigeration machine that made ice to cool its cellars, first with ether as the refrigerant; by the 1880s most breweries made their own ice by machine (Oxford Companion to Beer). The Swiss National Museum credits support from August Deiglmayr of the Dreher brewery and Gabriel Sedlmayr of Spaten, and notes that in 1883 the Swiss brewing industry still needed 100,000 tonnes of natural ice. Brewing lager all year, far from ice sources, became possible.
Credit
Carl von Linde; Gabriel Sedlmayr (Spaten); August Deiglmayr (Dreher)
1876
Spoiled beer has a cause
In Études sur la bière (1876, in English 1879), Louis Pasteur stated that "every unhealthy change in the quality of beer coincides with a development of microscopic germs which are alien to the pure ferment of beer." Sour or off batches now had a cause that could be seen under a microscope and kept out. The pure yeast culture that Emil Christian Hansen developed at Carlsberg in 1883 is covered under Food and beverage processing.
Credit
Louis Pasteur
1950s
Fermentation that never stops
Morton Coutts of Dominion Breweries in New Zealand developed continuous fermentation: wort flows steadily into one vessel where yeast grows, then into others where it ferments, and finished beer is drawn off without stopping, instead of brewing in separate batches. Coutts' later US patent (filed 1961) describes feeding boiling wort straight into "a continuous fermenting system". A beer historian dates the first continuous-fermentation brewery to 1957 at Palmerston North.
LessonThe method suits a brewery making one beer; in the early 1980s most brewers worldwide returned to batches, and by 2013 New Zealand was the only country still using a multi-vessel continuous system for beer (Stewart, Hill and Russell).
Credit
Morton Coutts; Dominion Breweries
1995
Engineered yeast stays in labs
In December 1995 BRF International, the British brewing research body, brewed a beer called Nutfield Lyte with genetically modified yeast, labeled "Brewed for research purposes"; the bottle is in the Science Museum's collection. A 2013 review found that genetically modified brewer's and distiller's yeasts were "still not routinely employed commercially, owing in large part because of opposition from public opinion."
LessonA yeast that works in trials still needs drinkers who will accept it.
Credit
BRF International
2011
Lager yeast's wild parent found
Lager yeast is a hybrid of ale yeast and an unknown cold-tolerant yeast. A team isolated that missing parent from southern beech forests in Patagonia and named it Saccharomyces eubayanus; its draft genome was 99.5% identical to the non-ale half of lager yeast's genome. This opened a source of new lager strains bred from the wild parent.
Credit
Diego Libkind, Chris Todd Hittinger, José Paulo Sampaio and colleagues (PNAS)
2022
Engineered yeast reaches breweries
In 2018 a University of California, Berkeley team engineered brewer's yeast with genes from mint and basil to make hop-flavor molecules, and a double-blind tasting panel judged its beer hoppier than hopped beer. In 2022 an Oregon State University study with the company Berkeley Yeast found that modified strains greatly raised two tropical-fruit thiols (sulfur aroma compounds) during fermentation; Berkeley Yeast said its strains were then used by more than 100 US breweries.
Credit
Charles Denby and colleagues (UC Berkeley); Berkeley Yeast; Tom Shellhammer (Oregon State University)
Watch for
Wild lager yeast made brewable
The wild lager parent, S. eubayanus, gives a clove-like off-flavor and ferments malt sugars poorly. A Delft University of Technology team reports non-genetically-modified strains, made by mutation, screening and laboratory evolution of a Himalayan isolate, that lack the off-flavor and consume more than 90 to 95% of the maltose and about 25 to 32% of the maltotriose in industrial wort.
Achieved whena brewery outside the lab brews and sells a lager with one of these strains, or another lab repeats the wort results.
Claimed by
E. Knibbe, J.-M. G. Daran and colleagues, Delft University of Technology, FEMS Yeast Research, 2026
Neural network fermentation forecasts
Sheffield Hallam University researchers trained an LSTM, a type of neural network built for data that changes over time, on 1,305 real fermentations of ales, IPAs, lagers and mixed-culture beers, and report that it predicts how gravity and other readings will change "with low errors and minimal residuals". The fermentation data came from the sensor maker Sennos.
Achieved whensomeone other than the authors tests it on fermentations from breweries that were not in its training data, or a working brewery publishes results from planning tanks with it.
Claimed by
O'Brien, Zhang and Allwood, Sheffield Hallam University, Processes, January 2026
Open
01
Open fermentation curves
A public dataset of fermentation logs (gravity, temperature and, where measured, pressure or CO2 flow, every few minutes) from many licensed breweries, with the yeast strain, starting gravity and the hour the brewer called fermentation finished; and a baseline model that predicts that hour early in the fermentation.
The briefCredit · unclaimed02
Thiol yeasts tested together
A side-by-side trial of the commercial thiol-releasing yeasts, the engineered strains sold to free tropical-fruit aroma from hops and malt, fermenting the same wort with the same hops in triplicate next to their unmodified parent strains, with the thiols 3MH and 3MHA measured by an outside lab and all raw data published.
The briefCredit · unclaimed
Open 01 · the brief
Open fermentation curves
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A public dataset of fermentation logs (gravity, temperature and, where measured, pressure or CO2 flow, every few minutes) from many licensed breweries, with the yeast strain, starting gravity and the hour the brewer called fermentation finished; and a baseline model that predicts that hour early in the fermentation.
Why it is still open
Published models are trained on data nobody else can see. A 2026 design study for data-driven micro-breweries says its data "cannot be made publicly accessible for legal reasons", and the Sheffield Hallam study offers only its pilot-scale data, on request. A Newcastle University student brewery logged a Tilt hydrometer on 400-liter batches and trained a small neural network to predict fermentation time; it was off by 9.8 hours on validation batches but did poorly on unseen ones, which the authors put down to training sets that "may be too small or batches too varied". Searches of Zenodo and GitHub found bridges and dashboards for wireless hydrometers but no shared multi-brewery logs.
Built with what exists
Wireless hydrometers already common in small breweries, such as the Tilt, the iSpindel and the RAPT Pill, with open-source bridges on GitHub (TiltBridge, SPINDELMATE); Python; a time-series model such as gradient boosting or an LSTM; Zenodo to publish under an open license.
First step
Ask three breweries that already log with wireless hydrometers to export one month of data, and write a one-page data-sharing agreement that strips names.
Done looks like
At least 300 fermentations from at least five licensed breweries published, and a model that, for breweries left out of training, predicts the finish within 12 hours for 80% of batches by the halfway point, with its error compared against each brewery's own rule for calling the end.
Open 02 · the brief
Thiol yeasts tested together
The first to publish is credited on this line, and the deliverable moves from Open to Achieved.
What to build
A side-by-side trial of the commercial thiol-releasing yeasts, the engineered strains sold to free tropical-fruit aroma from hops and malt, fermenting the same wort with the same hops in triplicate next to their unmodified parent strains, with the thiols 3MH and 3MHA measured by an outside lab and all raw data published.
Why it is still open
The figures brewers see come from the sellers. Omega Yeast says its strains reach about 600 parts per trillion of 3MH with a boosted yeast gene and over 6,000 with an added bacterial gene, against about 60 for a normal strain; Berkeley Yeast says its Tropics strain, launched in 2020, was the first commercial strain with high thiol release. The one peer-reviewed test found, Oregon State's 2022 study, used Berkeley Yeast strains and had Berkeley Yeast staff as co-authors. Brewers' own accounts in trade magazines compare strains by taste only. A search found no independent comparison across sellers.
Built with what exists
A licensed brewery's pilot system or a university brewing lab, pitch-ready packs of each strain from the sellers, a contract lab that measures thiols at parts-per-trillion levels, and a shared spreadsheet for the results.
First step
Get price quotes from two labs that measure 3MH and 3MHA, and ask a licensed brewery with a pilot system to split one batch of wort between two strains.
Done looks like
At least five thiol-releasing strains and their parents, each fermented in three separate vessels from one wort, with 3MH and 3MHA in parts per trillion, attenuation and fermentation time reported for every vessel, and each seller's stated figure set beside the measured one.