Skip to main content

Can AI perfect the IPA?

Data-driven brews. Climate-resilient hops. This is the future of beer.

By Tony Rehagen

Beer is about as low-tech as it gets. The oldest surviving beer recipe was unearthed in Mesopotamia, from 1800 B.C., when the Sumerians made a divine drink from fermented barley bread soaked in water with yeast. About 3,000 years later, Benedictine monks in a Bavarian abbey outside of Munich recorded the addition of hops as a bittering agent and preservative. From that day on, those four ingredients — water, fermentable starch (such as barley malt), yeast, and hops — have been the base for just about every beer ever quaffed by humankind.

Those Benedictine brothers could time-warp into a 21st-century brewery and figure out how to whip up a passable batch of ale. Despite the craft beer revolution’s explosion in flavors and additives (everything from doughnuts to cannabis to edible glitter), brewers have painted with the same palette of styles since the 1400s.

But today, as nearly 10,000 U.S. craft breweries jockey for attention and market share, they’re producing more technological innovations in beer-making than the industry has seen in millennia. An innovation race in brewing hardware, ingredients, and recipes aims to help smaller companies brew the same quality beer, batch after batch.

In Australia, brewers are using artificial intelligence to perfect an IPA with feedback collected through QR codes. Meanwhile, beer makers in the U.S. Pacific Northwest and beyond are using genetically modified hops to deal with shortages caused by climate change.

Such new technologies may help brewers survive and prosper in a competitive marketplace.

“Consistency is always king,” says Malcolm Purinton, a history professor at Northeastern University who specializes in the history of beer and brewing. “Especially now. Fifteen years ago, it didn’t matter if the beer was that good — people were excited to go local and experiment. But now, when you see companies buying up craft breweries and the fight over shelf space and tap handles, you have to be good.”

An artificially intelligent beer

Before you strive for consistency, you need to perfect your formula. Why lock in a recipe for a beer that no one likes? This is among the biggest obstacles facing craft brewers. As Purinton points out, macro breweries like Anheuser-Busch, Miller, and Coors have lager formulas calibrated from more than a century’s worth of consumer feedback. That’s why a fresh Budweiser tastes the same today as it did in 1975.

Big brewers have used AI to make beer before: In 2017, Carlsberg teamed with Microsoft for its “beer fingerprinting project,” which sought to detect flavors and aromas while creating new beers and ensuring air-tight quality control.

Until recently, small craft brewers have had no such access to that volume of market research and data. They work from personal taste, scant online reviews, and word of mouth to determine if their IPA is too hoppy or their kettle sour too tart.

An Australian platform is trying to change that. It’s using AI to help better connect mom-and-pop brew masters to their target audience and fine-tune and perfect recipes.

“We can say, hand on heart, that every review is going into a neural network that will affect the flavor of their beer.”

Denham D’Silva, co-founder of Australian company Deep Liquid

In 2021, Deep Liquid, an Adelaide-based company that partners with the Australian Institute for Machine Learning, helped nearby Barossa Valley Brewing create AI2PA: The Rodney, an AI-generated IPA. On each can of AI2PA, a QR code allows drinkers to submit their thoughts on the beer’s flavor, aroma, and mouthfeel. That real-time feedback goes straight into a data set that is then plugged into an algorithm that can adjust the recipe accordingly.

“Since the biggest companies collect the most data, they’re the ones who’ve been able to harness AI,” says Deep Liquid co-founder Denham D’Silva. “We’ve flipped this and put the technology in the hands of the consumer, giving them direct contact that enables smaller brewers to leverage AI.”

Purinton says making machine learning more accessible to smaller brewers will help them refine recipes, maintain quality control, and know exactly how much of each ingredient to use, limiting waste and saving money.

D’Silva sees an added bonus to Deep Liquid’s model for craft brewers: Increased consumer engagement. “People jump on these beer review sites and then realize they’re the only ones reading these reviews, so why bother?” says D’Silva. “But there is research that suggests that when customers believe their reviews are affecting change, they’re more engaged. We can say, hand on heart, that every review is going into a neural network that will affect the flavor of their beer.”

The future of IPA is GMO

Innovation is coming to beer’s ingredients, too. While water, barley, yeast, and hops have been on virtually every beer mash bill since the 8th century, the way brewers acquire and use these ingredients has shifted drastically in recent years. And that’s due largely to climate change.

Brewers have had to find creative alternatives for their foundational ingredients. Water shortages due to drought are hurting brewers, especially in the U.S. West. So San Francisco’s Devil’s Canyon Brewing Co. recently released a beer made entirely from the city’s new greywater recycling system, which repurposes water from sinks, showers, and washing machines. Shifting weather patterns can imperil America’s barley crops, but beer makers have a wide choice of fermentable alternatives, including wheat, rice, corn, oats, rye, and sorghum.

Hops, however, pose the biggest challenge: 96% of hops in the U.S. are grown in the Pacific Northwest, where drought and wildfires have ravaged the land and made year-to-year yields of the bittering flower unreliable.

So how can technology protect brewers’ access to the hoppy tastes and aromas in their wildly popular IPAs? The answer might be in the fourth ingredient: Yeast.

Brewers are teaming with bioengineers to explore the role of genetically modified yeast strains that can unlock more intense flavoring agents within hops during fermentation. This means that more of the tropical and fruity essences can be extracted from each flower, achieving the same big hop flavor with a much smaller load of hops.

For example, researchers at Berkeley Yeast, in collaboration with Oregon State University, modified a brewers’ yeast strain to boost the amount of two essential oils within hops, called thiols, which provide tropical flavors like guava and passion fruit. The result is a powerfully fruity beer made with a fraction of the ingredients.

Eventually, the technology could save money, energy, and resources. But most brewers, adventuresome chemists and cooks at heart, are more excited about what flavors these genetically modified yeasts might one day unleash.

“Savings on hops can be beneficial for big brewers,” says Scott Janish, co-founder of Maryland’s Sapwood Cellars Brewery and author of The New IPA: Scientific Guide to Hop Aroma and Flavor. “But a majority of brewers are doing it to create beers that are unique and have that ‘wow’ factor.”

Janish sees a possible future when genetically modified yeasts could give brewers even more control over their craft. “I think the technology will become crisper,” he says. “You’ll see more yeast strains created for specific purposes, some that create banana flavor, maybe more Fruity Pebble esters [flavor compounds that produce fruity aromas in beer], or glycerol with more mouthfeel. But we’re completely starting over — there’s a lot of trial and error.”

Climate change, supply chain snafus, and the constant pursuit of profit and market share will no doubt continue to drive brewers to delve further into new technologies. But Purinton says that even though brew masters are technically chemists, they don’t get into brewing to geek out on technology.

“People don’t become brewers because they wanted to use their software engineering degree,” he says. “And people aren’t going to visit a brewery that is just a wall of computer screens.”

Besides, who would explain that to the time-traveling Benedictine monks?

Published on

Tony Rehagen is a writer based in St. Louis. His work has appeared in Popular Mechanics, Bloomberg Businessweek, The Washington Post, The Boston Globe, and Pacific Standard.

Illustration by João Fazenda.


With help from AI, a Holocaust survivor’s story lives on

How 116 cameras and natural language processing turn museums into conversations.

By Tony Rehagen