A rainfall number is a surprisingly poor answer to a food question. Has it rained less? Perhaps. Will there be less corn? That depends on the soil, the crop’s condition, the timing, and what happened before the rain stopped. Gro Intelligence built a business around the distance between those two questions. Then, in 2024, the company itself ran short of something no satellite could supply: money.
- The invention: agricultural and climate data organized so analysts could compare it and build forecasts.
- The buyers: food businesses, financial institutions, agricultural companies and governments.
- The outcome: operations ended in May 2024; Almanac acquired the assets later that year.
01 / The trader who wanted the whole picture
Sara Menker came to agriculture through commodities trading at Morgan Stanley. She had grown up in Ethiopia, where shortages made the fragility of ordinary life visible. Finance supplied a professional vocabulary for that fragility: exposure, supply, downside risk. Gro, founded in 2014, was an attempt to make the underlying system easier to see.
The problem was not a world without information. Ministries published statistics. Satellites watched vegetation. Weather agencies measured precipitation. Trade organizations tracked shipments. The trouble was that each piece arrived speaking its own dialect. An analyst trying to connect them had to become a translator before becoming an analyst. A collection of numbers could be abundant and still be remarkably unhelpful.
02 / A common grammar for the harvest
Gro’s answer was an ontology, a systematic account of what things are and how they relate. A crop, a region, a measurement and its source could be connected within the same structure. In her 2017 fundraising announcement, Menker described turning scattered information into an organized whole. The machine learning depended on this less glamorous achievement. Before a model could find a relationship, the data had to be comparable.
That is where Gro fitted into the market. It sold analytical infrastructure between raw agricultural information and the people making decisions. A government report might answer one question in one country. Gro tried to let an analyst ask connected questions across countries and datasets without rebuilding the plumbing each time. Its expertise brought plant science, agriculture markets and geospatial science into the same room as engineering.
In a 2018 demonstration for Kansas State risk-management fellows, Menker searched decades of rainfall in India and produced charts and maps within seconds. The impressive part was the work the students did not have to do: find the files, reconcile them, and discover whether they meant the same thing.
03 / Put the forecast where the work happens
Gro offered a web platform, but it also provided an API and Python client for analysts building their own workflows. An Excel add-in let users find data series and pull them into spreadsheets. This was a sensible concession to how organizations operate. A procurement team’s habits do not vanish because someone has designed a handsome dashboard.
The DTN partnership made the practical use more specific. In 2023, Gro described a Supply Location Monitor combining predictive grain supply with DTN market and pricing information at more than 4,000 US facilities. For a grain buyer, location changes the question. The issue becomes the economics of buying, storing and handling grain near particular facilities, rather than an abstract national harvest forecast.
US grain facilities covered by the DTN-linked Supply Location Monitor, as described in 2023.
The commercial audience included Unilever, Yum! Brands, BNP Paribas and Wells Fargo, named by TIME in 2021. These were buyers with exposure to agricultural supply or financial risk. Gro’s advantage was the connection between physical conditions and business questions. Its alternative was often an internal analyst assembling public data, alongside specialist weather, market and climate-risk services.
04 / What the intelligence cost
Inc reported pricing starting at $1,000 a year in 2020, with publicly available satellite imagery helping keep costs down. That figure is a historical entry price, not a description of every enterprise contract. Gro sold access to its data and analytical capabilities; customized work could involve a different commercial proposition.
Investors financed the larger ambition. In January 2021, Gro raised $85 million in a Series B backed by Intel Capital, Africa Internet Ventures and family offices, among others. The announced plan was to expand adoption and improve machine learning. A large market supplied the argument: food security and climate risk touched governments, insurers, banks and food businesses alike.
“It’s about hedging for the downside risk.”
Sara Menker · TIME · 2021
The reach widened. Rockefeller Foundation support helped fund a public African food-security tracker in 2022. Gro and Kepos Capital launched the Carbon Barometer in 2023 to compare national carbon-policy incentives. Menker told TIME that climate data was finding uses beyond agricultural commodities. The same foundation could support more questions. Whether each question supported a repeatable sale was another matter.
05 / Cash failed before the idea disappeared
The visible breakdown was operational. By February 2024, Gro was struggling to make payroll, and James Cariello had replaced Menker as CEO. Emergency financing followed in March, alongside layoffs of approximately 60% of staff. By the end of May, the business was closing.
AgFunderNews quoted a former employee who described dependence on Unilever revenue and pursuit of projects resembling bespoke consultancy. That is an attributed diagnosis, not an audited account. Still, it identifies the commercial difficulty: a broad platform can keep generating interesting answers while sales remain concentrated and expensive to repeat. Financing could postpone that difficulty; it could not settle it.
06 / A smaller question for the next owner
In November 2024, Almanac, formerly Semios, acquired Gro’s assets through an auction. Its CEO, Sumer Johal, described applying the datasets and models to customers Almanac already served. Public information would complement private farm data. The acquisition preserved tools without restoring the old company.
The lesson a builder can copy is to connect evidence to a recurring decision inside an existing workflow. Start with a grain purchase, a supply plan, a risk review. This approach depends on comparable data, a buyer able to act, and a reason to return. It falters when observations are inadequate, implementation costs swamp the benefit, or every customer requires a new consulting project. Gro made the harvest more legible. Its successor’s task was to make that legibility easier to buy.