Dutch AI firm Source.ag has purchased Optimal, “closing the loop” on autonomous greenhouse control
Source.ag, the Amsterdam-based AI specialist for greenhouse growers, has announced the acquisition of Optimal, a London-based startup that has built one of the first AI systems to operate a commercial greenhouse autonomously.
“Modern greenhouses are critical for supplying affordable vegetables under extreme weather and water scarcity,” Source stated. “If food cannot be grown outdoors, it must be grown indoors. Another record summer of extreme heat, storms and volatile energy prices has sharpened the case. Today, greenhouses depend on a shrinking pool of skilled growers, which limits scalability. The missing piece has been AI that works in the messy physical world of agriculture. That is what Source and Optimal have built and now bring together.”
Optimal’s AI directly controls greenhouse heating, ventilation, screens, lighting, CO2 and irrigation via a grower’s existing hardware. Its technology utilises proprietary physics and machine learning models of the crop and the greenhouse to plan the growing environment days ahead, and re-optimising in real time as the weather changes.
In a commercial-scale greenhouse in the Netherlands, it delivered 13 per cent more yield with 27 per cent less energy than the benchmark Dutch grower, Source revealed. It has also run commercial operations in Canada.
Optimal’s founder and chief executive officer Dave Hunter joins Source as VP autonomous control, along with Optimal’s engineering team. Hunter is a former head of quantitative trading at Deutsche Bank and was a member of Oxford’s Machine Learning group.
”In 2016, DeepMind used similar algorithms to ones he was researching to cut the energy Google’s data centres spend on cooling by 40 per cent,” Source stated. ”If AI could run a data centre, he reasoned, it could run a greenhouse. He founded Optimal that year through Entrepreneurs First. Its team comes from Oxford, Imperial College, Airbus and Red Bull Racing. Backers include Founders Fund, Backed VC, 2050 and Acequia Capital, plus angel investors Charlie Songhurst (Meta board member), Ed Grefenstette of Google DeepMind and Wendy Tan White of Intrinsic, the Alphabet robotics company.”
Source’s software takes care of monitoring the crop, forecasting the harvest weeks ahead and automating irrigation, while Optimal adds control of the climate itself. Together they form a “closed loop”, according to Source: one system that observes, predicts and acts.
Source’s software currently runs in more than 500 greenhouses in 25 countries, supporting €2.5bn of tomato, pepper and cucumber production a year and working with 13 of the world’s 25 largest high-tech growers.
Source said it had raised €50mn from investors including Astanor Ventures, Acre Venture Partners, Enza Zaden and Harvest House.
“Everyone is talking about physical AI. With this deal we are putting our money where our mouth is,” said Rien Kamman, co-founder and chief executive officer of Source.ag. “Optimal’s AI has already improved yield and energy in real commercial greenhouses. Now it joins a platform that runs in more than 500 of them. We need many more greenhouses across the globe to secure access to healthy food for everyone. With the current technology stack, we do not have enough expert growers to run them all, but with AI we can solve this scalability problem.”
“A global shortage of skilled growers prevents new greenhouses from being built and leaves our food system exposed to extreme climate and energy events,” said Dave Hunter. “We have proven that applied AI can overcome this scarcity by running greenhouses autonomously, enabling growers to increase productivity and expand. By joining Source, we can now accelerate the deployment of this AI into greenhouses across the world and create a more resilient food system for the future.”
“As growers, we feel climate and labour pressure every season,” added Kees van Veen, chief executive officer of Agro Care, one of Europe’s largest greenhouse tomato producers and a Source customer. “We do not need more hype. We need tools that keep the climate stable, energy in check and quality high across hundreds of hectares. Autonomous control is the logical next step, and we intend to be among the first to use it.”
The first autonomous deployments at Source customers in Europe and North America are set to go live this winter. Source then intends to extend the technology to more crops, growing methods and regions. The overall goal is to enable greenhouses to be deployed anywhere in the world, Source said, allowing for affordable produce to be supplied year-round even in the face of extreme weather and energy volatility.
