Breaking the agtech lock: A Groundbreaker’s call for open-source innovation
Agtech is booming, but much of that innovation is trapped behind patents and proprietary datasets. Guy Coleman, a 2024 evokeAG. Groundbreaker is challenging that model, with an OpenWeedLocator that uses open-source approaches to unlock faster, cheaper, DIY precision weed control for farmers.
Have an idea that challenges thinking in our sector? Applications for the evokeAG. 2026 Groundbreaker program are now open.
Should agricultural innovation be locked behind proprietary data, or can we achieve bigger, faster outcomes by sharing knowledge and tools? That’s the question Guy Coleman, alumni of the evokeAG. 2024 Future Young Leaders (now Groundbreaker) program has been asking as he builds momentum behind his open-source OpenWeedLocator – or OWL – project.
“I wanted to frame OWL as more than just a research project,” said Guy. “I wanted to challenge the evokeAG. audience to think differently about the way we do agricultural research.”
A big platform for a big idea
Initially set on a career in medicine, Guy pivoted into agtech, swapping anatomy labs for ag science at the University of Sydney – which led to a PhD in machine learning and computer vision for weed recognition.
“At that point I hadn’t done much coding, so I used open-source tools to learn,” explained Guy. “But there weren’t any tools to learn with that applied an agricultural lens. That’s where the interest started: how can I fill this gap in agriculture?”
Guy’s Groundbreaking question was whether open-source approaches (the sharing of datasets, code, and ideas) could shift agtech from being locked away inside tech companies to a more collaborative, open model.
“On the surface, it’s a geeky, boring idea to talk to an audience about,” laughed Guy. “But at its core, there’s something really impactful about open source: where the future of agtech doesn’t necessarily have us waiting for a tech company to market a solution.”
That ‘geeky idea’ became the foundation for OWL, a weed detection device that farmers could build themselves. “It’s designed to be practical, scalable, and accessible,” explained Guy. “The latest OWL prototype is built from simple components; just a Raspberry Pi (a credit card-sized computer) with a camera on the front. The computer is about $100; the camera’s about the same. Then you have a driver board that provides power and connects to between 4 and 16 solenoids on a sprayer,” explained Guy.
“It’s all off-the-shelf parts, and everything’s 3D-printable, and replaceable for a few hundred dollars.”
Denmark: The right fit
Guy’s Groundbreaker project made waves on the Southern Ocean coastline of Western Australia, with an Esperance farmer reaching out to get involved. “He was keen to start collecting data for me and field testing the device,” recalled Guy. “And for the past year he’s been using his SwarmFarm robot to collect hundreds of thousands of images to build an open source image dataset.”
Now based in Copenhagen, Guy is focused on scaling OWL, and pursuing post-doctoral research on weed ecology and ‘detection resistance’.
Denmark, he said, is the right fit for his open-source philosophy. “Europe has a strong culture of funding social good projects. That mindset makes it much easier to get support for open-source tools, because they can benefit everyone, not just one company.”
Major funders like Novo Nordisk (the makers of Ozempic) are backing his work, including efforts to scale OWL from prototype to a broadacre-ready system.
Rethinking agtech models
Guy believes Australia’s agtech sector could benefit from a shift towards open standards and shared datasets.
“If farmers have access to the underlying data and code, they can build their own solutions,” he explained. “They don’t have to wait for a tech company to develop a specific algorithm that can detect wild radish in chickpea. They can collect their own image data, annotate it, and train their own algorithm specific to their region.”
The concept flips traditional agtech business models on their head. “As opposed to typical proprietary IP through a closed-source company, the solution is open, simple, and everyone can do it,” explained Guy. “But because there’s no differentiating factor IP, it’s harder to go to a VC and say, ‘Can you give me a million bucks?’”
Still, there’s a clear commercial upside: it lowers barriers for companies entering new markets. “Danish spot-spray companies could ask for a weed image library they can use before launching in Australia,” said Guy. “Right now, they have to start from scratch – and that’s a barrier. Meanwhile, the USDA has invested tens of millions into large-scale open-source weed image libraries across multiple states. We risk falling behind.”
Open source as a farmer safety net
Risk reduction is a selling point for farmers, too. “If a startup goes bust, farmers aren’t left with a $200,000 ‘brick’,” said Guy. “The technology can keep evolving, even if the original company doesn’t. The farmers I’ve spoken to are all keen for this type of solution. No one wants to pay twice.”
So keen was a group of West Australian vegetable growers that they travelled all the way to Denmark to workshop the technology. “A grower had read about my Groundbreaker project in VegetablesWA magazine, got in touch about working together on the idea, and secured funding for 11 growers to do a study trip here to build and test the OWL.”
Guy’s growing network of collaborators reflects a core advantage of open-source innovation. “With public image datasets, you can leverage global research and development, instead of just one team” said Guy. “Copenhagen researchers are already interested in working on the Australian image data collected on that SwarmFarm robot – that would never have happened if the data wasn’t public or the effort made to collect it.”
Getting OWL ready for flight
Commercialisation is on the horizon, but Guy is eager to keep OWL true to its open-source roots.
“Through Groundbreakers I explored ideas like keeping the base ‘recipe’ open and selling ready-to-go kits on top of that,” said Guy. “That’s the pathway we’re looking at in Denmark.”
Australia is on the roadmap, too. And it feels fitting, given that’s where OWL first took flight. “Open source can share technical innovation through the whole supply chain and across country borders, not just with people who have the money to buy solutions,” concluded Guy. “That’s the big benefit for agriculture. It’s not about picking winners; it’s about everyone having the tools to innovate.”
Applications for the 2026 Groundbreaker program close on 3 August 2025. If you’re aged 18-30 and working on an idea to improve agriculture or rural communities – no matter how early stage – the Groundbreaker program can help you grow it.
For more information and to apply, click here.
evokeᴬᴳ⋅ 2026 will be held on Tuesday 17 February and Wednesday 18 February 2026 at the Melbourne Convention and Exhibition Centre. evokeᴬᴳ⋅ is powered by AgriFutures Australia and funded by the Australian Government, Platinum Partner Elders and Host State Partner, Agriculture Victoria.
Pre-sale tickets are now on sale at evokeag.com