Naware is betting that the future of weed control doesn’t have to rely on chemical sprays. The startup has built a system that uses computer vision to identify weeds and then eliminates them with steam—vaporized water—rather than herbicides.
From lasers and drones to a safer idea
Founder Mark Boysen didn’t start with steam. In the early days, he experimented with using drones paired with a 200-watt laser to kill weeds. The approach didn’t last long. Boysen said the fire risk was simply too high, making the concept impractical for real-world conditions.
The motivation behind a chemical-free method was personal as well as practical. Boysen has described thinking about how his family in North Dakota lost three members to cancer, and how they suspected the illnesses could be connected to chemicals in the groundwater. Against that backdrop, building a non-chemical way to manage weeds became more than a product idea—it became a mission.
After abandoning the laser route, Boysen went through extensive prototyping and experimentation, including exploring cryogenics. Ultimately, the solution he landed on—and demonstrated earlier this year at TechCrunch Disrupt 2025—was steam.
How Naware’s steam-based weed control works
Naware’s system is designed to “see” weeds mixed into grass and other turf, then destroy them using heat from vaporized water. In practice, that means the company is combining two components:
- Computer vision that detects weeds among surrounding plants in real time
- Steam delivery hardware that targets identified weeds and kills them without chemical herbicides
The concept is aimed at large turf and land-management environments where weeds are costly and persistent—lawns, athletic fields, golf courses, and potentially broader agricultural settings.
Boysen has said the equipment can be mounted or attached to machines that are already commonly used for grounds work, including mowers, tractors, and ATVs. The idea is to fit weed control into existing maintenance routines rather than forcing customers to adopt an entirely new workflow.
A “garage startup” approach to prototyping the steamer
While Naware is operating in an era dominated by “agentic AI” and massive software valuations, Boysen has framed the company as something more old-school: a classic garage-style startup with hands-on tinkering.
Early testing reportedly began with a low-cost consumer garment steamer purchased on Amazon. The team then ordered seven more as they kept iterating. According to Boysen, those devices weren’t built for industrial use, which meant the company had to do significant research and development to make the approach effective, repeatable, and scalable.
That challenge—turning a simple steam concept into rugged, reliable equipment—remains a core part of the productization effort. For a tool like this to work across wide areas and across different weed pressures, the output can’t be inconsistent, and the system can’t be fragile.
The harder problem: identifying weeds in “green-on-green” conditions
In weed control, the hardware is only half the battle. Boysen has indicated that the tougher challenge has been teaching the software to reliably identify weeds when everything in view is green—what he called the “green-on-green” problem.
Object recognition is a well-established use of AI, but turf environments present an especially demanding version of it. A weed in a lawn isn’t a high-contrast target. It can be partially obscured, vary in shape and size, and appear under different lighting conditions. And for Naware’s approach to be practical, detection needs to happen in real time as the equipment moves across the ground.
To meet those compute demands, Boysen said the system uses Nvidia GPUs. That detail underscores the direction Naware is taking: bringing modern AI hardware and real-time perception techniques into a field that has historically relied on blanket chemical applications or labor-intensive spot treatments.
Who Naware is targeting—and the cost savings it claims
Naware is focusing on customers where chemical use can be extensive and expensive, and where the public-facing nature of the turf increases scrutiny around inputs. Boysen said the company is aiming at organizations that handle lawn care for athletic fields and golf courses.
In those environments, weed management often involves recurring chemical purchases, repeated applications, and staff time devoted specifically to spraying. Boysen has claimed that Naware can reduce costs substantially for such customers—saying the company can save them “anywhere from $100,000 to $250,000 on chemicals alone.”
He also argued that there is an additional labor component to the savings. If a customer can reduce or eliminate chemical application cycles, they may also reduce spending on personnel whose primary responsibility is spraying those products.
Why the turf market is a logical early wedge
Even without introducing new claims beyond what has been reported, it’s easy to see why athletic fields and golf courses are attractive early markets for a chemical-free weed killer:
- High maintenance standards: Weed pressure is often treated aggressively to maintain turf quality and appearance.
- Large surface areas: Costs add up quickly when chemicals are applied repeatedly across wide acreage.
- Operational routines: Many properties already use mowers and utility vehicles daily, creating a natural path for an attachable system.
Naware’s positioning suggests it wants to slot into those existing routines: detect weeds as the vehicle moves, apply steam immediately, and avoid storing, mixing, and spraying chemicals.
Paid pilots, early traction, and partnership talks
Boysen has said Naware is already running paid pilots as it tests the system in real environments and refines the product. These pilots serve two purposes: validating that the weed-killing performance holds up outside controlled settings, and helping the company dial in reliability, repeatability, and operational workflow.
At the same time, Boysen said the pitch is drawing attention from prospective partners. In particular, he described pursuing strategic partnerships and said Naware is in discussions with equipment manufacturers—specifically referencing “$5 billion companies” that are interested in the product. He did not name the businesses.
For a company building hardware designed to attach to mowers, tractors, and ATVs, partnerships with established manufacturers could matter as much as the underlying tech. Such relationships can accelerate distribution, reduce go-to-market friction, and help ensure the system integrates cleanly with equipment customers already trust and buy.
What Naware says it needs to win: partnerships, patents, and funding
Boysen has outlined three ingredients he believes will determine whether Naware succeeds:
- Strategic partnerships with major equipment companies
- Patents to protect the approach and product differentiation
- Funding sufficient to scale fast and outpace would-be competitors
So far, Naware has been bootstrapped, according to Boysen. But he has said the company plans to open its first fundraising round in the coming months.
His stated goal for that raise is ambitious: he wants a round that is strong enough to discourage others from trying to replicate the concept. At the same time, he has emphasized that the core promise—killing weeds effectively—has to be delivered consistently for the company to earn its place in the market.
The bigger picture: AI-guided weed control without chemicals
Naware’s approach sits at the intersection of two broader shifts. The first is the rise of computer vision moving from digital-only products into physical-world applications, where real-time perception can guide actions like targeting, spraying, cutting, or—in this case—steaming. The second is growing interest in reducing chemical use, whether for environmental reasons, health concerns, or simply cost control.
Weed management has historically been dominated by chemical herbicides because they scale: spray large areas quickly, repeat on schedule, and get predictable results. But the drawback is that chemical programs can be expensive, labor-intensive, and increasingly controversial in certain settings. A system that can target weeds precisely—especially if it can be bolted onto equipment crews already operate—could appeal to customers looking for a different tradeoff between cost, effectiveness, and operational complexity.
Conclusion
Naware is attempting to replace chemical weed control with an AI-guided steam system that identifies weeds in real time and eliminates them using vaporized water. With paid pilots underway, discussions with large equipment manufacturers, and plans to raise its first funding round, the startup is positioning itself to bring computer vision and heat-based weed killing into mainstream turf maintenance—starting with athletic fields and golf courses.
Related Articles
- Startups and VCs in 2026: What Investors Say Founders Must Prove, Where Capital Is Moving, and Whether IPOs Return
- Friendship Apps Are Booming: 222, BFF, Meetup, Meet5, and More Promise Easier Ways to Make New Friends
- Ÿnsect’s $600M insect-protein dream collapses as the French startup enters judicial liquidation
Based on reporting originally published by TechCrunch. See the sources section below.
