Why this former Microsoft distinguished engineer decided to build weed-zapping laser-armed robots

By: Lisa Stiffler

When Georgios Chrysanthakopoulos decided to launch a startup building small, weed-controlling robots, he wanted an enviro-friendly solution for flattening the leafy pests. His solution is not only non-toxic — it’s just plain cool.

“Heck, I’m doing robots. I might as well stick lasers on them,” Chrysanthakopoulos said.

So that’s what he did. Chrysanthakopoulos is the 43-year-old founder and CEO of dCentralized Systems, a Seattle-based company that’s developing solar-powered robots that use precisely aimed lasers to lay waste to weeds while sparing crops and other desirable plants.

Chrysanthakopoulos spent nearly 16 years at Microsoft, followed by shorter gigs at Google, F5 Networks and the software company VMware. “I felt like I had peaked in the big company space,” he said. “I wanted to learn something new.”

While at Microsoft, Chrysanthakopoulos said that he worked on a super-secret robotics and neural-networks project, but the company eventually scrapped it.

“It was way ahead of its time,” he said. The work was hard, but exciting. “I wanted to get back to that. It kicked my butt.”

Chrysanthakopoulos’s concept for dCentralized is building small, non-polluting, low-cost robots for use on farms and properties that are 20 acres or less — the folks who can’t afford massive capital investments into big machinery. He’s aiming for robots that are roughly 2-feet square, cost $1,000 and can be packed flat and easily shipped. Customers will assemble the robots.

“For a day, they’re robotics engineers,” Chrysanthakopoulos said. “It’s like putting together your BBQ grill, but it’s autonomous and lives in your backyard.”

The robots need to be trained onsite to identify the specific weeds. One approach that dCentralized is testing requires users to place a blue plastic ring that the robot recognizes around a weed in the farm or yard. The robot can find the ring and record the image of the weed it contains, learning which plants to target. Then it uses its 5 watt laser to cut down the weeds (by comparison, a laser pointer is about 5 milliwatts). The robots are also designed to communicate with and track each other, so multiple units can be set loose on a larger property to tackle the weeds in a shorter time.

As a bonus, the robots also can act as security guards, recognizing moving creatures such as people, cats and coyotes.

Chrysanthakopoulos started his company in February and has partnered with Adlai Felser, who is working as the robotics lead. They’ve sold five robots to friends and acquaintances. They plan to deliver the products to them by July. Chrysanthakopoulos said his goal is to make 100 robots total over the next year. Most of the parts are from off-the-shelf. In addition to selling the robots, the company wants to offer seasonal services, charging about $30 an acre to mow down weeds.

There are numerous companies in the agricultural robotics space. Some are focused on planting crops and harvesting them, while others are also tackling weed control. That includes San Franciso-based FarmWise and the international AGCO Corporation’s Xaver robots. GeekWire also just profiled Smartsheet co-founder Brent Frei, who is working on a new agriculture robotics company called TerraClear that the software entrepreneur dubs the “Roomba for rock picking” since his mechanized device is designed to remove rocks from land used for growing crops.

Despite his technical background, Chrysanthakopoulos himself has run into his own roadblocks on the fundraising trail, in part because of the unconventional idea. He sees less experienced colleagues with software startups landing millions of dollars in investments. His response from VCs? Check back when you’ve got this figured out and are ready for later stage, Series B funding.

He’s given himself two years to make the project work and for now, he’ll keep bankrolling his laser-wielding, enviro-bots. Because even when it’s challenging, Chrysanthakopoulos is happy to experiment with robots and incorporate all manner of technology into a project designed to do good.

“It’s living the engineer’s dream,” he said.

We caught up with Chrysanthakopoulos for this Startup Spotlight, a regular GeekWire feature. Continue reading for his answers to our questionnaire.

One of dCentralized’s solar-powered, weed-controlling robots. (dCentralized Photo)

Explain what you do so our parents can understand it: “We build lightweight, affordable and cooperative robots that eliminate pollution and poisonous chemicals from our food supply and backyards.”

Inspiration hit us when: “We wanted to build autonomous systems that solve a real, pressing need in society. We read an article about herbicide use in farmland and the conflict it’s creating between farmers. Drift of the chemicals from one farm to those adjoining it causes loss of crop for those not using the chemicals. Several other factors motivated us to tackle this domain: the inherent risk of using a pesticide/herbicide in the food supply, the resistance of the weeds to it, and the agro giant control of the farmer through GMO seeds and pesticides. After more background research and looking into companies entering this space, we saw an opportunity for affordable, cooperative robots that are 100 percent chemical free. We use engraver-class lasers to zap weeds from unstructured or structured fields, and if available, a solar lens to burn the target plant.”

VC, Angel or Bootstrap: “Bootstrap. It’s important to move fast and control our destiny while we build credible technology and a solid business plan. We can do this through quick iteration, field trials and research of the vast and difficult domain we chose to compete in.”

Our ‘secret sauce’ is: “We have experience in software and vehicle design so the combination of both, plus a our willingness to iterate quickly, has already given us a viable vehicle and a couple of field trials, all within three months of starting the company. We’re also mission driven. While it might be considered a weakness and a constraint, we want want to avoid waste and pollution in the production process, so we are using easily recycle-able metal and compostable materials, including wood! This mission-driven approach and our commitment to not use any hazardous chemicals has led to some interesting inventions that differentiate us from our competitors. Our robots are small, lightweight, can run 24/7, and since they can cooperate, we can vary the number of machines used on a backyard or field, versus requiring a larger, more expensive machine.”

The smartest move we’ve made so far: “We iterate quickly, have stayed small and kept our focus on the small, cheap and useful approach. We have had several frustrating moments since creating a company from scratch. In a difficult domain, it was bound to be hard. We are still early on but I think it paid off to build and test the machines in the beautiful Eastern Washington landscape. It’s so much better than a desk job.”

The biggest mistake we’ve made so far: “The company was officially formed and the first employee hired in February of 2018, so we are very young. It is hard to tell what our biggest mistake is yet. It might be how we approach the market, or even if we should start with farming versus a consumer approach (gardening/yard maintenance). It is also likely we will need to make some further hard choices in the vehicle design to achieve our very aggressive target price. I hope we do not end up regretting what we eliminate in terms of capability.”