Given satellites and environmental sensors, the world already has data about the world. But we still don't see much, and Fieldstone Bio believes that microbes can change this.

"They have evolved into perceived and responded to information. It's just all the calculations around us, always going on," Brandon Fields, co-founder and chief science officer of Bio, told TechCrunch. "How do we take this and actually manipulate it to benefit us?"

Fieldstone's technology comes from this problem. The startup was founded in 2023 after MIT left, and Professor Chris Voigt's lab has developed a way to turn microbes into sensors. Scientists programmed microorganisms to change colors, whether it is soil hidden in dirt or nutrients in landmines, and then figured out how to detect them.

"The key technology in Chris' lab is the idea, 'How do we really visualize these cells from a distance?'" Fields said.

Fieldstone Bio recently raised $5 million in seed funding with Ubiquity Ventures and LDV Capital, the company exclusively told TechCrunch. The startup has been testing its technology in the lab, and funds will allow it to test these microorganisms in the real world.

Each strain is tailored to sense specific compounds, such as nitrogen on farmland or TNT residues in landmines.

“We separate the microbes from the environment we want to perceive,” Field said. “We build the DNA parts of the sensors and then put them in these different sensors to see which sensors behave best and which last the longest.”

Once the microorganisms are ready, the live stone will be broadcast using drone. After the microbe has some time to sense the environment (based on the target, several hours to several days), the company will have another drone to take photos in the area.

These images are not the usual aerial photography seen on Google Maps. Instead, they use so-called hyperspectral cameras to divide it into visible and infrared light into up to 600 different colors. Since Fieldstone's microbes will reflect light at very specific wavelengths, it can train AI models to find these signals in the data torrent.

"That's where AI is powered, because we can start using this information to tease these really faint signals to produce a real hot map of microorganisms to perceive the environment," Fields said.

In addition to agricultural and national security applications, Fieldstone has programmed microorganisms to detect environmental pollutants such as arsenic, said CEO Patrick Stone.

“Instead of doing core soil samples every 100 feet and you have 100 feet of resolution, we can get an inch of resolution and really draw where they need to clean things up,” he said.

Genetically edited microbial sensors broadcast on the farm are bound to cause eyebrows in people who oppose genetic modifications. Fields said the company has maintained contact with the EPA to ensure the company complies with regulations.

Fields said over time he hopes the company's database will become large enough that the model can be trained to associate other signals in the environment with any data returned by the microbe. This will allow hyperspectral cameras to detect no need to spread engineering microorganisms without spreading arsenic contamination.

"Ultimately, you don't need to apply microorganisms at all," Fields said. "You have drones, aircraft and satellites now, and you're currently collecting information about chemical information around the world."