A new prototype of a small, solar-powered robotic vessel recently started sailing in the Pacific Ocean, pulling an underwater rack filled with seaweed. The startup developing the technology, called Phykos, says each platform holding the fast-growing kelp may be able to capture as much CO2 as 250 trees—and though the approach still needs to be proven, the company thinks that it could be a viable way to quickly sequester carbon by sinking the seaweed to the ocean floor.
The startup’s founders met at X, Google’s moonshot factory, before deciding to launch their own company. They started working on the technology after exploring how they might best be able to help address climate change. “We looked at the breadth of solutions out there, and then were really taking a first principles approach, and filtering down from that,” says Nico Julian, who cofounded Phykos with Jeff Zerger. “What can scale to a meaningful scale? What can ramp to that scale quickly, because we don’t have much time? And what can grow its own legs financially, and be a business?”
Seaweed along coastlines already captures an estimated 173 million metric tons of CO2 each year as it grows; some of that seaweed eventually sinks, trapping the carbon at the bottom of the ocean. Phykos wants to replicate the same process in the open ocean, where kelp doesn’t grow, to vastly increase seaweed’s global level of carbon sequestration. “Seaweeds have evolved to grow crazy fast and are fantastic at drawing out CO2,” Julian says. “Essentially what we’re just doing is giving them a bigger surface area, out in the open ocean, to do their same magic.” It could be a meaningful part of a larger carbon removal industry, he argues, which some experts have calculated will have to be as large as the oil and gas industry by the middle of the century.
The tech is modular: with the units that float on the surface, each the size of a small boat, and the lines of kelp underneath roughly the size of a single-family house. After seaweed “starts” from nurseries are planted on the lines, the vessels will navigate out to the open ocean. Software on each vessel is designed to steer toward the best areas for growth, moving throughout the year, and to automatically avoid areas like shipping lanes. Then it will harvest itself. “The seaweeds will grow and periodically get a haircut, so to speak, with an integrated harvest clipper mechanism,” says Julian. Unlike some types of kelp that float—picture the seaweed along California coastlines, which has small, round air-filled pockets to keep it near the surface—the company plans to work with species that naturally sink. A scale built into the platform will weigh the seaweed after each harvest to help calculate how much carbon has been captured.
Seaweed farming near the shore isn’t new, but the startup’s approach is different. “It’s a really challenging robotics problem,” says Marius Wiggert, a PhD researcher at the University of California, Berkeley, who is one of a team of researchers collaborating with the company. The device has to be able to navigate through unpredictable waves and wind and to operate on its own for multiple years. Because it’s becoming harder for seaweed to grow as the ocean warms, it will have to be able to navigate to cooler areas. (The system may be able to help make up some of the gap in sequestration that has occurred as naturally-grown seaweed has been declining.) The vessels can also navigate to particular areas to deposit the seaweed underwater.
The company will also have to prove that the seaweed can grow as expected, and its growing—and sinking—won’t have any unintended effects. “When we think about macroalgae, in particular, there are a bunch of open questions on the fundamental science side,” says Freya Chay, who works on the policy team at CarbonPlan, an organization focused on the transparency and scientific integrity of various approaches to carbon removal, from reforestation to regenerative agriculture. “Things like, how much can we grow, and how much of it actually sinks, and what eats it on the way down? And what are the ecosystemic impacts going to be at an intervention of this scale? That’s really basic, fundamental science that is still up in the air.” The company will also have to deal with other new challenges, like how growing and releasing seaweed in the open ocean may be regulated.
Phykos, like another startup called Running Tide that is also working on autonomous seaweed farms, is working with academic experts to study how the process works and to design a system that can avoid impacts on marine animals like migrating whales. It believes that the carbon in the seaweed will be locked away for 1,000 years (this is how long scientists have determined it stays in seaweed that naturally sinks).
If it works as expected, that’s a more permanent way to offset carbon than relying on forests that are increasingly likely to burn. (While some argue that it’s better to use seaweed for food, pharmaceutical products, and other uses rather than sinking it, Julian says that he thinks both are needed—but seaweed won’t decarbonize the atmosphere unless it’s sequestered.) Phykos is now working with a carbon registry to verify how it’s counting up the benefits so that carbon credits can eventually be sold.
The company, which just completed a stint at the tech accelerator Y Combinator, is still developing the technology. But they expect that the cost will be “probably on the lower end of the cost spectrum,” compared to other carbon removal technology, Julian says. Already, other companies have expressed that they’re interested in buying carbon offsets when they’re available. That’s partly because the system could have multiple benefits. The ocean has already absorbed most of humanity’s CO2 emissions, making the water more acidic, and it’s also quickly heating up. As the seaweed absorbs CO2 from seawater, it can also help reduce the problem of ocean acidity if it happens at a large scale. Seaweed can also help cool the water. “What’s exciting to us is that [seaweed] can have a certain restorative effect to the oceans,” Julian says.