On May 31, the Swiss company Climeworks will turn on the first commercial plant to suck carbon dioxide out of the air and feed it to vegetables in a neighboring greenhouse. Located in the tiny, agricultural municipality of Hinwil, Switzerland, the plant stands 40 feet tall and looks like a rectangular wall of oversized dryers stacked three-high.

It will be the first business to sell carbon dioxide drawn right out of its surroundings, using a technology called direct air capture. To date, carbon capture technologies have been restricted to areas where there are high concentrations — like the smokestacks of coal-fired plants. But the promise of direct air capture is to grab the kind of ambient carbon emitted by cars, aircraft, and trains.

Since it’s not happening at the source, direct air capture must be able to deal with lower concentrations (0.04 percent of air), making the technology inherently more difficult, and expensive. That’s why until now, it has been demonstrated only in small, experimental pilot plants.

What Climeworks is hoping to show in Hinwil isn’t just that the technology works but that captured carbon can then be resold as plant fertilizer, fuel, or even for carbonated beverages.

But climate scientists say so-called negative-emissions technologies — which remove pollutants from air — will be crucial to meet the goal of the Paris Agreement, the landmark 2015 deal co-signed by 196 nations, to hold the increase in global temperature to no more than 2 degrees Celsius.

“Climeworks is the first to scale up to substantive level,” said Julio Friedmann, a former principal deputy assistant for fossil energy for the U.S. Department of Energy and senior adviser at the Lawrence Livermore National Lab. “There’s almost no way to hit those targets without using negative emissions, and in some cases, quite soon.”

The plant is projected to capture 900 metric tons of the greenhouse gas, or about the emissions from 200 cars a year. It traps ambient carbon dioxide with absorbent filters inside the plant’s air collector. To release the carbon dioxide, the filters are heated to 212 °F with waste heat from a neighboring partner waste incinerator plant owned by the company Kezo. The freed carbon dioxide is then pumped over to the greenhouse operated by Gebrüder Meier to “enhance the growth of vegetables and lettuce by up to 20 percent,” according to a press release.

One study estimates that to avoid the two-degree global temperature rise by century’s end, negative emission technologies must scale up to capturing 5 billion tons of CO2 annually by 2050. That’s approximately twice the amount absorbed by all the planet’s oceans. And it’s a hefty ask for these fledgling technologies.

“If we want to do this by midcentury, we need to not only start developing these technologies but implementing and scaling up these technologies. That’s what we see as our main role,” said Jan Wurzbacher, who co-founded Climeworks in 2009 with Christoph Gebald at the University ETH Zurich. Wurzbacher and Gebald developed their process in collaboration with the Swiss Federal Laboratories for Materials Science and Technology; the startup has received more than $7 million in venture capital funding from Venture Kick, Gebert Ruf Stiftung, de Vigier Stiftung and ClimateKIC, private investors, and also grants from the Swiss government for various projects, including their new plant.

The company has an ambitious goal of capturing 1 percent of the global emissions of CO2 by 2025. That would require 750,000 of their modular plants in operation, Wurzbacher said. To get there, they need to cut costs by a quarter to a third and land a lot more customers, he said.

The plant cost somewhere in the single-digit millions to build, according to a communications representative, who declined to give an exact number. Climeworks says it’s selling the captured CO2 to the greenhouse around market price, which won’t cover the price of the plant because it includes research and development costs. The company estimates its next plant will cost about $2 million.

A 2015 U.S. National Academies report estimated that direct air capture costs somewhere from $300 to $1,000 per ton of CO2 captured, a wide range due to the lack of public, real-world data.

Noah Deich, the executive director of the nonprofit Center for Carbon Removal, likened current direct air capture companies to the renewable energy industry in the 1980s, in that focusing on small-scale projects would allow them to develop the technology and bring down the costs.

“People have started to build these units in ways that are able to take advantage of existing markets. If you tried to [sell] captured CO2 from a coal plant, you’d quickly overwhelm all the greenhouse need; there’s not a market at that scale. Climeworks has been able to find a solution for the problem that’s fit for today,” Deich said.

A major potential customer for the carbon-capture companies is the food-and-beverage industry, a big consumer of CO2 for things like carbonated drinks and dry-ice packing. A handful of other companies have proposed turning captured carbon into products such as fuels, building materials, and plastics, though they haven’t reached a commercial scale.

One of the leading competitors is Carbon Engineering. Based in Canada, the company aims to commercialize captured air-to-fuel technology to produce low-carbon fuels. Having demonstrated that their liquid solution technology can handle one ton of CO2 per day, the company has partnered with Greyrock Energy to start testing a process later this year to convert water and CO2 to fuel.

The fuel is estimated to cost about $3.79 a gallon, a price that’s “within striking distance,” according to Carbon Engineering’s business development manager, Geoff Holmes. That price isn’t competitive with fossil fuels, but Holmes cites programs like California’s low carbon fuel standard that has incentives or credits for low-carbon intensity fuels that could level the playing field in the future. Climeworks is also working on air-to-fuel projects and supplies some of the CO2 used to make synthetic diesel in a Dresden pilot plant for the carmaker Audi.

Business models for captured carbon aren’t well developed yet, noted Adrian Corless, CEO of Carbon Engineering. Ideally, one day it may be possible to have a business simply doing carbon removal, he said. “But I think we’re still a decade or more from that scenario.”

Startups aren’t the only ones responding to the call for carbon removal. Last week, the U.K. announced a $10 million greenhouse gas removal program to fund negative-emissions research.

It’s only been in the last couple of years that studies have shown that direct air capture will be absolutely necessary to meet the Paris Agreement goals, said Dr. Jennifer Wilcox, a professor at the Colorado School of Mines. She recently published a study on the most useful applications for captured carbon and has another accepted paper recommending locations for air capture plants in the U.S. “If we’re going to have to do direct air capture, let’s at least do it responsibly,” she said.

Though direct air capture shows some promise, industry and climate experts alike emphasized that reducing emissions remains a major hurdle that should be prioritized ahead of these technologies.

It’s too early to tell if carbon removal will be ready in time to make an impact on climate change. But, as Friedmann puts it: “Their chances are entirely a function of what the world decides is important.”