I had been following the project since early last year, after receiving roughly a dozen press releases from Deep Sky about all of the companies it was setting up partnerships with. But it was hard to believe the scope of the ambition until I saw it with my own eyes.
CarbonCapture Inc., one of the companies piloting its technology at Deep Sky, had originally planned to deploy in the U.S., but has since packed up and headed north. The Los Angeles-based startup recently shipped all the equipment for its first demonstration project from Arizona to the Deep Sky site on four flatbed trucks. On a crisp October day, under a bluebird sky, the company’s CEO Adrian Corless stood in front of the newly installed towering mass of metal fans and explained the move.
“Because of what’s been going on in the U.S. and the backing away from support of climate technology and carbon removal, we made a decision back in February that we were going to redirect our focus and effort to Canada,” he told an audience of Canadian officials who had come to see the tech up close.
“Eight weeks ago, this was just dirt,” Corless said. “Today, we’re actually going to bring the first of our modules to life.” Then he invited Danielle Smith, Alberta’s conservative Premier, to do the honors. She pointed her fingers like a pistol and yelled, “Hit it!”
Behind her, the fans started to whir.
Deep Sky is not like other companies working in direct air capture, or DAC. Whereas most startups are developing their own patented designs and then raising money to go out and build demonstrations, Deep Sky is solely a project developer. It buys DAC systems, operates them, and sells credits based on the amount of carbon it’s able to remove from the air and sequester underground. Other companies buy these credits to offset their own emissions.
In the spring of 2024, Damien Steel, Deep Sky’s then-CEO, explained the theory of the case to me. It takes a different set of skills to engineer the tech than to deploy it in the real world, he said, which requires procuring energy to run the system and developing storage sites for the captured CO2. “There’s a reason why renewable developers don’t build their own windmills and solar panels,” he told me.
DAC technology is nowhere near as advanced as solar panels or wind turbines. Removing carbon dioxide from the air, where it makes up just 0.04% of the total volume, is currently far too energy-intensive to be commercially viable. There are more than 100 companies around the world trying to crack it.
Deep Sky’s first ambition was to buy a bunch of prototypes, test them next to each other, and figure out which were the most promising. Steel told me he was in the process of acquiring 10 unique DAC systems to install at a “commercialization and innovation center” known as Deep Sky Labs.
Deep SkyImage: Emily Pontecorvo
By the end of that summer, the company had signed a lease for the site in Alberta. Less than a year later, this past June, it had completed initial construction and was ready to begin hooking up DAC systems. In August, it announced that it had successfully injected its first captured carbon into an underground storage well. I had never seen one DAC project in the real world, let alone five. The company suggested I come for a tour during CarbonCapture’s launch event in late October.
By then Steel, who joined Deep Sky after more than a decade in venture capital, had stepped down from the CEO role “for personal reasons,” he wrote in a LinkedIn post, though he stayed on as an advisor. My guide would be his successor, former Chief Operating Officer Alex Petre.
Deep Sky Labs, now called Deep Sky Alpha, is in Innisfail, a town of about 8,000 people surrounded by farmland and prairie. To get there, I flew to Calgary and drove 75 miles north on Highway 2, the primary throughway that connects to Edmonton. Innisfail is dense and suburban-looking, with an industrial corridor on the western edge of town. Deep Sky was on its outermost edge, on the site of a former sewage lagoon the town had recently reclaimed, and sat catty corner to a welding and manufacturing company, which, as I was later told — multiple times — was developing hydrogen-powered locomotives.
A bright white cylindrical building about the size of an airplane hangar, emblazoned with “Deep Sky” in big black letters, was visible from half a mile away. As I pulled up to the site, workers in neon vests and hard hats were scurrying among outcroppings of pipes and metal structures. Unsure of where to enter, I parked on the road and wandered up to some trailers outside the perimeter. Petre poked her head out of one and beckoned me inside an office, where she fitted me with my own vest and hard hat so I could get a closer look.
“This is the only place in the world where we are putting together different direct air capture technologies side by side,” she told me, as we passed through a gate and began walking the grounds. Other than the sound of trucks and excavators driving around, it was fairly quiet. None of the DAC units were operating that day — one was down for maintenance, one for the winter, and the rest were still under construction.
The first stop on the tour was a modest black shipping container labeled SkyRenu, a DAC company based in Quebec. It was the smallest system there, designed to capture just 50 tons of carbon per year — roughly the annual emissions from a dozen cars. Directly across from it, workers appeared to be fitting some pipe on a much larger and more complicated structure resembling Paris’ Pompidou Center. This was United Kingdom-based AirHive’s system, which would have the capacity to capture about 1,000 tons per year once completed.
AirhiveImage: Emily Pontecorvo
DAC systems are feats of chemistry and mechanical engineering. At their core is a special material called a sorbent, a liquid or solid designed to attract carbon dioxide molecules like a magnet. The process is generally as follows:. First, the sorbent is exposed to the air, often with the help of fans. Once saturated with carbon, the sorbent is heated or zapped with electricity to pry loose the CO2. The resulting pure CO2 gas then gets piped to a processing facility, where it’s prepared for its ultimate destination, whether that’s a product like cement or fuel or, in the case of Deep Sky, a deep underground rock formation where it will be stored permanently.
Deep Sky’s aim was to trial as many iterations of the tech as it could at Alpha, Petre told me. That’s because what works best in Alberta’s climate won’t necessarily be optimal in Quebec or British Columbia, let alone hotter, more humid zones. “When the feedstock, which is ambient air, ends up being so different, we need multiple different technologies to work,” she said.
Case in point: A DAC system designed by Mission Zero, another U.K company, was offline the day I visited — and would remain so until next spring. It utilized a liquid sorbent and had to be drained so that the sorbent wouldn’t freeze when temperatures dropped below freezing overnight. The challenge wasn’t entirely unique to Mission Zero, however. “Everyone is struggling with winter,” Petre told me.
Deep Sky CEO Alex Petre standing in front of Mission Zero’s installation.Image: Emily Pontecorvo
Alpha is piloting systems with liquid sorbents and solid sorbents, variations on the chemistry within each of those, and systems that use different processes to release the carbon after the fact. The development cost ran to “over $50 million” Canadian, Petre told me. The company raised about that amount in a Series A back in 2023. It also won a $40 million grant from Bill Gates’ venture capital firm Breakthrough Energy in December 2024, and this past June, the Province of Alberta awarded Deep Sky an additional $5 million from an emissions-reduction fund paid for by fees on the fossil fuel industry.
The company fully owns and operates almost all of the DAC units onsite, although it’s still working with the vendors to troubleshoot issues and sharing data with them to improve performance.
When it comes to Carbon Capture Inc., however, the arrangement is a bit different. Deep Sky has agreed to host the company’s tech, giving it access to power, water, and underground CO2 storage, but CarbonCapture will retain ownership and help with operations, and the two companies will share the proceeds from any revenue the unit generates.
Petre said the structure was mutually beneficial — Deep Sky gets to demonstrate its strengths as a full-service site developer, while CarbonCapture gets access to a plug-and-play spot to pilot its system in the real world. The U.S. company is also looking to expand in Canada. “There’s lots of potential collaboration down the line,” Petre said.
Before Trump arrived at the White House, CarbonCapture had been making aggressive plans to grow in the states. In the fall of 2022, before the company had even demonstrated its tech outside of a lab, it announced that it would build a project capable of removing 5 million tons of carbon per year in Wyoming by 2030. It later leased an 83,000-square-foot manufacturing facility in Arizona to produce the equipment for the project.
At the time, the Biden administration was integrating carbon removal — of which DAC is just one variety — into its “whole-of-governement” climate strategy. The Department of Energy rebranded its Office of Fossil Energy to reflect a new focus on “carbon management,” a broad term that encompasses carbon captured at fossil fuel plants as well as from the atmosphere. In addition to overseeing the development of the DAC Hubs, the agency was running more than a dozen other grant programs and research initiatives mandated by Congress that were intended to help the nascent industry get established in the U.S. Biden’s 2022 climate law, the Inflation Reduction Act, also increased the tax credit available to DAC projects from $50 for every ton of carbon stored underground to $180.
As helpful as all of that may have been for the nascent industry, Canada was arguably going further. In 2022, the country finalized its own tax credit — an investment tax credit — that would cover 60% of the capital cost of building a direct air capture plant. The approach, while inspired by the U.S. subsidy, is geared more at de-risking project development than rewarding project success. The following year, the province of Alberta said it would offer an additional 12% investment tax credit on top of that.
Alberta was also becoming a leader in developing carbon storage infrastructure. Despite — or, more likely, because of — its oil-based economy, the province views carbon capture and storage as a “necessary pathway” that “will help Alberta transition to a low-carbon future.” Canada is the fourth largest producer of crude oil in the world, and the bulk of it comes from Alberta’s environmentally destructive tar sands.
CarbonCapture Inc.’s Project TamarackImage: Emily Pontecorvo
The government of Alberta owns most of the subsurface rights there, unlike in the U.S., where such rights are bestowed to landowners. That meant the province could simply offer companies leases to develop carbon injection wells. After two requests for proposals, the province selected 24 projects to “begin exploring how to safely develop carbon storage hubs.” A few of them, including Deep Sky’s storage partner — the Meadowbrook Hub Project north of Edmonton — are now operating.
Corless, of CarbonCapture, told me he spent a lot of time in Washington talking to the new staff at the DOE after Trump’s inauguration. It became increasingly clear to him that the DAC Hubs funding — and the general support for the sector enjoyed under the previous administration — would be going away.
By that point, the company had already planned to move its Wyoming venture to Louisiana after struggling to secure a grid connection at its original site. CarbonCapture had been awarded a DAC Hubs grant to conduct an engineering study for the project, but it received a notice from the DOE that the grant was canceled earlier this month. The company is still considering its options for how or whether to move forward.
On the same day the news leaked, CarbonCapture announced that it was shifting its plans to build a separate, 2,000 ton-per-year pilot plant from Arizona to Canada. Corless told me the company had originally planned to partner with a cement company to store the captured carbon in building materials, but Alberta offered more attractive commercial prospects. The company could more quickly access geologic carbon storage there, enabling it to sell carbon credits, which command a higher price than experiments in carbon-cured cement.
The timing of the announcement was pure coincidence. The poor prospects for an American DAC industry under Trump weren’t not a factor in the move, however. CarbonCapture wanted its pilot project to be a “springboard” for its first commercial plant, and Canada was attractive “given the favorable economic incentives, favorable regulatory environment, and the general positive interest in deploying DAC,” the company’s marketing director, Ethan Stackpole, told me in an email. “This is in contrast to the current atmosphere in the U.S.”
CarbonCapture signed a contract with DeepSky to host the pilot, dubbed Project Tamarack, in May, and set up a Canadian business entity called True North to build it. When I visited the site, the company was in the final stages of “commissioning” the unit, i.e. getting it ready to operate. The equipment had been manufactured at the company’s factory in Arizona, but it may end up being the only system produced there. The facility is now sitting idle.
Petre and I followed the tidy rows of wires and pipes that wound through Deep Sky Alpha, carrying electricity, water, and compressed air to each DAC system. A set of return pipes delivers the captured CO2 to Deep Sky’s central processing facility — the big white cylindrical building — where the company measures the output from each system before combining it all into a single stream. Inside, she showed me how the gas moved between large, tubular instruments that measure, dry, compress, and cool it into a liquid.
“Everything outside is first of a kind,” she said. “All of this equipment in here is fairly standard energy oil and gas equipment, it’s just arranged in a very different way.”
Sensors monitoring the wires and pipes enable Deep Sky to measure how much energy and water goes into producing a ton of CO2. Finally, trucks carry away the liquid CO2 to the Meadowbrook storage hub about two hours north, where an underground carbon sequestration well operated by a separate company called Bison Low Carbon Ventures provides it a permanent home.
While trucking the CO2 wasn’t ideal, the amount Deep Sky would capture at Alpha was so small that it made more sense to partner with Bison, which already had a permitted well, than to try to build one itself, Petre explained. When Deep Sky scales up at its next facility, which it expects to build in Manitoba, the company aspires to drill its own carbon sequestration wells on site.
Despite Alberta’s advantages for DAC, the location is not without drawbacks. The province had imposed a seven-month moratorium on renewable energy approvals from 2023 to 2024, which led to project cancellations and put development on ice. When the ban lifted, new regulations restricting wind and solar on agricultural land and near designated “pristine viewscapes” continued to make it difficult to build. Petre told me Deep Sky was one of only two companies in Alberta to secure a power purchase agreement with a solar farm last year.
“If I said, ‘I need 150 megawatts for my next facility right now,’ it would be a fairly difficult process,” she said. “There isn’t that much capacity online, and I would have to compete with data centers and a whole bunch of other folks who are also looking to come here and develop.” The company has started looking into building its own renewable energy supply on site, she said.
That anti-renewable sentiment stems from the region’s strong oil and gas identity. After my tour with Petre, I sat through a short program celebrating Project Tamarack’s launch, where Alberta’s Premier Danielle Smith conveyed her excitement by asserting that the province was “working to phase out emissions, not oil and gas production.” Alberta would double its energy production in the coming years, she said, while still reaching a goal of carbon neutrality by 2050.
Of all the extraordinary things I had seen and heard that day, this was the most brazen. The promise of direct air capture — the entire reason to expend time and energy and funds on plucking CO2 molecules out of the air — is that it’s one of the few ways to clean up the carbon that’s already in the atmosphere. Using it to offset continued oil and gas production might slow climate change, but there are a lot of other cheaper, more efficient, and more effective ways to reduce emissions — like switching to carbon-free power and electric cars.
I asked Corless about Smith’s comments later that day over coffee. Was it realistic to double oil production and go carbon neutral? He was coy. It would be very hard, he said. But it also depends on whether you’re talking about neutralizing the emissions from producing the oil versus from burning it. Corless seemed to view the argument as a political necessity, if a dubious one, to win government support for scaling DAC.
“I was hopeful that when the new administration came in, we could create an economic argument and tie what we’re doing to energy dominance and energy security,” he said, of the Trump administration. “It was just, I think, a bridge too far. Whereas here, that narrative is landing.”
Petre was more equivocal, responding that Deep Sky acknowledges that “we are not going to move away from oil and gas tomorrow,” and takes this as motivation to “get direct air capture to as low cost as possible and as easy to deploy as possible.”
In addition to the five DAC units currently installed at Alpha — SkyRenu, Airhive, CarbonCapture, Mission Zero, and a system from a German company called Phlair — Deep Sky has announced plans to bring two more units to the site from Skytree and GE Vernova. A few other deals are in the works but not yet public, Petre told me.
Even once Deep Sky Alpha has enough capacity installed to be printing carbon credits by the day, it won’t have proven that DAC is viable at scale. It’s not meant to. Many aspects of the facility are intentionally inefficient because of its nature as a testing ground.
“We had to do a lot of overspec-ing and oversizing of things,” Petre said. All the excess makes her optimistic about Deep Sky’s next project, however, where it will scale up a smaller number of systems to a much larger capacity. “If we can do something this complex, there’s a lot of room to simplify,” she said.
The Chevrolet Bolt.Bill Pugliano/Getty Images
Pumpkin bumper crops? Quoth the farmer, nevermore. INOAA
Deep SkyImage: Emily Pontecorvo
AirhiveImage: Emily Pontecorvo
Deep Sky CEO Alex Petre standing in front of Mission Zero’s installation.Image: Emily Pontecorvo
CarbonCapture Inc.’s Project TamarackImage: Emily Pontecorvo
The rare earths will flow, for now. Cheng Xin/Getty Images