On the dark volcanic soil of Iceland, looming conspicuously is a nondescript plant that its creators hope may be the key to holding back global warming.
It's not much to look at – a series of fan-like units stacked one on top of each other, a large annex that wouldn't look out of place on a big Australian farm, and an assortment of pipes and steelwork.
But Mammoth, as it's been dubbed by Climeworks, the Swiss firm that operates the plant, is doing something that's remarkable in many respects.
It is sucking carbon dioxide – the greenhouse gas blamed for much of the heating taking place across the planet – directly out of the air in order to bury or store it somewhere or use it in some other way.
At the heart of the project, which was switched on just this week, is a process known as direct air capture.
Expect to hear the term much more in the years ahead as countries the world over cast around for ways to cut their pollution and clean up their economies.
What is direct air capture?
In layman's terms, direct air capture, or DAC, is the process of sucking in air through a giant vacuum and then capturing the carbon dioxide it contains.
The carbon dioxide is then stripped away using chemical filters that can snag any particles passing through the giant fans.
In the case of Mammoth, heat is applied to release the carbon once again, albeit as a stream of gas that is given to another firm for handling.
At this point, the gas is mixed with water and pumped underground, where the carbon dioxide reacts with minerals in Iceland's volcanic basalts to form calcium carbonate, a white crystal.
Effectively, carbon that was in the air is turned into rock – in this case a rock that is a key ingredient in limestone.
But it can also be used in other applications that already exist, such as adding the fizz to drinks or stimulating plant growth in greenhouses.
Thus, DAC is able to deliver what are known as "negative emissions", or the removal of greenhouse gases already in the air as opposed to the avoidance of yet more.
It's a technology, according to its boosters, that has a potential value measured in the trillions of dollars.
Where in the world is it used?
While Climeworks and its two Icelandic projects are the flag-bearers for the technology, DAC is a growing industry worldwide.
There are a raft of companies keen to develop the industry, scientists working on research and development, and deep-pocketed investors willing to spend big to make it happen.
Among them is 1PointFive, an arm of major American oil producer Occidental Engineering, and Canadian start-up Carbon Engineering, which is backed by tech billionaire Bill Gates.
At a site called Notrees – in a barren corner of the state of Texas – 1PointFive is building its own DAC plant to capture carbon.
Some of the carbon will reportedly be used by Occidental for purposes that can hardly be described as climate action – by being injected into oilfields to squeeze out extra production.
But some will be pumped underground in an end of itself and there are hopes the carbon will stay buried indefinitely.
Eventually, Occidental plans to build as many as 100 large-scale DAC plants in a little over a decade, while Climeworks is among others with plans for even bigger facilities in the US and elsewhere.
How much does it cost?
Climeworks, for its part, has been somewhat coy about the cost of the technology to date.
But, at the opening of Mammoth earlier this month, US media reported the company as saying the cost per tonne of carbon removed was closer to $US1000 than $US100.
And that matters greatly because $US100 a tonne is seen as a crucial benchmark for the viability of DAC.
At that point, the technology would theoretically become competitive with other options, such as planting trees, where a price was put on carbon.
But whereas trees burn in bushfires or can be cut down – releasing much of the stored CO2 – carbon buried underground through DAC could be kept there.
And polluting companies, perhaps wary of the bona fides of tree-planting schemes, for example, may be more willing to put their money into the technology.
That the industry is in its infancy doesn't help with its economics.
Still, there are signs that costs are coming down as more and more companies enter the fray and the technology evolves.
By accounts, the system developed by Carbon Engineering should be able to bury a tonne of carbon at a cost of between $US94 and $232, at least when deployed at scale.
So far, though, the economic goals of the industry have proved elusive.
Is it the same as carbon capture and storage?
There are similarities but DAC and carbon capture and storage differ in a critical way.
In theory, DAC has negative emissions so long as the energy that's used to power the process causes fewer emissions than it reduces from the atmosphere.
That is the case for Mammoth and its sister project Orca, which are both powered by the abundant and clean geothermal energy on tap in Iceland.
For carbon capture and storage (CCS), on the other hand, there are no such illusions.
The aim of the game is merely to minimise as much as possible the pollution that would ordinarily occur in many industries.
Take, as an example, the giant Gorgon liquefied natural gas plant in Western Australia.
A condition of the project's approval by the WA government was the construction of a CCS plant to bury much of the carbon from the gas fields.
Even were the project to be working properly, it would still only be sequestering about 80 per cent of the "reservoir" CO2.
But the project isn't working as intended, and just 34 per cent of the five million tonnes of CO2 it captured last financial year was injected underground.
Then there's the matter of carbon capture utilisation and storage, or CCUS, as if another acronym was needed.
This is the process of taking carbon and pumping it into oil and gas fields to extract greater volumes of fossil fuels.
Climeworks, for the record, says it has nothing to do with fossil fuels.
How much difference can it make?
Given it's minuscule current size, it's hard to see how the DAC industry can have any effect on global warming at all.
Despite its name and large size compared with other projects in operation, Mammoth will be able to pull just 36,000 tonnes of carbon from the atmosphere a year – and that's running flat out.
Climeworks points out that such a reduction is equivalent to taking almost 8,000 gas guzzling cars off the road.
The company says it wants to be cutting one million tonnes of carbon from the atmosphere every year by the end of the decade.
Occidental, meanwhile, says it's designing another plant for Texas that will have the capacity to remove 500,000 tonnes of carbon a year from the air.
All of these numbers and goals and claims need to be seen in context, though.
Every year, the world burns fossil fuels that produce about 35 billion tonnes of carbon dioxide.
Roughly one trillion tonnes of carbon dioxide have accumulated in the atmosphere thanks to the activities of humankind.
Even the International Energy Agency's target of removing 70 million tonnes of carbon a year by 2030 – modest though that may seem – is way off in the distance.
What's more, the world does not have time on its side.
Global temperatures are already rising, and fast.
According to the Copernicus Climate Change Service, April marked the 11th straight month in which global average surface temperatures hit a record.
For the past 12 months, the agency noted, the global average temperature was the highest on record at 1.5 degrees Celsius above pre-industrial levels.
A big challenge facing DAC, in particular, is physics.
Carbon concentrations in the air, although rising, are tiny and extracting meaningful amounts of the gas using that method is difficult.
Will it be a fig leaf for fossil fuels?
A constant criticism of carbon capture – of any sort – is that it legitimises the burning of fossil fuels by providing a veneer of respectability.
And, certainly, the industry itself is becoming keener by the day on CCS and CCUS, otherwise known as enhanced oil recovery.
Compared with DAC, it's relatively easy to capture high concentrations of carbon with CCS, too, given the plants tend to be built alongside industrial facilities.
But proponents of direct air capture stress the technology shouldn't be conflated with the others and will be critical to industries that can't easily go green.
These industries are a big deal, and a big source of emissions, in their own right.
Think cement producers or steelmakers or chemical and fertiliser manufacturers.
Often referred to as "hard-to-abate", these are industries that cannot decarbonise simply by using renewable electricity rather than dirty power.
For them, fossil fuels such as natural gas often are an essential input for a chemical process – required for the molecules instead of the energy they might provide.
Perhaps the last word should go to the United Nations.
Two years ago, it found the world was so far behind on reducing emissions that some sort of carbon capture was "unavoidable" so long as the world wanted to meet its own climate goals.