
Helix Carbon proposes recycling carbon dioxide from blast furnaces into carbon monoxide to reduce coke use in iron production, claiming the method could cut steel industry emissions and save producers $14 per ton of iron while targeting specialty chemicals first for early revenue.
AI-generated summary
The iron and steel industries are major global polluters, and many startups have attempted to reduce emissions by developing new metal-making methods, but Helix Carbon argues that addressing the blast furnace is essential for meaningful decarbonization.
The iron and steel industries are among the biggest polluters in the world. Plenty of startups have tried to cut the industry’s emissions, mostly by developing new ways to make the metals. Helix Carbon thinks those startups have been going about it all wrong.
“Any attempt to decarbonize the steel industry without dealing with the blast furnace is not going to work,” David Brown, Helix Carbon’s co-founder and CEO, told TechCrunch. A blast furnace is the giant furnace that turns iron ore into the iron used to make steel. Rather than replace blast furnaces, Helix wants to clean them up. Its approach promises to slash the steel industry’s carbon emissions while boosting profit margins for iron producers, a rare win-win for a dirty industrial process.
Blast furnaces have made iron for centuries, using coke, a processed form of coal, to help turn raw iron ore (iron bound up with oxygen) into metallic iron. The burning coke supplies heat for the reaction but doesn’t take part in it directly. Instead, carbon monoxide produced from the combustion does the work of reducing the iron, meaning stripping away its oxygen atoms.
The process releases tremendous amounts of carbon dioxide, or CO2, which is normally dumped straight into the atmosphere. But Helix Carbon has devised a way to turn that waste carbon dioxide into carbon monoxide and recycle it back into the blast furnace.
Helix Carbon’s process relies on an electrolyzer, a device that uses electricity and a catalyst to break off an oxygen atom from the carbon dioxide. Electrolyzers aren’t new, but Brown said the startup’s design can handle the entire exhaust stream from the blast furnace. Most electrolyzers are too delicate for raw exhaust like that. They work only on purified, concentrated carbon dioxide, so the exhaust typically has to be processed before they can use it.
By injecting carbon monoxide into a blast furnace, iron producers can use less coke. How much, exactly, is still debated, Brown said. Some scientific papers suggest as much as 90% of the coke can be eliminated, while other studies suggest it’s more like 30% to 40%.
Helix Carbon is proposing that iron producers start by replacing 33% of their coke. The company says using its technology to do so would save them $14 per ton of iron. “That may not seem like a lot, but margins in the iron industry are famously thin,” Brown said.
The iron and steel industries are a combined $2 trillion market, according to Grand View Research, which makes the sector an appealing target for a young startup to chase. But it also moves slowly, in part because its equipment lasts so long. “There are blast furnaces that have lived for over 100 years,” Brown said.
To bring in revenue earlier, Helix Carbon plans to court the specialty chemicals industry first. “We don’t need to capture the CO2 first,” Brown said. “We just utilize it, and that opens up the door of making chemicals on site right out of waste streams.”
AI outlook — possibilities, not facts
Helix Carbon will secure partnerships with iron producers to pilot its blast furnace emissions recycling technology within the next 12 months.
Likely · Within months
Helix Carbon will generate initial revenue from the specialty chemicals industry before expanding into iron and steel applications.
Very likely · Within months

At TechCrunch Disrupt 2026, Cerebras Systems CEO Andrew Feldman criticized the growing AI doomer narrative, arguing that AI leaders should highlight the technology's potential benefits, such as ending diseases like cancer, while acknowledging risks. He emphasized the need for thoughtful regulation and community engagement in data center development, citing better practices in countries like Brazil, Canada, and Nordic nations compared to U.S. NIMBYism.

OpenAI said it alerted the FBI after discovering an employee, identified as Saudi national Abdulrahman Alfozan, planted a hidden recording device in an executive conference room. The device was found in a TV USB port in OpenAI's San Francisco office. Alfozan was fired and arrested on October 9 and charged with intercepting oral communications. OpenAI stated its internal AI models helped analyze the 'sophisticated' device, which had advanced encryption and data transfer mechanisms. The FBI said the scheme aimed to capture sensitive communications about the company's super intelligence technology.

A company discovered a sophisticated hidden recording device in one of its offices, used internal models to trace it to a specific employee who planted it days earlier, referred the matter to the FBI leading to an arrest, and confirmed no evidence of data exfiltration; the employee was terminated and the FBI investigation remains ongoing.

A company discovered a sophisticated hidden recording device in one of its offices, used internal models to trace it to a specific employee who planted it days earlier, reported the incident to the FBI leading to an arrest, and confirmed the employee was terminated with no evidence of data exfiltration.

A company discovered a sophisticated hidden recording device in one of its offices, used internal models to trace it to a specific employee who planted it days earlier, referred the matter to the FBI leading to an arrest, and confirmed no data exfiltration occurred. The employee was terminated and the FBI investigation remains ongoing.

A U.S. judge has ruled that Apple will gain expedited access to OpenAI's hiring materials related to interviews with former Apple employees, as part of Apple's trade secrets lawsuit filed in July. The court ordered OpenAI to turn over interview documents within five days and other related materials within 14 days, while denying broader requests for access to OpenAI's research and development.