Anthropic has become the first pure AI startup to join Frontier, the carbon removal collective founded by major tech companies. The company contributed to a new $915 million tranche of funding, bringing total pledges to $1.8 billion. This marks a significant milestone for both Anthropic and the broader AI industry, as tech firms face increasing scrutiny over their energy consumption and carbon footprints.
Frontier was launched in 2022 by Stripe, Google, and Shopify to help tech companies fulfill their climate pledges by funding carbon removal projects. The collective vets emerging carbon removal technologies and signs contracts with those it deems most promising. To date, Frontier has committed nearly $700 million across more than 50 projects, resulting in 1.8 million tons of carbon dioxide removal. With the new funding, Frontier aims to scale its efforts, focusing on fewer but larger contracts that target technologies capable of removing one gigaton of CO2 annually.
Anthropic's membership is particularly notable because it is the first AI company without a broader tech portfolio to join. While Google is a founding member, Anthropic represents a new category of participants: companies whose core business is artificial intelligence. This comes at a time when AI companies have been on an energy buying spree, often securing power from fossil fuel sources to fuel the massive data centers required for training and running large language models. Anthropic itself has not yet published a sustainability report and has publicly endorsed an “all of the above” energy strategy, which critics argue can dilute commitments to renewable energy.
The company’s decision to join Frontier may signal a shift in attitude. By investing in carbon removal, Anthropic acknowledges that some emissions are unavoidable in the near term and that offsetting them through verified removal is a pragmatic step. Carbon removal credits function much like financial offsets: companies purchase credits representing a ton of CO2 removed from the atmosphere, and these credits can be subtracted from their reported carbon footprint. However, the integrity of such credits depends heavily on the quality and permanence of the removal projects, which is why Frontier’s vetting process is crucial.
The landscape of carbon removal technologies
Carbon removal encompasses a range of approaches, each at different stages of commercial readiness. Direct air capture (DAC) uses chemical processes to pull CO2 directly from ambient air. Enhanced rock weathering accelerates natural geological processes by spreading crushed silicate minerals on land, where they react with CO2. Bio-oil production involves converting biomass into a stable oil that can be injected underground. Ocean antacids aim to increase the alkalinity of seawater to absorb more CO2. Bioenergy with carbon capture and sequestration (BECCS) combines biomass energy generation with capturing the resulting CO2 emissions. Frontier has backed all these technologies, but its new strategy will concentrate on those with the greatest potential for gigaton-scale deployment.
The shift from many small bets to fewer large ones mirrors a similar evolution at Microsoft, which is the largest corporate buyer of carbon removal credits. Microsoft has signed multi-year contracts with companies like Climeworks and Carbfix, but it is also demanding that its suppliers demonstrate a path to government subsidies. Frontier made clear in its announcement that any new contract must include a plan for eventual government support. This reflects a broader recognition that carbon removal, like clean water or renewable energy infrastructure, is a public good that will likely require sustained public funding to reach climate-relevant scales.
The United Nations Intergovernmental Panel on Climate Change (IPCC) has stated that carbon dioxide removal technologies are essential to achieving net-zero emissions by mid-century. However, the cost remains a barrier. DAC currently costs hundreds of dollars per ton of CO2, far above the price of many emissions reduction strategies. Frontier’s contracts help bring costs down by providing early revenue to developers, but the collective has also indicated it does not intend to underwrite the industry indefinitely.
Why AI companies face a unique climate challenge
AI companies are particularly energy-intensive. Training a single large model can consume as much electricity as hundreds of households use in a year. Inference—the process of generating outputs from a trained model—also demands significant computational power, especially as AI applications become widespread. Data centers that host these computations are increasingly being built near renewable energy sources, but many still rely on natural gas or even coal to ensure reliability. The result is a growing carbon footprint for the AI industry, even as many of these companies publicly commit to climate goals.
Anthropic’s move into carbon removal may encourage other AI firms to follow suit. Companies like OpenAI, Cohere, and Mistral have not yet joined similar initiatives. By taking this step, Anthropic positions itself as a climate-conscious leader in a sector often criticized for its environmental impact. However, critics note that carbon removal should complement, not replace, direct reductions in emissions. Anthropic’s “all of the above” energy approach still allows for continued reliance on fossil fuels, which means the company will need to purchase a large volume of removal credits to balance its books.
The road ahead for Frontier
Frontier’s new funding cycle will prioritize projects that demonstrate a clear path to gigaton-scale removal and a strategy for securing government funding. Contracts will run between eight and ten years, longer than previous agreements. This extended timeline gives developers stability to plan and build their facilities. Frontier expects to contract as far out as 2040, but it has not specified what happens after that date. The implicit assumption is that by then, governments will have stepped in with robust carbon pricing or direct procurement programs to sustain the industry.
If governments do not take over, the carbon removal sector could stall, leaving the world short of the negative emissions needed to keep warming below 1.5 degrees Celsius. The IPCC models that require between 100 billion and 1 trillion tons of CO2 removal this century. Current capacity is negligible compared to that need. Frontier’s work is a crucial bridge, but it is not a permanent solution. The collective hopes that by demonstrating the viability of these technologies, it will create the political and economic conditions for government action. If that doesn’t happen, the climate impacts will accelerate, and society will face far greater challenges than the cost of removing carbon.
Source: TechCrunch News