A pioneering carbon removal coalition has secured a second major round of funding commitments and welcomed Anthropic, the company behind the Claude AI assistant, as a new member — a move that places one of the world’s most closely watched AI labs squarely inside the climate technology funding conversation for the first time.
Frontier was launched in 2022 with a pledge from Google, Stripe, and other founding members to spend $1 billion on carbon removal by 2030. In the years since, the initiative has signed contracts worth nearly $700 million covering 1.8 million tons of carbon removal across 53 separate projects. The model underpinning Frontier’s approach is known as an Advanced Market Commitment, or AMC — a funding structure designed specifically to de-risk early-stage technologies that would otherwise struggle to attract conventional investment because the market for their output doesn’t yet exist at scale.
A “Baton Pass” to a New Phase
Frontier’s second phase marks a deliberate narrowing of focus. Rather than spreading capital across a wide field of early-stage experiments, the coalition will now concentrate on a smaller group of projects that have a credible path toward government support once Frontier’s own funding commitment ends. The group has described this transition internally as a “baton pass” — handing proven approaches off to public-sector backing once they’ve cleared the highest-risk stage of development.
To fund this next stage, members have committed $915 million to what Frontier is calling the “Growth AMC.” This new pool of capital will target offtake agreements spanning eight to ten years, structured around partnerships with 10 to 15 companies operating in the carbon removal space. Alongside this larger structured commitment, Frontier will continue supporting what it calls “high-potential breakthrough ideas” through smaller mechanisms — prepurchases of removal credits, limited offtake agreements, and direct research grants aimed at technologies still too early-stage for the Growth AMC structure.
Timing Matters: A Market Under Pressure
The new funding arrives at a moment when the carbon removal sector badly needed a vote of confidence. The market was rattled in April by reports that Microsoft — by far the largest single purchaser of carbon removal credits to date — was slowing its buying activity, raising concerns about demand-side momentum for an industry still dependent on a small number of large corporate buyers.
That uncertainty was partially offset just last week, when the Science Based Targets initiative announced that companies seeking net-zero certification will be required to use carbon removals to cover a small but steadily increasing share of their ongoing emissions starting in 2035 — a policy shift that effectively builds long-term, structural demand for removal credits into the corporate climate compliance system itself, rather than leaving that demand purely voluntary.
Frontier’s $915 million announcement, in that context, functions as a counterweight: a signal that committed capital is still flowing into the sector even as its largest historical buyer reassesses its pace.
What Frontier Is Betting On
While Frontier did not disclose the specific projects it intends to fund under the new Growth AMC, the coalition shared cost and scale estimates for three carbon removal pathways based on lessons learned during its first funding phase.
Surficial mineralization tops Frontier’s list of priority approaches. The technique involves grinding rock into fine particles and exposing them to air, allowing the minerals to chemically absorb carbon dioxide directly from the atmosphere. Notably, much of the heavy machinery required already exists — the mining industry has spent decades developing the equipment needed to crush and process rock at scale, and large volumes of mining waste suitable for this purpose are already sitting in stockpiles worldwide. Frontier estimates this pathway could eventually capture more than 10 gigatons of carbon dioxide annually, at a projected cost of $80 to $120 per ton. The approach is currently being validated through small-scale field trials.
Ocean-based mineralization represents a related but distinct pathway, in which crushed rock is added directly to ocean water, triggering chemical reactions that draw down atmospheric CO2 and store it in dissolved form in seawater. Frontier estimates this approach also carries a theoretical ceiling above 10 gigatons annually, though at a somewhat higher projected cost of $100 to $200 per ton. The approach has already attracted corporate backing, with British Airways, Stripe, and Shopify among the companies that have funded early-stage ocean alkalinity projects.
Biomass carbon removal and storage is the third pathway Frontier highlighted, in which plants capture carbon through natural growth, and that carbon is then stored — typically by burning the biomass for energy while capturing and sequestering the resulting CO2 emissions, rather than allowing them to re-enter the atmosphere. Frontier estimates this approach could remove 1 to 5 gigatons annually, at a cost ranging from $60 to $200 per ton. Microsoft is among the most prominent backers in this category, having supported Stockholm Exergi, a Swedish utility that burns forest residues and other organic waste for electricity generation while capturing the resulting carbon emissions — one of the largest projects of its kind currently operating.
Why Anthropic’s Involvement Stands Out
Anthropic’s decision to join Frontier is notable largely because of how unusual it is within its peer group. Major AI model developers have, collectively, announced remarkably few sustainability commitments relative to the scale of computing infrastructure their industry now consumes — a gap that has drawn increasing scrutiny as data center energy demand tied to AI training and inference continues to climb globally.
That scrutiny is sharpened by a transparency gap that extends well beyond Anthropic alone. Neither Anthropic nor OpenAI, widely regarded as its closest competitor, disclosed emissions data to the Foundation Model Transparency Index, a research initiative run by Stanford’s Center for Research on Foundation Models that compiles and publishes public information on AI safety practices, training data sourcing, environmental impact, and related disclosures across the industry’s leading model developers.
Against that backdrop, Anthropic’s move to back a carbon removal initiative directly — rather than simply publishing voluntary sustainability commitments — represents a more concrete and financially binding form of climate engagement than most of its AI industry peers have offered to date. Whether other major AI labs follow with similar commitments, particularly as questions about the sector’s energy and carbon footprint intensify, remains an open question that Frontier’s next round of member announcements may help answer.
This is a sensitive topic involving climate policy and corporate environmental claims; the information above reflects publicly reported figures and statements from Frontier and related parties as of this writing.
