The Carbon Removal Investment Thesis: What 2026 Data Shows

Carbon removal scored 7.0/10 in Fathom's Q1 2026 analysis — the highest commercial viability of any sector across 87 scored peer-reviewed papers. That number matters. The average sector in our dataset scores 5.5. Carbon removal isn't close to average.

The question isn't whether carbon removal is important. Every serious climate model requires it. The question investors have been asking for five years is: is it investable yet? The 2026 data suggests the answer is shifting from "not quite" to "yes, selectively."

Why Carbon Removal Kept Failing the Investability Test

Carbon removal has a permanent place in IPCC scenarios, but it has historically struggled to clear the bar for institutional capital. The reasons are specific and worth understanding before making the investment case:

1. Cost — Direct air capture (DAC) cost estimates ranged from $250–$1,000/tCO₂ as recently as 2022. At those prices, the market requires either carbon tax equivalents north of $200/ton or heavy subsidy reliance. Neither has characterized most carbon markets to date.
2. Permanence uncertainty — Verification of long-term carbon storage (100+ years) has relied on physical monitoring that is expensive, sparse, and hard to audit. Voluntary carbon market scandals in 2022–2023 burned institutional buyers, making permanence evidence non-negotiable.
3. Site selection risk — Identifying suitable geological formations for CO₂ injection requires expensive seismic surveys. A bad storage site isn't just a technical failure — it's regulatory and reputational damage.
4. Scale gap — The technology readiness level (TRL) of most DAC approaches sat at 3–5 as of 2023, meaning proof-of-concept at best. Capital-efficient scale-up pathways were unclear.

Our Q1 2026 paper dataset shows each of these bottlenecks is now experiencing targeted technical progress. The signals aren't uniform — but they're present, and they're concentrated enough to identify specific investable thesis vectors.

What the Research Actually Shows in 2026

1. AI Is Solving Subsurface Characterization

The single highest-signal finding in our carbon removal dataset: machine learning models are dramatically reducing the cost and uncertainty of subsurface geological characterization for CO₂ storage.

Traditional site characterization for geological sequestration requires physical seismic surveys, core sampling, and modeling that can cost $5–15M per site and take 2–4 years. New approaches using physics-informed neural networks (PINNs) trained on existing subsurface data are compressing this to months and reducing capital requirements by an order of magnitude.

The papers scoring highest in our dataset describe systems that combine satellite-derived surface deformation data, existing seismic archives, and ML-based inversion models to produce high-confidence subsurface maps without new physical surveys. This isn't incremental — it's a fundamental cost structure change for the storage side of carbon removal.

Companies building in this layer — AI-native geological intelligence for carbon storage — represent the first investable thesis vector. They're picks-and-shovels plays that benefit regardless of which DAC technology wins.

2. Carbon Mineralization Is De-Risking Faster Than DAC

Direct air capture (the Climeworks/Heirloom model) gets the headlines. Carbon mineralization is moving faster on the investability curve.

Mineralization converts CO₂ to stable carbonate minerals — rock, essentially. The permanence story is unambiguous. The challenge has been process efficiency: getting CO₂ to react at meaningful scale and speed. Our dataset shows active research progress on three fronts:

• Enhanced weathering — Spreading silicate minerals (basalt, olivine) on agricultural land to accelerate natural CO₂ absorption. TRL is now 4–5 for field deployment, with multiple startups running commercial pilots. The agricultural co-benefits (improved soil health, crop yields) create a partial revenue offset that changes the economics.
• Ocean-based mineralization — Electrochemical processes that enhance ocean alkalinity, accelerating the ocean's natural carbon absorption. TRL is lower (3–4), but the research velocity in our dataset is the highest we've seen in this sub-category.
• Concrete and construction — Injecting CO₂ into concrete during curing (carbon-cured concrete) is TRL 7–8 and actively commercializing. This isn't "carbon removal" in the atmospheric sense, but it's permanent utilization with immediate economics.

The investability signal for mineralization: multiple viable pathways, lower permanence risk than geological storage, and active policy support (IRA's 45Q tax credits apply to enhanced weathering as of 2025 guidance).

3. Policy Architecture Is Locking In Demand

The research papers in our dataset don't analyze policy — but the commercial viability scores implicitly reflect it. A technology's TRL trajectory is meaningless without a demand signal. Carbon removal now has one.

Policy Signal	Impact	Timeline
IRA 45Q tax credits	$180/tCO₂ for DAC, $85/tCO₂ for geological storage	Active through 2032+
DOE Carbon Negative Shot	$3.5B committed to reducing DAC cost to <$100/tCO₂	Active
EU Carbon Border Adjustment	Creates compliance demand for verifiable removal credits	Phase-in 2026–2034
Voluntary corporate commitments	Stripe, Microsoft, Shopify pre-purchasing removal at $200-1000/tCO₂	Ongoing

The 45Q credit alone — $180/tCO₂ for point-source DAC with geological storage — puts the breakeven cost threshold for first-generation commercial DAC within reach for well-capitalized operators. That's a structural change, not a cyclical signal.

The Technology Landscape: Where Capital Is Flowing

Carbon removal is not one technology. It's a sector with five distinct pathways, each at different TRL stages and with different capital requirements:

Pathway	TRL (2026)	Capital Intensity	Investability Signal
Direct Air Capture (solid sorbent)	5–6	Very High	Strong (subsidies critical)
Direct Air Capture (liquid solvent)	4–5	Very High	Moderate
Enhanced Weathering	4–5	Medium	Strong (field-proven)
Bioenergy + CCS (BECCS)	4–6	High	Moderate (land-use constraints)
Ocean Alkalinity Enhancement	3–4	Medium	Early (monitoring gap)
Geological Storage AI Tools	5–7	Low–Medium	Strong (picks-and-shovels)

Our scoring methodology rewards technologies where: (1) the technical path is clear, (2) a commercial model exists without indefinite subsidy dependence, and (3) the competitive timing is right for early-stage capital.

On that framework, three sub-sectors stand out from the Q1 dataset: enhanced weathering at commercial pilot stage, geological storage AI tooling, and modular DAC systems targeting the 45Q credit window.

The Specific Investment Thesis

Here is what the 2026 research data supports — and what it doesn't:

✅ Where the Data Supports Investing

• AI-native geological intelligence — Companies using ML to slash subsurface characterization cost. Picks-and-shovels positioning. Benefits from every ton of CO₂ that goes underground, regardless of capture pathway. Defensible IP, high-margin SaaS-like revenue model.
• Enhanced weathering at scale — The agricultural co-benefit model creates partial revenue offset. Field trials are producing MRV data. The verification gap (proving permanence) is being solved by remote sensing + ML — the same technology convergence driving the Remote Sensing sector.
• Modular DAC platforms — Not billion-dollar greenfield DAC plants. Modular, containerized systems that can deploy alongside industrial emission sources to capture point-source CO₂ while the scale-up learning curve compresses costs. 45Q-eligible, deployable on existing industrial land.
• Carbon removal MRV infrastructure — The bottleneck for every removal pathway is verifiable, permanent, third-party auditable credits. Companies building the measurement and verification layer (satellite + ML + ground truth) are infrastructure plays with multiple exit vectors.

❌ Where the Data Doesn't Support Investing (Yet)

• Large-scale DAC plants without subsidy lock-in — Projects requiring $500M+ in capital before 45Q credits are secured carry unacceptable development risk. The policy window is real, but financing structures need to be airtight.
• Ocean-based approaches without MRV infrastructure — The science is promising. The verification gap is not close to solved. Institutional buyers won't touch ocean carbon credits until permanence is demonstrable at scale.
• Carbon removal "funds" with no technical differentiation — Purchasing diversified baskets of removal credits isn't venture. The alpha is in backing the companies de-risking the technology, not in the credit market itself.

Why the Timing Signal Is Strong in 2026

Our viability scoring penalizes technologies that are either too early (no commercial path) or too late (market already priced in). Carbon removal's 7.0/10 reflects a specific timing window: technical readiness has crossed the threshold for selective capital deployment, but the sector is not yet priced as a mainstream climate bet.

The comparable moment in solar was 2009–2012: technology readiness was proven, policy support was in place, but institutional capital hadn't yet recognized the sector as a core portfolio allocation. The investors who entered during that window captured outsized returns.

Carbon removal isn't solar — the capital intensity is higher, the verification infrastructure is earlier, and the market structure is more complex. But the pattern recognition is similar: a technology that works, policy that creates demand, and a market that hasn't fully priced the convergence.

How This Fits the Broader Q1 2026 Picture

Carbon removal's 7.0 viability score doesn't exist in isolation. In our Q1 2026 sector overview, we identified carbon removal as one of three sectors where research is graduating from theoretical to operational. The other two — AI for Climate (6.9) and Remote Sensing (6.1) — are actually enabling technologies for carbon removal itself.

The convergence is specific: AI for Climate provides the subsurface modeling and optimization capabilities that de-risk storage. Remote Sensing provides the MRV infrastructure that makes removal credits credible. Carbon removal is the asset class; AI and remote sensing are the infrastructure that makes it investable.

This isn't coincidence. It's the reason carbon removal scored first in the ranking.

What to Watch in Q2 2026

Based on research velocity in our dataset, these are the signals that will determine whether carbon removal's investability score holds or accelerates:

1. Permanence data from enhanced weathering pilots — Companies running 2–3 year field trials should have MRV data maturing. If permanence coefficients hold, institutional demand for enhanced weathering credits accelerates.
2. Modular DAC cost announcements — Several companies have committed to $/ton cost targets for 2026 modular systems. These numbers, when public, will either validate or reset the commercial timeline.
3. EU CBAM credit guidance — How the EU Carbon Border Adjustment Mechanism credits removal versus reduction will significantly affect corporate demand structure.
4. First commercial geological storage completions — The first fully-verified, commercially-operating geological storage sites create the proof point that MRV infrastructure needs to attract institutional buyers.

Next Steps

If you want to go deeper on carbon removal or the broader climate tech landscape:

• Explore the Carbon Removal sector page — live paper counts, average viability scores, and our AI-generated investment memo
• Read the Q1 2026 sector overview — carbon removal in context with the other nine sectors
• Browse individual scored papers — every paper with a thesis brief and viability rationale

— Fathom is an AI-powered climate venture scientist that scans and synthesizes 50,000+ research papers per week to surface high-signal investment opportunities. This analysis reflects data from 87 papers scored through Q1 2026.