Arnav Chandra

Carbon Capture Research

2024-06

Independent research into direct air capture technologies, focusing on cost reduction pathways and scalability.


Carbon capture is one of the most important and most underfunded areas of climate technology. The dominant narrative treats it as a backup plan, something we'll need if emissions reductions fail. The more I researched, the more I became convinced it's not a backup. It's a requirement.

This research project ran from October 2023 through summer 2024 and focused specifically on direct air capture (DAC): technologies that pull CO₂ directly from the atmosphere rather than from point sources like power plants. DAC is more expensive and energy-intensive than point-source capture, but it's the only approach that actually draws down historical emissions.

The core question I was trying to answer: what are the real cost reduction pathways? The current cost per ton of CO₂ captured via DAC is prohibitively high: $400–$1,000 depending on the system. Learning curve analysis, materials innovation, and process heat integration were the most promising levers.

I also spent time understanding the policy and market dynamics: carbon credit markets, government incentives like the 45Q tax credit, and how different regulatory regimes affect the economics.

The research sharpened my thinking on how hard problems with long time horizons actually get solved, and how rarely the path looks like what anyone predicted.


Outcome: Produced an independent research report on direct air capture (DAC), covering cost reduction pathways, materials innovation, and the policy and market dynamics shaping the space.

What I took from it:

  • Carbon capture isn't a backup plan. The emissions math doesn't close without it, and treating it as optional is a mistake the field is slowly correcting.
  • The real levers for cost reduction are learning curves, materials innovation, and process heat integration, not just subsidies. Understanding why something is expensive tells you where to push.
  • Hard problems with long time horizons rarely solve the way anyone predicted. The research made me more comfortable sitting with that uncertainty instead of needing the path to be clear before starting.