Psst… We received a tip…

You’re in the service corridor of a data center. Concrete underfoot, fluorescent buzz overhead, warning labels on every door. 

Then you feel that familiar sensation. Hot air streaming through a vent seam. The building breathes. Fans pulling, pushing, balancing pressure. Cold aisle, hot aisle. A constant inhale and exhale that never stops. 

Every second, data centers push air through themselves. Most people think that airflow is just the tax you pay for compute. This patent treats it like an asset. The building already moves huge volumes of air as a non-negotiable operating cost. 

This patent taps into the air flow machine that’s already roaring behind the walls, to tackle one of the biggest environmental challenges of our generation. The system captures carbon as the air passes, then regenerates captured material. It turns the data center’s breath into a carbon capture beast.

This patent comes from Provocative Science Holdings, a Massachusetts outfit that pitches “carbon capture built into data centers”, swapping the usual cooling story for a capture story.

One of the reasons carbon capture struggles outside controlled lab conditions is humidity. Ambient air carries water vapor, and many capture materials bind water along with CO₂. That can lower effective capacity, slow cycles, and spike the energy needed to process the material.

This patent is built around that problem. It features a moisture datum (a measured data point) inside the capture chamber, feeds that information to a controller, and adjusts heat output as a function of that moisture datum. The aim is predictable carbon capture cycling even when the air is wet, sticky, and variable.

Now that is clear, let’s look at the rest of the machine step by step from the beginning.

Start with the part data centers already do better than almost any building on Earth. They move air constantly.

The patent wraps a carbon-capture loop around that forced airflow. The air intake receives a gas stream that contains a “carbonic substance”, coming directly from an exhaust system. The patent proposes an ideal use case is an air cooling system of a data center.  

From there, the patent’s system runs two functions.

The reason the moisture controller we mentioned before matters is that this whole cycle lives and dies on repeatability. The patent’s control loop treats humidity as an input to the regeneration step.

Then comes the part that changes the business conversation. It captures carbon and compresses that carbonic substance into a “carbonic liquid.” This raises questions about what happens to this output, and is it a product in itself?

The patent claims the intake can receive gas from exhaust systems generally, not only data centers. 

There are two compatible adjacent categories.

First are other giant-airflow environments where the air is relatively clean, closer to ambient, and where humidity swings are part of life. Big-building HVAC exhaust, large-scale ventilation, campus systems. The tech story stays intact because the entire bet is that air is already moving.

Second are true combustion exhaust sources, backup generators, gas turbines, marine engines. Those streams have higher CO₂ concentration, which helps, but they are hotter and dirtier, and they come with contaminants that can poison capture materials. Using this system there is plausible, but it is not a drop-in.

The patent triggers a commercialization instinct that the capture unit should behave like facility infrastructure.

The IEA projects global data center electricity consumption roughly doubling to around 945 TWh by 2030. (IEA) Uptime Institute’s 2024 survey shows “supporting infrastructure” like cooling still eats a big slice of the bill. (Uptime Institute)

This function accounts for approximately 35.9% of total facility energy. This shows us that the building infrastructure already inherently does a lot of non-compute work to move heat and air. 

On the U.S. policy side, the implementation of the 45Q tax code is still the loudest price signal. Congressional Research Service summarizes the post-IRA structure as up to $180 per metric ton for direct air capture with prevailing wage and apprenticeship requirements met, and up to $85 per ton for point-source capture with geologic storage under the same labor conditions. Those are big numbers if you can actually deliver verifiable tons.

Demand for carbon removal is trying to show up too. Frontier, founded by Stripe, Alphabet, Shopify, Meta and McKinsey Sustainability, is an advance market commitment to buy $1B+ of permanent carbon removal between 2022 and 2030. And Climeworks’ Mammoth plant is a 36,000 tCO₂ per year facility, a useful “current-gen” benchmark for what large means in standard carbon capture systems today. (Reuters)

Shipping is the headline because the emissions pool is enormous and the sector is under tightening regulation. The European Commission pegs global shipping at 1,076 million tons of CO₂ in 2018, about 2.9% of global human-caused emissions. If even 5% of that were captured onboard, that is roughly 54 Mt CO₂ per year. At $100 per ton in value, credits, avoided penalties, fuel standards, that is a $5B per year prize. 

Cement is bigger in raw emissions, commonly cited around 7% to 8% of global CO₂. It is also a harder retrofit story. High temperatures, dusty exhaust, brutal uptime requirements. The market is huge, but a data-center-airflow-shaped system is not the obvious wedge…. Sounds like an interesting problem to ideate around!

The third adjacent is the one people forget because it sounds boring. Power plants built to feed data centers. Google just signed an agreement tied to a new 400 MW gas plant in Illinois designed to capture roughly 90% of emissions. That deal is not this patent’s architecture, but it is the same macro signal. Compute growth is pulling in carbon management infrastructure, either at the data center, at the power plant, or both. (Reuters)

Finally, the downstream constraint hiding in plain sight.

The patent’s method compresses the captured CO₂ into a liquid. That sounds like you just created a product. The IEA estimates about 230 Mt of CO₂ is used each year, mostly for urea and enhanced oil recovery. That market is real, but it is not infinite, and not all of it counts as durable climate removal.

So the commercial question is not only “can you capture?”, it also raises questions of “who takes custody of the molecule next?”

The investment signals point towards direct air capture, durable removals, and the infrastructure that makes custody and verification possible.

The cleanest signal is late-stage capital showing up for permanent removal solutions

Heirloom uses limestone to permanently remove billions of tons of carbon dioxide from the air while creating new energy pathways. They closed a $150M Series B co-led by Future Positive and Lowercarbon Capital, with strategics like Japan Airlines, Mitsubishi Corporation, Mitsui, and Siemens Financial Services joining.

Svante (point-source capture) reported a C$435M (US$318M) Series E led by Chevron New Energies, and Canada’s Growth Fund announced an investment of up to US$100M (C$137M). (BetaKit)

Neustark (mineralization-based removal) raised $69M in a growth equity round led by Decarbonization Partners (BlackRock and Temasek).

Charm Industrial (bio-oil sequestration) raised a $100M Series B, and stated it had delivered over 6,200 tons of carbon removal at the time of announcement.

On the M&A side, Occidental agreed to buy Carbon Engineering for $1.1B to accelerate its direct air capture buildout. (Reuters)

The market for carbon removal credits could grow from $2.7B (2023) to as much as $100B per year (2030 to 2035) if key barriers are solved, and it estimates $32B has already been invested in removal projects, including $21B into engineered removals. (Reuters)

Scaling capture and storage reshapes jobs and procurement, but the numbers vary by scenario. A US DOE workforce analysis estimates 116,200 to 174,300 average annual “investment jobs” over a five-year construction period for one CCS (Carbon Capture & Storage) retrofit scenario. NETL has cited much larger “up to 1.8 million jobs” potential in a major CCS buildout.

The hard constraint remains cost. The IEA notes current projects in this industry cost an estimated $500 to $1,900 per tons today. That is why the patent’s obsession with moisture control and predictable cycling is a direct attempt to push down operating costs in messy real air.

When a building is forced to breathe, you can either treat that airflow as pure overhead, or turn it into throughput.

This patent routes a data center’s constant ventilation through capture chambers, then cycles regeneration with heat and vacuum so the captured material can become productive.

If it works the way it’s claimed, you get carbon removal that rides on infrastructure already running 24/7, instead of paying for a whole new airflow machine.

Ready to learn more? Dive into the details: US12491467B1.