Charles Cadieu - Co-Founder and CEO of Spiritus
Climate Tech / AI Pioneer

Charles
Cadieu

The neuroscientist who decided to fix the sky.

He modeled how brains process light. Then he taught computers to see hearts. Now he's building orchards that breathe CO2 - at a cost that changes what's economically possible.

Co-Founder & CEO Spiritus MIT + UC Berkeley 2x Exits San Francisco
$41M+ Total Raised - Spiritus
3 Companies Founded
2 Successful Exits
2MT Annual CO2 Target - Wyoming
$100 Target Cost per Ton CO2

The Lung Architect

There's a moment in every Charles Cadieu company where the science does something the market didn't expect. At IQ Engines, in 2013, it was image recognition fast enough to organize a photo library in real time. At Caption Health, it was an AI that guided a nurse with no ultrasound training through a cardiac scan good enough for clinical use. At Spiritus, the moment is simpler and stranger: a material that sucks carbon dioxide out of the air without fans. No motors. No forced airflow. Just passive contact - the way a lung works.

The PhD Cadieu earned at UC Berkeley's Redwood Center for Theoretical Neuroscience was titled, in the dry language of academia, "Probabilistic Models of Phase Variables for Visual Representation and Neural Dynamics." What it actually studied was how biological systems extract signal from noise - how a brain decides what it's seeing, under uncertainty, using probabilistic inference over incomplete data. That mathematical vocabulary has followed him across every company he's built.

We're seeing soaring demand for data centers and heavy industries, yet we can't ignore the carbon that comes with it. Our DAC technology brings large-scale decarbonization within reach.

- Charles Cadieu, Series A Announcement, March 2025

Yahoo, Then GE, Then the Atmosphere

The arc is useful to hold in mind. Cadieu didn't fall into climate tech after some environmental epiphany. He followed the logic of the problem. IQ Engines built computer vision in a world before it was common. Yahoo bought it in 2013 - before the deep learning revolution made image recognition a commodity. He was early.

Caption Health applied that same early-to-market instinct to medical imaging. The company developed AI guidance that could walk a non-specialist through a cardiac ultrasound acquisition - point the probe here, adjust the angle there - and produce images good enough for clinical decision-making. The FDA's landmark 2020 authorization for Caption Guidance was the first time regulators blessed an AI system for guiding image acquisition (not just analyzing the resulting image). GE HealthCare acquired the company in February 2023. TIME Magazine had already called it one of the 100 Best Inventions.

Within months of the GE deal closing, Cadieu was co-founding Spiritus with Matthew Lee, a former staff scientist at Los Alamos National Laboratory where the underlying sorbent technology had been developed. The company came out of stealth in September 2023 with an $11 million seed round led by Khosla Ventures and support from Frontier - the advance market commitment backed by Stripe, Alphabet, McKinsey, Shopify, and Meta. It was not a typical seed story.

SPIRITUS = FROM THE LATIN: BREATH, AIR, SPIRIT

The Carbon Orchard

Direct air capture is not a new idea. Strip CO2 from the air, store it underground - simple in concept, ruinously expensive in practice. Current industrial approaches run $600 to $1,000 per metric ton. At those prices, meaningful scale is a budget problem before it's a technology problem. Spiritus's claim - and it's a significant claim - is a path to $100 per ton through passive air contacting and low-temperature desorption.

The "Carbon Orchard" is the physical instantiation of this: an array of passive air contactors filled with sorbent material that binds CO2 from ambient air without powered airflow. Once the sorbent is saturated, low-energy heat regenerates it, releasing concentrated CO2 for geological sequestration. No fans means dramatically lower operating energy. The structure is modular and can be deployed at scale in open terrain - Wyoming rangeland, New Mexico desert, eventually anywhere with the right geology for underground storage.

In March 2024, Spiritus announced "Orchard One" in Wyoming - a facility designed to sequester 2 megatons of CO2 annually, among the largest planned DAC projects in the United States. A 1,000-ton pilot facility is under development on Nambé Pueblo tribal land in New Mexico, expected to begin operations in early 2026. TIME Magazine named the technology one of its 200 Best Inventions of 2024.

Series A and Strategic Alignment

March 2025 brought a $30 million Series A led by Aramco Ventures. The co-investors - Khosla Ventures, Mitsubishi Heavy Industries America, TDK Ventures - are not passive financial bets. Aramco comes with Middle East deployment potential. Mitsubishi brings heavy industrial infrastructure and Japanese market access. The round gave Spiritus $41 million in total funding and validated a strategy of pairing American carbon removal innovation with global industrial partners who have the physical footprint to deploy it.

Around the same funding announcement, Spiritus expanded its leadership team with executives from Tesla, Meta, and ExxonMobil - a mix that signals the company is past the research phase and into the build-and-deploy phase. Dorian West and Kip Hensley joined to lead commercial scale-up. The talent profile of a company often tells you more about its current problems than its press releases do.

Cadieu has been direct about the policy and industrial context: the same AI infrastructure build-out and industrial expansion that's driving energy demand is also driving carbon emissions. His bet is that the economics of DAC, with the right technology, can reach a point where removing carbon becomes a viable industrial service - not charity, not government program, but a market.

America can - and must - grow its industrial base, but we can't let carbon emissions run unchecked. Our technology ensures large-scale progress does not mean large-scale emissions, preserving both economic opportunity and environmental responsibility.

- Charles Cadieu, Spiritus CEO

The Through-Line

What connects image recognition (2013), AI-guided cardiac ultrasound (2023), and passive direct air capture (2025)? Cadieu's answer, implicit in the pattern if not always stated explicitly, is that they're all problems of sensing and inference at scale. How do you extract useful signal from an overwhelming environment? IQ Engines did it with pixels. Caption Health did it with ultrasound waveforms. Spiritus does it with molecules in the atmosphere.

His academic work studied how the brain resolves ambiguity under uncertainty - specifically how "phase variables" in neural representation encode the relationships between visual features, not just the features themselves. It's a subtle point, but the insight is about structure: the patterns in patterns. The Carbon Orchard's passive architecture is, in a different register, the same kind of insight. Instead of forcing air through a system (imposing structure from outside), you let the environment do the work. The sorbent material is structured to capture what's already there.

Whether Cadieu would frame it that way is another matter. Founders rarely narrate their own arcs in the way observers do. What's clear is the empirical record: he has consistently identified technologies that were technically plausible but economically premature, built companies around them before the market was ready, and exited when the market caught up. At Spiritus, he's making the same bet with higher stakes and a longer time horizon than anything he's attempted before.

What Comes Next

The Nambé Pueblo pilot facility in New Mexico is the near-term proof point. If 1,000 tons of CO2 can be removed at a cost trajectory consistent with the $100/ton target, the Wyoming Orchard One project becomes fundable at the scale required. The Saudi Arabia partnership with Aramco opens a separate deployment pathway in a region with abundant sun (heat for desorption) and specific geology favorable for sequestration.

Cadieu has said he expects "a handful" of Carbon Orchard projects before the end of the decade. In direct air capture terms, that would be a significant milestone. The industry has been promising scale for years; few have delivered it. The question is whether Spiritus's technology advantage - specifically the energy cost of passive contacting versus fan-driven systems - holds up at operational scale. The pilot will answer that question. The Series A says investors think the odds are favorable.

Cost Per Ton of CO2 Removed

The industry's biggest barrier isn't physics - it's price. Spiritus is targeting a 90% cost reduction through passive air contacting and low-energy desorption.

Industry Standard (Current) $600-$1,000/ton
Spiritus Trajectory (Near-Term) ~$300-400/ton
Spiritus Long-Term Goal $100/ton

From MIT to the Atmosphere

2004-2009
UC Berkeley PhD - Computational Neuroscience at the Redwood Center for Theoretical Neuroscience under Bruno Olshausen. Dissertation on probabilistic visual representation and neural dynamics - the mathematical foundation for everything that followed.
2012-2014
MIT McGovern Institute - Postdoctoral research with James DiCarlo studying how the brain's visual cortex processes objects. Deep learning was just emerging as a serious field; Cadieu was already in the middle of it.
2013
IQ Engines acquired by Yahoo!/Flickr - AI image recognition startup exits in one of the early acquisitions in the modern computer vision era. The founding predates widespread deep learning adoption.
2017-2023
Caption Health - Co-founded with Kilian Koespell. Developed AI that guided untrained users through cardiac ultrasound acquisition. FDA authorization in 2020 was a first in medical AI. GE HealthCare acquisition closes February 2023.
2023
Spiritus founded and out of stealth - Co-founded with Matthew Lee (ex-Los Alamos National Lab). Emerged September 2023 with $11M seed led by Khosla Ventures and Frontier advance market commitment participation.
2024
Orchard One announced + TIME recognition - Wyoming Carbon Orchard designed to remove 2 megatons CO2 annually. TIME Magazine names Spiritus technology among 200 Best Inventions of 2024.
2025
$30M Series A - Led by Aramco Ventures with Khosla, Mitsubishi Heavy Industries America, and TDK Ventures. Nambé Pueblo pilot facility on track for 2026 operations. Leadership team expands with Tesla, Meta, and ExxonMobil alumni.

Spiritus Carbon Orchard - Anatomy of Passive DAC

The Carbon Orchard takes its structural inspiration from biology - specifically the passive, high-surface-area architecture of the lung.

Passive Air Contacting

No fans or blowers. The sorbent material is arranged in open-air structures that allow ambient wind to carry CO2-laden air across the capture surface. Energy consumption drops dramatically compared to fan-driven systems.

Sorbent Technology

The proprietary sorbent material (developed at Los Alamos National Lab) binds CO2 from ambient air at concentration levels of approximately 420 parts per million. Sorbent durability is key to unit economics at scale.

Low-Temperature Desorption

Once saturated, sorbent is regenerated using low-temperature heat - potentially from waste heat or solar thermal - rather than energy-intensive high-temperature processes. This is the second major lever on cost.

Geological Sequestration

Concentrated CO2 is injected into geological formations for permanent storage. Site selection prioritizes geology that meets compliance and verification requirements. Spiritus uses real-time MRV (monitoring, reporting, verification) for carbon credit quality.

Verifiable Milestones

Exit #1

Yahoo/Flickr Acquires IQ Engines

August 2013. One of the early exits in the modern computer vision era, before deep learning became mainstream.

FDA First

First AI-Guided Image Acquisition Authorization

Caption Health received the first-ever FDA authorization for an AI system that guides image acquisition (not just analysis). A regulatory landmark in medical AI.

TIME 100

TIME Magazine Best Inventions

Caption Guidance named among TIME's 100 Best Inventions following FDA authorization and clinical deployment.

Exit #2

GE HealthCare Acquires Caption Health

February 2023. Strategic acquisition by one of the world's largest medical device companies, expanding AI-guided ultrasound access globally.

TIME 200

TIME 200 Best Inventions 2024

Spiritus Carbon Orchard named among TIME's 200 Best Inventions of 2024 for the passive direct air capture technology.

Series A

$30M Led by Aramco Ventures

March 2025. Strategic Series A with Aramco, Mitsubishi, Khosla Ventures and TDK - giving Spiritus both capital and global industrial deployment partners.

The Through-Line

10

Years between exits: Yahoo acquired IQ Engines in 2013; GE HealthCare acquired Caption Health in 2023. Cadieu appears to operate on decade-scale cycles.

3

Consecutive companies all involving pattern recognition: pixels (IQ Engines), ultrasound waveforms (Caption Health), atmospheric molecules (Spiritus).

$75M

Total capital raised across all three startups, including VC and non-dilutive sources like the Bill & Melinda Gates Foundation and NSF.

420

Parts per million CO2 in the atmosphere - the concentration Spiritus's sorbent must capture from. This is a very dilute signal, which is why passive air contacting architecture matters so much.

2026

Expected start of operations at the Nambé Pueblo pilot facility in New Mexico - the first real-world proof point for Carbon Orchard technology at meaningful scale.

Share This Profile