Shrey Agarwal
I am the Co-Founder & CEO of Alt Carbon, where we are building the AI infra to understand and explore Earth systems.
I was brought up amidst the forests & tea gardens of Darjeeling (India) and now divide my time across Alt’s research labs at Darjeeling, Bangalore and Cambridge (UK).
I studied Chemistry & Mechanical Engineering at BITS Pilani and led consumer AI at SuperShare (0 → Series A) and seekho.ai (-1 → Series A, acquired), alongside my double major.
I love chemistry, deep time & technology.
I’m interested in:
- Exploring new institutional forms for scientific innovation, from FROs, automated labs, cities of abundance, to modern lunar societies
- Creating different forms of civilization backup storage, like the Svalbard Seed Vault
- Building inter-disciplinary labs inspired by Bell Labs, Xerox PARC, and MIT's Building 20
- Building and applying Natural General Intelligence (NGI)
- Drilling deeper than the Kola Borehole in my lifetime to unlock novel subsurface insights
- Improving my understanding of deep time: Snowball Earth, Year Without a Summer
- Finding samples like the Apollo lunar rocks that were sealed for 50 years so future scientists with better tools could open them
- All things science fiction (my favourite short story is Asimov's "The Last Question")
- Signing up for quests that involve singing sand dunes, moonbow & synchronous fireflies
- Studying modern chess openings and endgames
I try my best to respond to every thoughtful email I receive at shrey@altcarbon.com.
Research
My research group chases open questions whose answers could become revenue generating verticals, deepen our understanding of Earth systems, or ideally both. A few we're working on now:
Enhanced Rock Weathering
- How much of the weathering flux can we attribute to the applied basalt rather than to soil and carbonate backgrounds, and can radiogenic isotope ratios like ⁸⁷Sr/⁸⁶Sr make that attribution unambiguous?
- Can reactive transport models calibrated at catchment outlets reliably predict alkalinity export, given cation exchange lags and secondary carbonate formation?
- How long does ERW-derived bicarbonate reside in groundwater, and along which flow paths does it reach rivers?
- What fraction of captured carbon is lost between farm and ocean to CO₂ degassing and carbonate precipitation?
- Can we predict a feedstock's carbon removal potential from its mineralogy, grain size, and dissolution kinetics before it leaves the quarry?
Mineral Exploration
- Can joint Bayesian inversion of gravity, magnetics, magnetotellurics, and seismic data produce posterior ore-body geometries sharp enough to site a single drill hole?
- Can geological priors, built by simulating the source, pathway, trap, and preservation of ore-forming systems, generate synthetic worlds realistic enough to train models that find deposits under cover?
- Can drilling campaigns be designed as sequential decisions, where each hole is chosen to maximize the expected value of information rather than the odds of a single hit?
- How do we train prospectivity models when known deposits number in the dozens and any unlabeled cell may hide an ore body, and can positive-unlabeled learning handle that?
- Can deep-crustal and lithospheric conductors imaged by MT, the fossil fluid pathways of ancient mineral systems, narrow a country-scale search to a district-scale one?
Over the next decade, I'd love to go deeper into natural hydrogen and the serpentinization reactions that produce it, life in the deep biosphere, why Himalayan springs are drying up, and new ways of imaging the subsurface with muons and quantum sensors.
Investing
My brother & I started investing together in 2026. We like to support big if true ideas, difficult problem statements, and ambitious, passionate founders. We are agnostic to sectors, and would like to be a part of wild journeys ranging from building an ice cream for Indian summers to bioacoustic models to speak to animals and birds.