← Back to Darwin Oceans Open the platform →

The data platform for aquaculture.

We turn public and private ocean data into farm-level decisions for shellfish growers today — and we are building the data flywheel toward an ocean foundation model tomorrow.
Company & product overview  ·  June 2026
The problem

The data to de-risk a farm is public, and not used.

Disease wipes out crops

Dermo and MSX ride warm, salty water north. Growers find out after the die-off.

Closures hit blind

One rain event spikes coliform and the harvest stops.

Siting is a coin-flip

A lease is a 20-year bet on water no one has measured.

150+
contaminants tracked, unused
12
agencies, incompatible formats
>20°C
where disease takes hold
Darwin Oceans · Confidential01  /  The problem
The opportunity

A booming farm-gate economy, starting in Maine.

  • $16.2M Maine oyster value in 2025 — the state's 4th most valuable fishery (Maine DMR).
  • +539% growth in farmed-oyster value, 2011–2022 (Maine DMR / Aquaculture Reports). One of the Northeast's fastest-growing food economies.
  • ~70–80% of Maine's farmed oysters come from the Damariscotta River — the "Napa Valley of oysters," our beachhead.
  • Maine oysters command a quality premium over commodity Gulf oysters (~24% on US menus); demand outruns supply.
  • Chesapeake Bay adds a $58M+ fishery and a 10-billion-oyster restoration push (Chesapeake Oyster Alliance).

Who pays

Growers & co-opsLenders & insurers State agenciesRestoration NGOs

Every lease, loan, policy and restoration site is a decision our suitability and disease layers de-risk — the screening layer they cannot build themselves.

$1.4B
precision-aquaculture market by 2030 — hardware, software & services (MarketsandMarkets)
$30B
global shellfish-farming market scale (QY Research, 2024)
$11.1B
World Bank blue-economy portfolio of active projects (FY25) funding modernization
Darwin Oceans · Confidential02  /  Market
The product · live today

A working visor, on real data. Not a demo.

Real Maine DMR, NOAA and NERACOOS data, served live as COG tiles. Running today.
  • 994 real stations + 135 oyster leases, scored across Maine. Live data, not samples.
  • Suitability, disease & closure risk on one map: Dermo, MSX, Vibrio.
  • Click any site for a plain-language read.
  • Maine live today; Chesapeake is the second region.
Try it live → darwin-oceans-visor.web.app
Darwin Oceans · Confidential03  /  Product
The science · today

How we score a site today.

HSI = 100 × √( SIsalinity · SItemperature )
Salinity suitability27 ppt → SI 1.00
27
stressednormal 14–27 ppthyper-saline
Temperature suitability19 °C → SI 0.96
19
feeding stopsoptimal 18–26 °Cdisturbed
Damariscotta River site
Mean salinity27 ppt  →  SI 1.00
Mean temperature19 °C  →  SI 0.96
Geometric mean × 100100 × √(1.00 · 0.96)
98/100
Highly suitableGreen zone on the visor. Every pixel on the map is scored the same way.
Piecewise-linear suitability curves calibrated to published Crassostrea virginica thresholds (geometric-mean HSI, Swannack / Theuerkauf style); inputs from Maine DMR and NERACOOS. A transparent screening index, not ground truth. The same function runs disease pressure (Dermo, MSX, Vibrio).
Darwin Oceans · Confidential04  /  Suitability index
The 20-year view · roadmap

The best site today is often not the best site over a 20-year lease.

Maine DMR leases run up to 20 years — gear, moorings and a fixed footprint a grower cannot move. The Gulf of Maine warms faster than ~99% of the global ocean (Saba et al. 2016), at 0.03–0.06°C/yr. Over a 20-year lease: 0.6–1.2°C of mean summer warming. The spatial pattern of suitability shifts within the estuary.

20yr
standard Maine DMR lease — a capital bet on water that has not been measured yet
×3
rate the Gulf of Maine warms vs. the global ocean average (Saba et al. 2016)

Method: delta-downscaling. Prithvi supplies 30m fine spatial structure; NWA12-MOM6 / CMIP6 supplies the decadal warming trend. Fused per-pixel at 2030 / 2035 / 2040 / 2045 horizons with quantified uncertainty.

osi_trend.tif
Winner vs. loser map

Per-pixel OSI slope 2025→2045. Which parcels warm into the ideal band — and which warm through it.

lease_persistence.tif
Will it last the lease?

Fraction of the 20-year term a pixel stays ≥ Moderate suitability. The question every lease decision asks.

crossover_year.tif
When to enter or exit

The year a site rises above or drops below the suitability threshold — timing, not just ranking.

lease_rank.tif
Risk-adjusted parcel rank

Term-mean OSI minus the lower ensemble bound. Ranks candidate sites by expected suitability net of climate downside.

Darwin Oceans · Confidential05  /  20-year view
The method · tomorrow

How we will score a site in the future.

01
Mine

Data mining at scale

Every public ocean signal, plus the grower outcomes no one else captures.

NOAAMaine DMRNERACOOSSatellite SSTGrower outcomes
›
02
Fuse

One corpus

A dozen sources on one grid, in space and time.

Common gridSpace × time
›
Only the best US talent can build it.
03
Fine-tune

Adapt a foundation model

Fine-tune an ocean foundation model on that corpus. Commercial + scientific value.

AlphaEarthPrithviSamudra→ fine-tuned
›
04
Decide

Anywhere on the coast

Suitability, disease and closures off-station, served by API.

Off-stationWhole coastAPI

Proven method, new domain: the same data-to-foundation-model playbook Darwin Geospatial ran for soil, forests and habitat, now pointed at the ocean.

Darwin Oceans · Confidential06  /  Method
The roadmap

The platform is the wedge. The model is the prize.

Now · 2026

The data platform

Public ocean data in one visor. Live today.

+6 months to first ARR

Ground-truth network

Customers feed proprietary data no competitor can buy.

The prize

Ocean foundation model

Predicts suitability, disease & closures anywhere. Licensed by API.

Each stage funds the next and earns the proprietary data for the one after it.

Darwin Oceans · Confidential07  /  Roadmap
The build plan

Three technical tracks, built in parallel, all connected.

The visor wins customers, the customers feed the corpus, the corpus trains the model. Each track powers the next.

Track A

Database

The corpus no one else holds.
  • Now: public ocean data fused (994 stations, 135 leases)
  • Next: satellite feeds + captured grower outcomes
  • Then: whole-coast, ground-truthed dataset
Track B

Visor

What customers touch.
  • Now: visor + COG-tile API, live
  • Next: hosted accounts, alerts, assistant
  • Then: dashboards + predictions by API
Track C

AI models

Index today, foundation model tomorrow.
  • Now: transparent suitability + disease index
  • Next: more variables, off-station ML
  • Then: fine-tuned ocean foundation model
Darwin Oceans · Confidential08  /  Build plan
The partnership · Why us

Top executives, engineers and ocean scientists, aligned.

Chief Executive Officer

Gabriel Díaz Ireland

Founder of Darwin Geospatial. Geospatial & ML engineer building satellite-data decision platforms across soil, forests and habitat. Leads product, data and commercial.

Founding Engineer · CTO

Sandupal Dutta

Earth & planetary science and software engineering (Johns Hopkins). Owns the data-fusion pipeline, the suitability/disease models and the visor + API.

Scientific Sponsor · JHU

Anand Gnanadesikan

Senior Scientist & Professor, Earth & Planetary Sciences, Johns Hopkins University. Ocean and climate modeling. Guides the science behind our models.

Darwin Oceans · Confidential09  /  Team
The trajectory

Growers + agencies are the plan. Insurers + the model are the upside.

YearARRComposition / driversProduct & data state
Y1 · 2026$30–75KMaine co-op pilots, semi-paid3 pilot co-ops (Damariscotta, Casco Bay, New Meadows); hosted visor, accounts & alerts live
Y2 · 2027$150–350KMaine expansion + Chesapeake entry + 1 agency~30–60 grower seats + first state agency; Chesapeake on live data
Y3 · 2028$0.4–0.9MMulti-region growers + 2–3 agencies + first insurerOff-station predictions in-product; metered B2B API live
Y4 · 2029$1.2–1.8MRecurring base + agency renewals + 2nd insurer + restorationFine-tuned ocean foundation model serving predictions by API
Y5 · 2030$2–3M base / $5–8M upsideGrowers $0.6–1.2M · agencies $0.3–0.5M · insurers $0.5–1.5M · restoration $0.2–0.5MFM the licensed, high-margin asset; whole-coast, multi-region

Base case, good execution. Figures indicative. The recurring grower and agency base is the plan we raise against; insurers and the foundation-model API land late and are the upside multiplier, not the assumption.

Darwin Oceans · Confidential10  /  Trajectory
Next steps · StartBlue

Two non-dilutive tracks, keyed to readiness.

StartBlue (UC San Diego: Scripps + Rady) takes no equity. Two tracks by technology readiness: Launch for TRL 2–5, Scale for TRL 6+. We enter on the Scale track; our dilutive founding round runs separately.

Launch track
$20KNon-dilutive · TRL 2–5
What it gives
  • Scripps + Rady cohort, mentorship and the blue-economy network
  • Pilot and customer introductions across the ocean enterprise
  • Up to $20K non-dilutive on completing the program
Where we are
  • Working visor on real data, ready today
  • Maine pilot intent across three co-ops
  • Our 6-month ground-truth plan
Scale track
$150KNon-dilutive · TRL 6+ · our target
What it gives
  • Training to advance TRL, scale operations and business development
  • Deeper StartBlue, investor and agency connections
  • Up to $150K non-dilutive on completing the program
Why we qualify
  • TRL 6: an integrated prototype on real public data
  • Two regions live, COG tiles + API, science reviewed by JHU
  • The dilutive founding round runs separately, alongside
Darwin Oceans · Confidential11  /  Next steps

Let's prioritize the working coast, together.

Gabriel Díaz Ireland · gabriel.ireland@darwingeospatial.com

Darwin Oceans — a Darwin Geospatial venture · darwingeospatial.com