The company building underwater drones for $50,000 Summary — Stripe's Cheeky Pint

Will O’Brien, cofounder and president of Ulysses, is building a fleet of low-cost, modular underwater drones to make ocean operations routine and affordable enough to unlock entirely new industries, targeting a 100x cost reduction over existing autonomous underwater vehicles.

The core thesis: just as reusable launch lowered the cost to orbit and enabled Starlink, drastically cheaper underwater robotics will turn the ocean from a $3 trillion economy into a $10 trillion-plus domain for commerce, conservation, and defense.
Over 75% of the seafloor remains unmapped; fewer than 100 people work full-time on undersea robotics, which O’Brien sees as a wide-open frontier comparable to the American West or early space exploration.
The company started with seagrass restoration — seagrass covers 2% of the seafloor but holds 20% of ocean carbon and supports 20% of commercially important fish stocks — but quickly realized the same robotic platform could serve defense, offshore energy, and subsea cable protection.

A cofounder met a marine biologist on a surf trip whose job involved manual underwater photography, seed planting, and water sampling; he saw an opportunity to build an “underwater equivalent of an aerial delivery drone.”
The initial task chain: map the seabed, locate flowering seagrass, collect seeds, run habitat-suitability models, then deposit seeds in restoration zones — all currently done by divers at 100–1,000x slower than the rate of loss.
Regulatory “green tape” is the single biggest barrier to seagrass projects; permitting can take years even for high-priority sites.
Ulysses still runs seagrass contracts (Virginia, Florida, California, Great Barrier Reef) but treats them as operational scaling of an already-built tech stack rather than R&D.

Leviathan: autonomous surface vehicle, 2,000 nautical mile range, carries and deploys Mako, provides wireless charging, Starlink connectivity, and future module-swapping/repair capability.
Mako: modular underwater vehicle built from “Lego-like” blocks — swappable battery packs for range, interchangeable thrust modules for speed, plug-in stereo cameras for 3D imaging, sonar nose cones for bathymetry, manipulator arms for physical intervention.
Design principle: full XYZ access — transit efficiently on the surface (low drag), drop Mako for precise underwater work, recover and recharge.
Comms constraint: acoustic bandwidth ~10 KB/s over ~5 km (dial-up speeds), so autonomy must run locally; Mako carries an onboard NVIDIA GPU (5090-class) + Intel CPU — ~100x more compute than typical AUVs — for real-time sense-making and obstacle avoidance.
Current autonomy: polygon-based lawnmower surveys with rudimentary obstacle avoidance; complex behaviors not yet needed for contracted tasks.

Generation 1 vehicle was ~80% off-the-shelf; now on Gen 5, moving aggressively toward in-house production of sensors, inertial navigation, pressure housings, and compute.
Target: a factory capable of producing thousands of units/year — one of the world’s largest underwater drone factories by volume.
Supply-chain logic: maritime suppliers build 100–200 units/year; drone, EV, and consumer-electronics supply chains build millions. Ulysses designs to the latter.
Inertial navigation, sonar, and pressure-tolerant electronics are being brought in-house to cut cost, lead time, and enable million-unit scale.
Cofounder Colm focuses full-time on manufacturing — a pattern O’Brien notes at Anduril and SpaceX/Tesla.

Offshore energy (floating wind, subsea cables) buys services: daily cable integrity monitoring, fishing-line detection and cutting, structural inspections — replacing $200k–$300k/day crewed ships.
Navies buy hardware: base Mako at ~$50k vs. $5M for incumbent AUVs; even with advanced sensors, stays an order of magnitude cheaper.
Long procurement cycles (MSAs, clearances, health-safety reviews) are accepted because won contracts are extremely sticky and hard to displace.
Department of Defense inbound interest is a new, sober, apolitical buying signal; once contracted, revenue is highly bankable.

US Navy: mine hunting/countermeasures — contracted before Red Sea/Hormuz tensions escalated.
Floating offshore wind (California): power cables attract abandoned fishing gear that entangles whales; Ulysses maps cables daily, detects fouling, and cuts lines autonomously.
Subsea cables: ~200 cables carry nearly all intercontinental data; vulnerable to anchor drags, tapping, and targeted cuts. Persistent 24/7 robotic patrols at choke points (e.g., west of Ireland, Asian straits) are only feasible at Ulysses’ cost structure.
Vision: permanent robotic “settlements” in critical waterways, analogous to persistent orbital constellations.

Irish entrepreneurial family (pub, farm, multiple businesses); grew up treating business as dinner-table conversation.
Guinness as a case study in undirected R&D: 1890s in-house lab produced Student’s t-test (small-sample statistics for hop-quality testing), the nitrogen widget (voted top-100 20th-century invention), and iconic marketing — all kept secret from competitors.
O’Brien identifies as a “pro-growth environmentalist / eco-pragmatist”: flourishing future requires both a growing economy and a biosphere that supports it.
Sees a cohort shift: founders in their late 20s, shaped by COVID, “woke era,” and a return to faith/community, building “heart-based” companies (SpaceX, Anduril, Ulysses) motivated by what the world needs rather than just what people want.
Faith community in SF has been a primary network for cofounders, investors, and early hires — a noticeable correlation among hard-tech founders.