Akhil Voorakkara Is Building the Ocean Company One Hard Lesson at a Time

Salt water has a cruel way of turning confidence into chemistry. On an early Ulysses mission in Australia, current found an unintended path through an underwater robot. The electrical short attacked the machine's anodized surface until its face began to disappear. The team called it Benji, after Benjamin Franklin. Their customer was less amused. The machine had failed, the planting season was closing, and the next useful window would not arrive for a year.

For Akhil Voorakkara, the company's co-founder and CEO, the failure became a clean sentence: “We basically overpromised and underdelivered. It was our first reality check.” It was also an expensive organizational diagram. Four ambitious founders had moved quickly enough to win a customer, build a robot and send it across the world. They had not yet decided, with equal precision, who owned what.

A year later, they returned to the same coast. This time the system worked. That passage from dissolved metal to a functioning field mission explains more about Voorakkara than a list of roles can. His career has been a series of machines exposed to public tests: open-data prototypes, a student Hyperloop pod, delivery drones and now autonomous vehicles operating in one of engineering's least forgiving environments. The pattern is not polish. It is a willingness to let reality grade the work.

Sufficiently weird

Ulysses did not begin with a defense memo or a grand platform map. It began with a complaint in wet Scottish weather. In early 2023, engineer Jamie Wedderburn heard a marine-biologist friend describe a miserable day restoring seagrass by hand. Dozens of volunteers had worked in cold shallows, only to see rough conditions undo the planting. Wedderburn went home, researched the problem and called his former colleague Voorakkara.

The pair had worked together straight out of college at Manna, an Irish drone-delivery startup. Voorakkara handled electronics; Wedderburn handled mechanics. By the time of the call, Voorakkara had moved into consulting at McKinsey and Wedderburn into satellite work. Both missed building. They briefly considered another aerial-drone company, but the ocean problem had a stranger pull. As Voorakkara put it, “We could do drones, but underwater robots are so much cooler.”

Their first machine was barely a vehicle. It was assembled from 3D-printed parts and electronics bought online. It could not swim. It leaked. Yet it could inject sesame seeds into sediment, which answered the narrow question in front of them: a repeatable mechanism might replace some of the human labor involved in planting seagrass. The team did not pretend to understand the biology. “None of us are marine biologists,” Voorakkara said. They asked restoration experts early, then built around what those experts knew.

“We can literally build the ocean company. I'm making the group chat now.”Akhil Voorakkara, recalling the founding team

The group chat was the first operating system

Voorakkara knew how to assemble this particular crew because their histories already overlapped. He had been friends with Colm O'Brien since they were 12. Colm's father ran a coding club for children in Dublin; both boys later taught there. At Dublin City University, Voorakkara helped found Éirloop, Ireland's student Hyperloop team. Colm later led it. Will O'Brien entered their orbit by asking a short, useful question: how could he help?

Éirloop was selected as one of 20 finalists from more than 700 entries for SpaceX's 2018 Hyperloop Pod Competition. Voorakkara served as head of electronics. The achievement sits in a longer trail of student projects. His public record includes an open-data service called LifeStats, an accessible-parking tool, a quieter-route navigation concept based on noise monitors, and a computer-vision drone designed to follow cats. The projects mixed hardware, data and a slightly mischievous sense of utility.

When Wedderburn's seagrass idea arrived, Voorakkara saw complementary abilities rather than four copies of himself. Colm had worked in Formula One aerodynamics. Will understood sales, government and growth. Jamie could turn a biological task into a mechanical system. Akhil could bridge electronics, commercial logic and the people involved. Excitement from the others became its own evidence. The problem was strange, but the group was credible to itself.

The season that would not move faster

The early experiments produced enough evidence to attract partners and capital. In late 2023, Akhil and Colm took a week-old seed-injection device to Western Australia. Researchers had doubts that the young company would even arrive. It did, and the underlying injection method worked. Ulysses raised a $2 million pre-seed round, established itself in San Francisco and built toward a complete robotic system.

Then came Benji. In software, teams speak casually about shortening feedback loops. In seagrass restoration, biology and weather control the calendar. Missing a species' seeding window can turn one mistake into a year of waiting. The damaged robot exposed more than insufficient electrical isolation. It exposed a company that had sprinted ahead of its internal agreements.

Voorakkara's useful move was not to convert the failure into folklore too quickly. The founders spent months working through it. They clarified their lanes. They dealt with the distance between the promise and what the product had done in salt water. When they returned to Australia the next year, the successful deployment felt like redemption because it closed a loop that no motivational speech could.

That patience coexisted with faster progress elsewhere. Ulysses closed an $8 million seed round in May 2025, moved into a larger San Francisco headquarters and expanded from five people to 16. The workspace itself reads like the company: ocean-blue walls, sharks, machine tools and an industrial inflatable pool where a Mako vehicle can be tethered to a rugged laptop and controlled with a PlayStation controller. Serious machines often grow from ordinary interfaces.

From planting to presence

Seagrass was a narrow first job with unusually demanding requirements. A restoration robot must sense its surroundings, navigate moving water and physically interact with the seabed without destroying the thing it is meant to help. That “read-write” capability became the base of Ulysses' larger platform thesis. Swap a payload and the same core vehicle can survey a reef, inspect infrastructure, map the seabed or support a maritime-security mission.

The economic argument is as important as autonomy. Ocean work still depends heavily on crewed vessels with day rates in the tens of thousands of dollars and specialized vehicles priced for rare, extreme missions. Ulysses builds smaller modular machines intended for higher-volume production. Its Mako is designed to carry interchangeable sensors and tools, run missions autonomously, and work as part of a broader launch, recovery and recharge system.

By April 2026, the company said it had tested five generations of underwater vehicles and generated more than $5 million in revenue. Its named customers and partners spanned the U.S. Navy, the Australian government, the Great Barrier Reef Foundation, The Nature Conservancy, the Virginia Institute of Marine Science, Mote Marine Laboratory and Florida's environmental agency. A $38 million Series A led by Andreessen Horowitz, with Booz Allen Ventures and Harpoon Ventures participating, was meant to scale Mako production and move Kraken and Leviathan from prototypes toward production.

A coastline, not a demo

Voorakkara's stated horizon has remained longer than a product launch. When Ulysses first emerged from stealth, he said that within five years he wanted the company managing hundreds of kilometers of coastline and giving organizations such as NOAA and the U.S. Coast Guard more efficient tools. The company now describes its destination as networked fleets operating above and below the surface.

That ambition has widened the company's work from ecological restoration toward infrastructure, science and defense. The common denominator is access. “The ocean is the last great frontier,” Voorakkara has said. Land, air and space have gained cheaper sensors, vehicles and computing; much of the ocean remains expensive to reach and hard to monitor. Ulysses is betting that lower-cost hardware and onboard autonomy can change how often people are able to act there.

The bet still has salt water in it. Vehicles must survive pressure, corrosion, currents and the awkward logistics of leaving shore. Customers must trust machines whose mistakes may disappear beneath the surface. Production has to turn prototypes into repeatable tools. The first chapter of Ulysses offers no shortcut around those facts. It offers a working method: recruit people whose skills interlock, ask specialists before certainty hardens, build a cheap test, admit what failed, and return when the season allows.

Voorakkara's story feels personal because the machines keep carrying earlier parts of his life. The coding club reappears as a co-founder relationship. The Hyperloop team reappears as comfort with public engineering trials. Drone delivery reappears as autonomy in a harsher medium. Consulting reappears in the effort to make an expensive industry legible. Even Benji remains useful, a corroded reminder that a founder can move quickly while the ocean keeps its own clock.