The Navy Just Bet $300 Million on Robots That Crawl Its Warships

A $71 million contract was put in place by the U.S. Navy six months ago with a Pittsburgh firm that has robots to scale the hulls of warships. The Navy has now raised the cap on that arrangement fourfold, to $300 million. With its new accord, Gecko Robotics has secured what is one of the biggest robotics inspection deals the service has ever put on paper. The company, worth $1.25 billion last year and born in a college dorm, also has an objective to live up to: the Chief of Naval Operations is looking for 80 percent of the fleet to be deployable by 2027, something it is not close to being at present.

What the Navy Actually Bought

Under an indefinite delivery, indefinite quantity contract with the Navy and the General Services Administration, the $300 million figure is a cap and not a disbursement. The Navy will put in for inspections when the need arises. Back in the spring the original $71 million award was made public to cover Pacific Fleet vessels including destroyers, amphibious warships and littoral combat ships. On October 20 the company announced the expansion, putting it down to the fleet's maintenance backlog as opposed to any research initiative.

Gecko's product is data. With its wall-climbing robots, drones and stationary sensors, the firm can get ultrasonic thickness and other kinds of readings off hulls, decks, welds and components at a density that eludes a human inspector with a handheld gauge. That information goes into Cantilever, the software platform Gecko dubs "The Operating Platform for the Built World." From there Cantilever builds a model of the ship's structural health for the repair planners, indicating what steel has thinned, which welds are questionable and what requires attention first.

The Robot That Reads Through Paint

For shipyards the most important part of the deal is a system known as Komodo. Warship decks are covered with a heavy non-skid coating and, up to this point, the only means of seeing what the steel was like below it was to grind the coating away and have a look, which usually happened once a ship was already in drydock and the maintenance clock had started. Komodo scans through the coating without taking it off and puts out a complete deck heatmap of corrosion and material thickness.

In practice the Navy can now make its repair plans ahead of a ship's arrival instead of waiting to see what needs doing once it is there. According to Gecko, that can mean as much as $8 million in savings on repairs for a given ship, structural problems pinpointed up to 50 times faster than by the old methods, and inspection timelines reduced from months to days. The company puts forward those numbers as ceilings, not averages, in the same release as the contract announcement. The Navy has not put out any figures of its own as to what the robots have turned up to date.

A Dorm Room Company With a Billion Dollar Valuation

Defense was not where Gecko began. Co-founder and chief executive Jake Loosararian made the first robot for the inspection of power plant boilers, and for ten years the company's business was with utilities, steel and paper mills, refineries and the like, places where a boiler tube can go bad and put a facility out of commission for weeks. Today its roster of customers and partners has expanded to include the Navy, the Air Force, Anduril and Siemens as well as Fortune 100 industrial companies; the defense page on the company website touts work on everything from submarine hulls and ballast tanks to missile silos and the airframes of military cargo aircraft.

The funding came in its wake. In June 2025 Gecko raised $125 million in a Series D round at a $1.25 billion valuation, with Cox Enterprises at the helm and Founders Fund, Y Combinator, USIT and XN coming back on board. The Navy deal is the first public indication of how that money is being put to use: to make the transition from a firm that inspects a few hundred industrial assets a year to one capable of putting robots on a fleet.

In making the announcement, Loosararian was direct about the strategy. "The conversation around AI and robotics in defense is no longer about experimental technology," he said. "It's about delivering measurable and impactful returns." He went on to say that "the U.S. Navy is demonstrating to the world how to deploy technology at scale," and that the company was "proud to give our servicemembers the tools they need to stay ahead of emerging threats."

Why 80 Percent Is So Hard

What is at the heart of the contract is readiness, that is to say, how many ships are in a position to deploy. Admiral Daryl Caudle, the Chief of Naval Operations, has an 80 percent target to hit by 2027 that he took on from his predecessor, Admiral Lisa Franchetti. The fleet is nowhere near that mark. Over the ten years up to 2020 amphibious ship readiness fell to 46 percent, and it has since gone lower, to 41 percent as of August 2025; for more than five months that year the Marine Corps had no expeditionary unit deployed.

Ships have a way of vanishing when they go in for maintenance. A destroyer put in a yard for its scheduled availability will often be there longer than planned once inspectors come across corrosion or cracked welds that were not in the budget, a category the Navy puts under the heading of growth work. Each week of that is a week the vessel is not at sea, and the shipyards are already at capacity. The Navy's thinking is that if it can learn the state of every deck and hull in advance, it can have the steel and crews in place and take the surprises out of the timetable. According to Gecko's own defense page, its inspections do just that: they cut inspection span by 90 percent and bring the maintenance backlog down from two weeks per ship to zero.

What Happens When Fleet Maintenance Becomes a Software Business?

What Silicon Valley sees in this contract is a template. The defense tech narrative of the past ten years has been dominated by new weapons: drones, missiles, autonomous boats. Gecko's arrangement is of a different order; it deals with the unglamorous middle of the Navy budget, the operations and maintenance money required to keep existing ships usable. If a robotics firm can earn a $300 million ceiling by making maintenance predictable, then the same logic holds for the Air Force cargo fleet, the Army vehicle depots and every drydock in the shipbuilding industrial base. Founders and investors who have shunned inspection robots as dull will want to take another look, and the big names in industrial software can be expected to notice a startup building the operating system for a market they considered their own.

Then there are the numbers to consider. All the figures in the announcement, from $8 million per ship to 50 times faster, are the vendor's own; the Navy has put out no independent review of what the first six months found. An IDIQ ceiling is not revenue. The way the contract is written leaves the Navy free to order much less than $300 million if the early results are wanting. And there is the matter of the workforce. At the shipyards, hull inspectors and non-destructive testing technicians are skilled trades, and some of them will read a system that does their job 50 times faster as a plan for layoffs, however the company frames it.

Where controversy is to be found, it is over dependence. Critics of the Navy's haste to go commercial say that giving one venture backed startup the structural data of the Pacific Fleet makes for a vendor that is hard to replace and a dataset any adversary would covet. Their case is met by those who say the only other option is the status quo, in which the Navy learns the condition of its ships by cutting into them in drydock, and that a startup with investors to answer to has every incentive to deliver. On one point both sides agree: the Navy has waited too long to modernize the way it inspects steel.

The signals to watch are concrete. Will the Navy publish readiness or drydock turnaround figures that can be tied to robotic inspection, which would turn the vendor's claims into evidence? How much of the $300 million is actually ordered over the coming year? Does the Komodo approach spread to aircraft carriers and submarines, where the coatings are thicker and the stakes higher? And does a rival appear? In a field as small as inspection robotics, a $300 million ceiling is the kind of number that creates competitors.

The Fleet That Needs Fixing First

Sparknify has been following the Navy's move to autonomy from the water side. In The Next Frontier of Autonomous Vehicles Isn't on the Road. It's at Sea. we looked at the uncrewed ships being built by startups and the Silicon Valley capital that funds them. Gecko's contract is the other half of that story. For the Navy to put an autonomous fleet into the field, it first has to keep the crewed vessels it already has out of the repair yard, and the means to do so is not a new hull but a robot with a very good sensor and magnets on its wheels.

The contract also serves as a reminder of how things go in defense, much like every other market Silicon Valley has entered: the money tends to arrive first for the problem nobody wanted to look at.

The Navy has opted to let a robot tell it where the rust is instead of finding out in drydock; the next two years will show whether that $300 million call was well made.

Subscribe to Sparknify for Silicon Valley tech news and the people behind it.