The Navy says AI cut a 160-hour submarine-planning job down to just 10 minutes — now it's investing $448 million to go bigger

[Daniel Nenni]

• - The Navy's investing almost half a billion dollars on a new, Palantir-powered AI system for shipbuilding.
• - Ship OS sped up the workflow for submarine jobs.
• - This capability is starting with private and public yards and will steadily expand into other shipbuilding programs.

The Navy is pouring hundreds of millions of dollars into an artificial intelligence system that it says has sped up key shipbuilding processes.

In one case, the AI cut painstaking processes of submarine schedule planning — mapping out how the many pieces of construction fit together and making sure people, parts, and yard space are available at the right time — from many hours to only minutes.

The Navy is launching the new Shipbuilding Operating System, or Ship OS, as it tries to break out of decades-old shipbuilding problems rooted in outdated technologies and work practices. The service announced a $448 million investment Thursday, saying it will accelerate the adoption of AI and autonomy across the industrial base.

The Ship OS technology is powered by Palantir's Foundry and Artificial Intelligence Platform and began in pilot programs at submarine shipyards.

At General Dynamics Electric Boat, a long-time submarine yard located in Connecticut, submarine schedule planning saw a dramatic reduction from 160 manual hours down to under 10 minutes. And at Portsmouth Naval Shipyard in Maine, material review times for submarines went from taking weeks to under an hour.

The $448 million investment will go toward the submarine industrial base and then expand. It'll be deployed across two major shipbuilders, three public yards, and 100 suppliers, Palantir said in a press release.

General Dynamics Electric Boat, a shipbuilder who tested the AI pilot, saw major decreases in time for submarine scheduling.US Navy photo by John Narewski

"This investment provides the resources our shipbuilders, shipyards, and suppliers need to modernize their operations and succeed in meeting our nation's defense requirements," said Navy Secretary John Phelan in a statement.

"By enabling industry to adopt AI and autonomy tools at scale, we're helping the shipbuilding industry improve schedules, increase capacity, and reduce costs," he added, explaining "this is about doing business smarter and building the industrial capability our Navy and nation require."

Maritime Industrial Base Program, a Navy initiative to revitalize US shipbuilding and repair capabilities, and Naval Sea Systems Command are overseeing the implementation of Ship OS. Both are gathering data from multiple sources to identify where the hiccups are in submarine shipbuilding, how the processes, including engineering, can be sped up, and what specific risks can be mitigated through technology.

 

[Barnsley]

Palantir-agogo

 

[KevinK]

Hopefully it will work for ships, too, and get subs and ships across the finish line. Lots of programs die on the vine.

Opinion | America Can’t Make What the Military Needs

The U.S. defense industry has lost the ability to build quickly and effectively.

www.nytimes.com

 

[Xebec]

I know the article says 'this is only one example', but running the math - the article is implying $640,000 savings in one process justifies $448,000,000 in taxpayer spending:

i.e. If we even built 20 warships a year, * 160 hours, * $200 an hour (burden rate is lower than this..) = $640,000 saved.

To get to $64M (~13% of $448M) in savings we would need 100 more of these processes..