Boeing's ORCA was visionary in 2019. An extra-large unmanned underwater vehicle that could deploy from a pier, transit thousands of miles autonomously, execute its mission, and return — all without a crew. It was going to change naval warfare.
It's 2026. After $885 million and eight years, the GAO says it's "unclear whether the Navy will transition the XLUUV to a program of record."
And the threat moved on.
What Happened to ORCA
The GAO documented it extensively. The first prototype was supposed to be delivered in December 2020. All five were due by end of 2022. By September 2022, the program was 64% over its original cost estimate and at least 3 years behind schedule.
The first prototype finally arrived in December 2023 — three years late. As of mid-2025, the GAO reported ongoing technical challenges with autonomy, battery endurance, and navigation. The Navy expects all five prototypes by end of 2025 but still hasn't committed to making ORCA a production program.
The XLUUV isn't a bad concept. It's a case study in what happens when the acquisition timeline exceeds the relevance timeline.
Enter Ghost Shark
While Boeing was still in development, Anduril took a different approach. Ghost Shark — an extra-large autonomous undersea vehicle for the Royal Australian Navy — went from concept to three delivered prototypes in approximately three years, for roughly $100 million in R&D (split evenly between Anduril and the Australian government).
In September 2025, Australia signed an A$1.7 billion (~US$1.1B) production contract. The system is designed for ISR and strike operations at long range, with operational capability from January 2026. On time. On budget.
Anduril's Chris Brose publicly noted that the US has "spent a significantly greater amount of money on [XLUUV] than the Australian government and Anduril have spent developing Ghost Shark capability and it's further behind."
The Surface Revolution
It's not just undersea. On the surface, Saronic went from founding in 2022 to a $4 billion valuation and a several-hundred-million-dollar Navy contract for the Corsair autonomous surface vessel (reported at $392M per Saronic's announcement) — with delivery in under 12 months. Secretary of the Navy John Phelan announced it at the Reagan National Defense Forum in December 2025, calling it proof of "how we'll build a hybrid manned-unmanned fleet."
And in the Black Sea, Ukraine proved the operational concept with improvised naval drones that helped sink or disable more than 20 Russian warships — forcing the Black Sea Fleet to withdraw from Crimea to Novorossiysk. They did this without a navy, using commercial components and startup speed.
The defense startups didn't win because they had better engineers than Boeing. They won because they shipped while Boeing was still in CDR.
The Math Argument
By the time you deliver a system 5+ years late, three things have changed:
- The threat it was designed for has evolved. The undersea domain in 2026 looks nothing like 2019.
- The technology landscape has made your design choices obsolete. Battery, autonomy, and AI capabilities have leapfrogged what ORCA was built to do.
- Smaller, faster competitors have filled the gap with 80% solutions at 10% of the cost.
The Navy doesn't need one exquisite submarine drone. It needs volume. Attritable platforms it can afford to lose. When you lose an XLUUV, you lose your strategy. When you lose a Corsair, you lose a line item. Replicator was supposed to deliver this at scale — thousands of attritable systems. We got hundreds.
That's not a technology argument. It's a math argument. And the math favors speed and volume over perfection and scarcity.
Anduril is now in talks for a $60 billion+ valuation — double its June 2025 figure. The market is telling us something the acquisition system hasn't fully absorbed yet: speed of delivery is the new competitive advantage, and the companies that ship are the ones that win.
Is the traditional defense acquisition model capable of producing systems at the speed the threat requires? Or has it been permanently disrupted by the startup model?
Sources: GAO-22-105974 (Sept 2022), GAO-25-107569 (June 2025), Breaking Defense (June 2025), USNI News (Sept 2025), Reuters (Feb 2026), Naval News (Sept/Dec 2025), Saronic PRNewswire (Feb 2025)