Software-Defined Warships: The Navy’s Digital Future of Necessity
In 2024, U.S. Navy warships received their first ever over-the-air software updates in combat. In the Red Sea, these upgrades improved the Aegis combat systems’ capabilities against cruise and ballistic missile attacks by Houthi militants.
Over-the-air software updates are nothing new. In the commercial world, they happen every night — on your laptop, phone, car, or even refrigerator. These updates work on any hardware that runs software, but have the most impact on software-defined platforms. Here, software sits at the center of the design, and can dynamically manage sensors, computing power, or weapons.
This was a departure from how traditional industries viewed platforms — change came in the form of physical alterations, which can be slow and expensive to develop and implement. For example, a car with defective brakes would require an automaker to issue a recall costing millions of dollars and tens of thousands of maintenance hours to replace physical parts. Meanwhile Tesla, whose cars are software-defined, “recalled” millions of its vehicles in this situation using an over-the-air update. Likewise, in the Russo-Ukrainian war, necessity has driven Ukrainian technologists to push nightly software updates to their frontline technology. Thus, the primary advantage of a software-defined platform is speed of upgradability.
In both the marketplace and in war, the speed, ease, and cost of improvement dictate success or failure. John Boyd called this the “observe-orient-decide-act (OODA) loop.” Digital technologies have a tighter OODA loop. Consequently, warfighting platforms will rapidly increase in capability due to the rate of improvement of software. This “first derivative phenomenon” may sound technical, but is quite simple: Production and update timelines for software are far shorter than for hardware. War is now software-defined.
These are issues we both deal with in our day-to-day work. One of us, Artem, is working as a surface warfare officer in the U.S. Navy to realize this future. The other, Austin, is building this technology in the private sector for defense and commercial markets. Our service as naval officers gives us a high level understanding of the technology and policy, and our current roles show us how seapower is delivered at the tactical level. To be sure, Austin has a commercial interest in this space but thinks about this technology every hour of every day, as anyone who has started a company and bet their career must. More importantly, he sits alongside engineers from commercial robotics, automotive, and autonomous vehicles — builders unconstrained by the cultural inertia in naval requirements, and only driven to deliver a working, software-defined ship. Despite sitting on opposite sides of the metaphorical government-industry contracting table, we found it easy to find a common argument.
It’s time warships move into the software-centric reality of the 21st century. Should their designs continue to center on humans and hardware, they will not evolve fast enough and – as a result — will become steel tombs.
Software-Defined Warfare: What It Is and Why We Care
Although software is proving its primacy on the battlefields from the Black Sea to the Red Sea, legacy surface fleets remain lagging adopters. Software evolves overnight while warships evolve over 40-year capital cycles – by their very nature, they are anchored in decades-old hardware. For example, Industrial Revolution-era advances in steel and steam engines, available in the 1880s, took until the military buildups of the 1910s and 1930s to fully arrive in navies.
Software will generate similar capability increases continuously, because the “learn-update-employ” cycle is so much faster. Warships will not wait for a lengthy post-battle refit — they will be updated days or hours after an engagement. Moreover, software updates compound with steady hardware improvement and tactical innovation, thereby producing exponential increases in lethality.
Figure 1: Surface warship capability increases since the industrial age (source: authors)
Digital warfare, fought from software-defined naval platforms, offers the only rapidly evolving toolkit that can match the blistering pace of technological change — the U.S. Navy’s warships can and should be software-defined. We’ll digest this hypothesis platform by platform. We highlight ongoing digitization efforts, suggest some more quick wins, and analyze potential next steps to keep America’s surface fleet operational and lethal for decades to come. The impetus for immediate change cannot be greater — a new fleet will not be available to meet the 2027 challenge, but software updates can be ready in time to make America’s Navy exponentially more lethal.
As the United States pursues speed, it should play to its strengths. America has the best software engineers, the world’s most valuable tech companies, and the intellectual property that other economies pay for. Just as industrial warfare was uniquely suited to America’s strengths in World War II, a digital war is one that the United States is well-equipped to win. America is good at software. The U.S. Navy should be too.
Yet bringing software-defined warships to the fleet will be challenging. Even with thoughtful hardware upgrades and deep tech refreshes, the bones of ship-board computing — including the critical power, cooling, and wiring to support new shipboard servers and processors, and by extension compute-intensive software such as artificial intelligence — are brittle.
Computers on Keels
Future warships should be software-defined. Just as Tesla offered the first “sophisticated computer on wheels,” new classes of warships will be much more sophisticated computers on keels. To achieve this end state, software-defined warships will need easily upgradable computing and networking hardware, standardized software architectures, and a digitally savvy acquisitions workforce.
First, software-defined warships should have powerful computing and networking hardware such as graphical processing units, central processing units, routers, switches, and more. These pieces of hardware should be easy to install and upgrade to account for the rate of change in commercial technologies. Moreover, shipboard compute hardware should prioritize commercial architectures such as x86 or ARM, as opposed to custom chip sets to enable any American company to build software for America’s warships. Vertical integration between hardware and software, like Apple, may seem preferable, especially if you are a prime defense contractor, but it rarely achieves the same effects in military systems as building a common platform for a wide range of developers to build on top of, like Android.
To further leverage America’s commercial leadership in software, software-defined warships should have standardized and modular software architectures for their information technology and operations technology systems. In other words, the software and data layers of warships’ overall technology stacks will become the “digital keel” of their present and future capabilities. Leveraging commercial standards and open-source tool sets will be critical to shortening the time between updates. Just like in World War II when manufacturing standardization enabled American shipyard workers to produce Liberty-class ships in just five days, digital standardization will open the U.S. Navy to a wider commercial market, and enable American software engineers to produce software updates for its warships in even shorter cycles. Crucially, those software engineers are essential both in and out of government.
Lastly, as warships become increasingly digital, they will simultaneously become increasingly vulnerable to cyber threats. On Sept. 21, 1997, the USS Yorktown, a Ticonderoga-class guided missile cruiser was serving as a test bed for the Navy’s “smart ship” program, when it suddenly lost power, and remained “dark” for over two hours. The culprit — a divide-by-zero arithmetic exception in the engineering control system. In 2024 and beyond, hackers could force similar errors, neutralizing warships at sea without firing a kinetic shot. Accordingly, the adoption of cyber principles such as zero-trust, will be critical to hardening digital infrastructure at sea. Equally, some systems, such as navigation, engineering, and combat systems may remain air-gapped, or at best communicate with the outside world via a one-directional data diode — new data is extracted, but systems updates are performed manually vice over the air.
Tomorrow’s Digital Warships
Unmanned ship design offers an early view of the software-centric platforms that will enable digital warfighting at sea. Without crew, these vessels should have vast amounts of sensing, compute, and software — enough to replace the crew and pass actionable information to off-ship humans through a narrow data pipe. In 2022, China offered the first software-defined warship combat system architecture, which may have been for the JARI unmanned surface vehicle. Ships that have this architecture can change their tasks or mission focus with a quick upgrade, perhaps even over-the-air.
Although most of the attention for new ships has centered on design maturity rather than digital leap-ahead, DDG-X, the Navy’s future destroyer, offers a view into how Navy leaders are internalizing the power of software-defined warships. Chief of Naval Operations Adm. Michael Gilday, discussing DDG-X alongside existing fleet capabilities, explained that more compute and software would sit in ship server rooms, rather than on the weapons themselves. This design choice allows quick upgrades across hardware resources. Additionally, sailors can recommend, choose, and download apps on ships, a software update familiar from commercial practice. The Navy’s research and development budget justification for DDG-X design emphasizes “flexibility to rapidly and affordably upgrade.” However, if Navy engineers ignore software architecture choices in the design phase, the service risks the desired flexibility and upgradability end states.
The Navy’s medium unmanned surface vessel program is already headed in a software-centric direction. Although early prototypes were software-skinny retrofits, DARPA’s No Manning Required Ship program started from a cleansheet, and keeps computers, not humans, at the center of its design. The Navy’s unmanned ship program office maintains an unmanned software architecture, with dedicated software interfaces defined for navigation, sensing, communications, engineering operations, auxiliary systems, payload management, and more. All these logical tasks will be performed by software, as by humans on manned ships. All of them will be adaptable and reconfigurable once digitized.
As software becomes core to warship design, operations, and maintenance, the automotive industry will remain a guiding light. Tesla, updating wirelessly from the driveway while you sleep, may be the standard bearer, but all cars are getting a lot more software. Over the next decade, industry trends suggest software could become over a third of the value in the automotive industry. Today, it is still a sliver.
Tonight’s Digital Fight: The 2027 Upgrades
If cars still have a long way to go down the digitization path, ships are decades further behind. The average U.S. car has been on the road for 12.6 years, making them four years older, on average, than the oldest Apple AirPods. Ships, on average, have been at sea for 22 years, and many U.S. warships were built in the 1990s, all before the oldest MacBook. Although the structural bones of naval ships are old, the U.S. fleet carries advanced digital technology. The challenge is how long it takes to get those advanced systems onboard each warship. Today, the U.S. Navy’s warships are the antithesis of easily upgradable. Seemingly digital systems, such as warships’ afloat network and computing environment, take months to install, and software updates arrive at the pace of hardware changes. As a sage robotics chief technology officer once told us, “early hardware decisions cast long shadows.” In naval architecture, these shadows often last 20 to 40 years.
The U.S. Navy has so far not adopted software-defined principles as part of the design of its warships. In fact, its warships — the most advanced in the world — are largely no different than those built to fight the Soviet Union in the late 1970s and early 1980s. Some of those warships are still active today, such as the aging Ticonderoga-class cruisers, whose AEGIS combat systems are hardware-centric and do not have software-focused metrics, dollars, or design. “Prime” vendors’ preference for vertical integration between hardware and software– such as AEGIS – further complicates the path towards software defined warships.
But many things can be done to drive warship digitization today. One priority on legacy warships is upgrading sensors and systems to keep, process, and share more of their data. Efforts like this enable fusion, whether it be for target tracking or maintenance insight. Another effort seeks to balance the compute load between ashore cloud and afloat servers — an area where many of the upgrades can already be digital. And yet another is the constant struggle of bandwidth management, using software to help ship and satellite antennas prioritize the data packets passing through their narrow pipes.
Digitizing U.S. warships is an urgent priority. The Navy cannot wait a decade for the Integrated Combat System , which is set to finally decouple software from hardware. The chief of naval operations outlined her vision for a high-tech Navy in Project 33. Alongside new platforms, such as software-defined unmanned ships, she emphasizes reducing maintenance delays. On older ships, maintenance should wait for specialized hardware technicians and replacement part backlogs. Getting to a fleet that can stay agile with rapid software updates will keep us ready for 2027, 2028, and whatever comes next. When the balloon goes up, ships can’t turn back to port for another upgrade — they need updates, tweaks, and agility that can arrive in their server room in seconds.
“The [littoral combat ship] is back,” Secretary of the Navy Del Toro stated emphatically from an event last month. His enthusiasm for the Navy’s most troubled (funded) ship program arose for one reason: upgrades. After experimentation, littoral combat ships now equipped with MK70 missile launchers can help address the Navy’s missile magazine challenges in the Indo-Pacific. However, this upgrade took years. Hardware upgrades, especially when the hardware baseline is not ready for them, take vast amounts of time and resources. Software upgrades on digital platforms – where the hardware is ready to interface with evolving software – takes minutes. The Navy is right to celebrate ship upgrades, but needs to push this culture deeper into acquisition and requirements offices where platform hardware choices lock in the “upgradeability” for decades.
Shipbuilding Woes versus Software-Defined Ships
While warfighters focus on upgrading the existing fleet for tonight’s fight, requirements and program officers have a generational opportunity to shape the first software-defined ships with industry. Not only will these OODAloop-closing warships be far more lethal — they will be much easier to build.
Applying the principles of consumer and industrial robotics to shipbuilding will deliver a producible, upgradeable, and highly reliable warship. Because hardware will stay on the vessel for a long time, early engineering should focus on designing systems that will accept many future software upgrades, even when all the future requirements are not known a priori. Much of the hardware should still be flexible, so the design shouldn’t include only bespoke, high-lead time components.
Today, most major components on U.S. warships have 18-month lead times because they are only produced for U.S. warships by specialty defense manufacturers. These bespoke components are also very expensive. During installation, bespoke components require much more specialized labor — the number-one cost in shipbuilding — for wiring, calibration, and fine-tuning of interfaces.
Software-defined ships will be assembled from commercial hardware supply chains, where everything is already mass-producible and interfaces are standardized. As more of the value of the ship becomes software and replaceable hardware, ships will become more attritable, or risk-worthy, too.
Lastly, the core systems of the ship, such as power generation, cooling, and other critical hull, mechanical, and engineering hardware — these should be very close to 100 percent reliability. The rest of the software-defined ship — the compute, the upgrades, the software — will all rely on the core hull, mechanical, and engineering systems. The best software won’t make a difference if the hardware is broken.
Full Steam Ahead
As an ambitious new executive branch looks for opportunities to reinvigorate U.S. naval power, software-defined warships have never held such tangible promise. With tech-savvy leaders in the Department of the Navy, software can eat even cold, grey, steel ships. Pentagon leaders should empower the clean sheet innovators in uniform who can help sell the vision of this technology inside the building. They can promote the civilian leaders who understand it to drive change for decades. They can give both budget authority. Moving real programmatic dollars will send industry the signal that government customers need this capability. Crucially, this does not mean only adding to the budget, but also making thoughtful divestments from legacy capabilities that are no longer suited to an era of software defined warfare.
Congress has an important role too in digitizing the fleet. Congressional experts have long scrutinized expensive shipbuilding investments and will treat software-centric ships no differently. Of late, the frigate and littoral combat ship have come under withering review for the lack of design maturity. If “tech bros” appointed to Pentagon jobs arrive on Capitol Hill with new software-first destroyer designs, they will face the same questions as any new shipbuilding program. Sen. Roger Wicker’s FORGE Act offers an example of how Congress can facilitate adoption of commercial technology in the Defense Department. A similar bill to facilitate and incentivize the digitization of large capital assets such as warships would be a welcome addition to the next National Defense Authorization Act. Appropriators can de-risk some of the authorizers’ ideas by committing one of the U.S. Navy’s aging Ticonderoga-class cruisers to serve as a “digital test ship” to inform future software-defined warship designs.
Today’s warship design choices cast decades-long shadows. But today’s, unlike last decade’s, offer the opportunity to bake software into the heart and bones of warships — this is the only way U.S. warships will evolve with the demands of this dangerous decade and those that follow. China is moving out on software-defined warships. It’s time the U.S. Navy did too.
Artem Sherbinin is an active-duty surface warfare officer currently serving as chief technology officer at U.S. Naval Surface Forces Command’s Task Force Hopper. He holds degrees from the U.S. Naval Academy and Georgetown University.
Austin Gray is co-founder and chief strategy officer of Blue Water Autonomy, a venture-capital backed defense technology shipbuilder. He previously worked in a drone factory in Kyiv and serves in the U.S. Naval Reserves as an intelligence officer. He holds degrees from Davidson College, Massachusetts Institute of Technology, and Harvard University.
The authors deployed together to the Western Pacific as part of the USS Theodore Roosevelt Carrier Strike Group in 2020.
The views and opinions expressed in this article are their own, and do not reflect the views or official policies of the U.S. Navy, the Department of Defense, or any part of the U.S. government.
Image: Navy Petty Officer 1st Class Gregory Johnson via Department of Defense
