Soaring demand for rocket motors used to power missiles and other weapons has spurred new thinking about supply chains. Seeking big returns, Silicon Valley-style startups are now taking on defense companies that have long dominated the industry, pulled into the competition by a need for production speed, high volume and lower costs, according to 10 industry executives, experts and U.S officials interviewed by Reuters.

The U.S. has plowed through over 50,000 rockets, missiles and other projectiles propelled by rocket motors since the Russian invasion of Ukraine in 2022 through the U.S. attack on Iran, Pentagon data shows. Washington is setting aside $53 billion and simplifying procurement rules to increase critical missile and rocket production.

CEOs from Lockheed, Boeing and Raytheon parent RTX, among the largest military contractors, have warned that solid rocket motor shortages were hurting missile production.

Now, defense entrepreneurs must prove they can deliver. Pleasing the Pentagon brings huge benefits, including contracts with a government agency that has an annual budget of more than a trillion dollars and a seal of approval other governments want to see before buying from new contractors.

But challenges are ahead. All the new entrants will need to produce enough of the new weapons to meet growing demand. Many new entrants are making rocket motors for existing missiles, some are making the entire missile, but none of the companies have scaled up production to replace legacy contractors.

Legacy solid rocket motor makers Northrop Grumman and L3Harris said they have been pushing their own research and development to pull these new technologies like 3D printing and new mixing technologies.

FROM STEERING CARS TO STEERING MISSILES

California-based Castelion, which makes solid rocket motors and hypersonic weapons, turned to the auto industry for sophisticated electronic components used in advanced driver assistance systems and electric vehicles to help steer its missiles to targets.

These auto industry processors, known as Field-Programmable Gate Arrays, can be bought at a tenth of the cost and obtained six times faster than comparable versions used in the aerospace industry, Chief Operating Officer Sean Pitt said.

The oil and gas industry has been another important supply chain resource for Castelion. Rather than sourcing high-pressure metal tubes from aerospace vendors with long lead times, the company is using high-temperature, stress-rated precision machined tubes used to help crack open rocks in the fracking process.

These tubes are built to handle heat and pressure levels comparable to what is required for a rocket motor, but are sold by far more vendors, at lower prices, than the aerospace industry equivalent, Pitt said.

Castelion, recently valued at nearly $3 billion, has won big Pentagon contracts to make over 500 hypersonic weapons.

Mixing rocket motor propellant is another area of innovation for startup Anduril. The company, among the more successful recent defense industry entrants with several billion dollars of contracts under its belt, is using a pharmaceutical industry technique to mix chemicals used in rocket motors.

Anduril, valued at $61 billion, has purchased Colorado-based FlackTek’s bladeless mixers capable of processing multi-hundred-kilogram propellant batches in minutes rather than hours.

Anduril says the machine delivers more than a tenfold increase in production throughput compared to its previous mixing systems. The bladeless mixers produce more than 24 times the output of conventional industrial mixers, which are more like a giant kitchen mixer with paddles that require time for cleaning.

The same bladeless centrifugal technology is used to produce precision compounds including liposome-based cancer treatments — applications where batch consistency and contamination control are as unforgiving as in weapons production.

Still, challenges lie ahead for new entrants trying to break into the solid rocket motor business.

Tom Karako, director of the Missile Defense Project at the think tank Center for Strategic and International Studies, said they include “the painstaking, multi-step manufacturing process of casting, curing, baking, x-raying and sanding that solid-fuel rocket motors require — followed by rigorous inspection.”

Curing ovens and X-ray equipment remain a bottleneck for the industry, Karako added.

3D PRINTING TO CUT PRODUCTION TIMES

Innovations have already had a dramatic impact on arms production. A 2024 case study from traditional rocket maker Northrop Grumman estimated that replacing conventionally machined metal tooling with 3-D printed polymer tools reduces the amount of time it takes to create a production line from roughly a year to about six weeks, enabling a new rocket motor to be developed from scratch much more quickly.

X-Bow Systems is a New Mexico-based company that specializes in low-cost solid rocket motor (SRM) production through a process that utilizes 3D printing of propellants and motors.

The technology has the potential to reduce the time and cost of SRM production drastically. X-Bow has also said it can shorten the creation of a new production line — which makes 3D printed motors — from a three- to six-year timeline to about 12 months. X-Bow already has a $191 million Pentagon contract for hundreds of solid rocket motors.

Texas-based Firehawk Aerospace, founded in 2020, also uses 3D printing to manufacture SRMs at a fraction of the cost of traditional production methods.

Firehawk says its manufacturing process cuts rocket fuel production time from up to 60 days to just 7 hours, at one-tenth the traditional cost.

It enables custom designed missiles to be test-ready in a matter of months. Firehawk is backed by funding from venture capital firm 1789 Capital, a fund where President Donald Trump’s son is a partner.

Government purchasing patterns will always be a limiting factor for startups. The Pentagon has traditionally bought rocket quantities annually, resulting in unpredictable demand shifts from year to year.

“How can we get good multi-year agreements that don’t roll off when administration changes?” said Lukas Czinger, CEO of Divergent Technologies, which makes parts for missiles. “That’s what businesses need to perform at low cost.”

The testing, which involved the variant of the future advanced reconnaissance vehicle, or ARV, equipped with a 30mm autocannon, took place in January and February at Camp Pendleton, California, according to a recent Marine Corps news release.

Prototypes by two designers, Textron Systems and General Dynamics Land Systems, underwent the water entry, swim and firing drills ahead of an upcoming competition period set to end with a down-select to one company around 2030.

In an exclusive interview with Military Times this month, Phil Skuta, GDLS’s director of Strategy and Business Development for the Marine Corps and Navy, described some of the challenges the new vehicle must overcome, including high-speed water entry and firing the autocannon on the water, both of which can involve weathering large and unpredictable waves.

He also described a new feature GDLS is asking the Marine Corps to evaluate: an “autotrim function” that would allow the ARV’s crew to outsource some of the work of driving the vessel in the water.

“You can set the course, the azimuth that you want to follow, and the vehicle will automatically stay on that course,” Skuta said. “So, we’re starting to build in a level of automation so that, while they’re doing, say, a long water-borne movement, that’s just one less thing the crew has to think about.

“They can think about the other activities involved in not only operating the vehicle but most importantly focusing on the enemy, so they don’t have to focus so much on, ‘Am I going in the correct direction? And is the vehicle being stable in the water?’”

Skuta said the function was somewhat similar to an aircraft autopilot function, but also like cruise control and automated lane-keeping in a ground vehicle.

“If you’re out of the lane, you might get a little buzz, so to speak,” he said. That’s more what this is right now — the first step.”

That’s particularly useful, from GDLS’ perspective, because the ARV’s swim system, separately from the ground automotive system, is piloted via a “small joystick” by a driver looking at internal screens, underscoring what a chore it is to do manual course corrections. The feature could also help the vehicle adapt to the increasingly automated future, Skuta said.

“We’re demonstrating by [incorporating] the automated trim course and azimuth function that we can eventually, in anticipation of future requirements, put more automation into the system, as well as robotic controls,” he said.

The ARV is expected to come in three variants: in addition to the ARV-30 autocannon platform there will be a Command, Control, Communications and Computers/Unmanned Aircraft Systems model, or C4/UAS, which has previously undergone testing; and a logistics variant, which will be featured in upcoming swim tests, Skuta said.

Textron and GDLS are both set to contract with the Marine Corps as soon as this fiscal quarter for 16 additional prototypes that will enable testing in coming years ahead of the final down-select.

Textron’s vehicle, which it calls the Cottonmouth, has six wheels compared to GDLS’ eight.

During the recent evolution, Skuta said, shoreside testing included a bilge pump demonstration, in which the ARV-30 was flooded with water, triggering safety sensors that automatically pump it out “like a water fountain.”

Water safety has been an increased focus since a July 2020 assault amphibious vehicle sinking off the coast of Pendleton, resulting in the deaths of nine service members. A malfunctioning bilge pump was found to have contributed to the tragedy.

Once in the water, the new vehicle was put through its paces on a five-kilometer swim course to test its maneuverability and stopping distance in calm conditions.

“You’re just kind of pivoting in the water to show the characteristics of how you can do very tight maneuvering in water spaces,” Skuta said. “And you know, that’s pretty important, because the Marines will find themselves crossing rivers and water obstacles while they’re in combat zones. And they need to have a good appreciation for how tight a turning radius the vehicle has in the water.”

Crews also practiced moving the turret around as if preparing for live fire on the water, which affects the vehicle’s center of gravity, he said.

“In a river-crossing scenario, [if the] enemy presents themselves on the far shoreline the Marines would fire from the vehicle in the water,” Skuta said. “So, that gives a good indication of how the turret and the fire control system will stabilize on a target while they’re in that water environment.”

Another tested skill, high-speed water entry, made for one of the “more fun” events to observe, he said.

The ARV-30 entered the water at progressively higher speeds, topping out at 25 miles per hour and sending a massive water plume into the air.

Notably, for all the testing with GDLS and Textron, Marines from the amphibious vehicles community were able to observe and monitor testing, Skuta said.

While the Marine Corps has not released any specific feedback from the demos, he said the onsite response from the Marines was encouraging.

“A lot of smiles, a lot of thumbs up from what they saw,” Skuta said.

The sea service reached an agreement with the Pittsburgh-based Gecko Robotics, the company confirmed Tuesday, to deploy tech capable of streamlining repairs and reducing maintenance delays for a surface fleet that continues to be stretched thin.

The contract will begin as a 5-year, $54 million indefinite delivery, indefinite quantity deal that will soon see Gecko begin work on 18 ships assigned to the Navy’s U.S. Pacific Fleet.

To expedite what has in recent years become a headache for naval readiness, Gecko uses drones, wall-climbing robots and fixed sensors to gather data on components, decks, welds and hulls.

That information, paired with AI tools, is used to identify current and potential structural issues that may remain hidden to the naked eye.

“A single robotic evaluation and digital rendering of a flight deck eliminated over three months of potential maintenance delay days,” the company release stated about one such procedure.

These measures have expedited maintenance “up to 50 times faster and more accurately than manual methods,” the company added.

“Readiness isn’t just a metric, it’s all that matters,” Jake Loosararian, co-founder and CEO of Gecko, said in the release. “This growing partnership is about unfair advantages Gecko is deploying to our Navy; and how prediction, through our robotics and AI products, ensure our brave men and women are the most advantaged in the world in their fight to defend freedom.”

In fall 2024, then-Chief of Naval Operations Adm. Lisa Franchetti unveiled a goal of having 80% of the Navy’s fleet ready to deploy at any given time by 2027.

Obstacles to reaching that goal materialized quickly, however, with a Government Accountability Office report in December 2024 highlighting a readiness rate among the Navy’s amphibious warfare ships of just 46% between 2011 and 2020. This rate, in turn, significantly hindered Marine Corps deployment and training plans.

In August 2025, that rate reportedly dipped to just 41%, resulting in a more than five-month gap in Marine Expeditionary Unit deployments that year and further straining resources amid the Trump administration’s push to counter the illicit drug trade in Latin America and the Caribbean.

The 80% plan, meanwhile, was picked up by Franchetti’s successor, Adm. Daryl Caudle, who called the rate “an ambitious but essential readiness goal.”

“Achieving this requires shorter maintenance cycles, increased spare-parts availability, improved training pipelines and targeted upgrades across the fleet,” Caudle wrote for Military Times in December.

“Readiness is not a budget line — it is a promise to the American people that their Navy will never arrive late to a fight."

The new tactical loitering munition from Singapore’s best-known defense company — dubbed the EagleStrike, and the first weapon in this category from ST Engineering — should be ready for production in early 2027.

The beyond-line-of-sight airborne weapon was developed at the company’s initiative, but it will presumably be offered to the Singapore Army in due course.

Jackson Tean, assistant vice president of International Business, Advanced Material Engineering, Land Systems, said the loitering munition has reached Technology Readiness Level 7. Further testing and certification will occur this year to fully mature it.

The company listed the munition’s range at 12.4 miles, with a 30-minute loitering time. Traveling at speeds of 67 miles per hour, the EagleStrike possesses a 7-ounce dual-mode shaped-charge warhead optimized for top attack against lightly protected targets.

The EagleStrike is launched from a canister. It is expected that a pod of 16 such canisters could be carried on a suitable vehicle.

ST Engineering has a range of drones in its portfolio, but Singapore’s military has shown a predilection for Israeli-made unmanned aircraft.

At Singapore Airshow 2026, for example, it publicly exhibited for the first time its large Hermes 900 drone from Elbit Systems as well as the Orbiter 4 from Aeronautics. The Hermes 900 is replacing the Singapore air force’s Hermes 450, which was in service for more than 20 years.

Turning to small arms, ST Engineering unveiled a new family of 5.56mm-caliber rifles based on Colt’s AR-15 design. This rifle family contrasts with the SAR-21 bullpup rifle that ST Engineering supplies to the Singapore Armed Forces.

This AME modular family of rifles is designed to equip an army squad with a range of five weapons comprising a carbine, assault rifle, grenade launcher, light machine gun and sharpshooter rifle.

The smallest member is the AME-A510 with a 10.5-inch barrel, which could be ideal for a vehicle crewman, for example. Next, the standard AME-A514 squad rifle boasts a 14.5-inch barrel. Specifications released by ST Engineering state a length of 32.3 inches and a weight of 6.4 pounds unloaded.

The same rifle can have an under-barrel 40mm grenade launcher attached, which changes its nomenclature to AME-A514GL.

More capable is the AME-A516L, which possesses a 16-inch barrel and integrated heat sink so it can act as a squad’s light machine gun.

Finally, the family boasts the AME-A520SS designated marksman rifle with a bipod.

A spokesman told Defense News that these rifles are aimed primarily at the export market. He also confirmed international sales have been achieved already.

Enhancements compared to the original AR-15 include enhanced barrel harmonics, a two-stage trigger, a muzzle brake that reduces recoil by 20% and ambidextrous controls.

A representative said the rifles are designed to be reliable, simple and cost-effective.

Housed within the Virginia Space Authority’s Mid-Atlantic Regional Spaceport in Wallops Island, the facility, dubbed Launch Complex 3, will support testing, launch and return missions for the reusable rocket — the largest vehicle to ever fly from the spaceport.

“As we prepare for our next-generation rocket, it makes sense that we have a world-class launch facility and landing facility,” Rocket Lab CEO Peter Beck said at the event, which was attended by several Virginian politicians, including Gov. Glenn Youngkin (R). “That high cadence for launch vehicle Neutron will expand indigenous capability and enable the United States to quickly and reliably reach the International Space Station, Earth orbit, the moon and beyond.”

The company sees Neutron as a contender to help ease the bottleneck in demand from both commercial and military customers for a ride to space. Today, that demand is in large part being met by a single provider in the medium-lift market, Elon Musk’s SpaceX.

The firm has its sights on eventually supporting the Defense Department’s most important and complex national security missions. Today, only SpaceX and United Launch Alliance are eligible to launch those missions, though Jeff Bezos’ Blue Origin will join them once its New Glenn vehicle meets all of its certification requirements.

Last May, the U.S. Space Force brought Rocket Lab into the fold of its multibillion-dollar National Security Space Launch program, which allows the firm to offer the Neutron to fly a less-strenuous slate of military missions.

But first, Neutron needs to take flight at all. Beck has said on several occasions that while Rocket Lab believes its plan to launch this year is within reach, the schedule is aggressive with no margin for error.

Speaking with reporters at the launch site, Beck said the company has some key testing in the coming months to qualify key stages of the rocket, which will give it a better idea of whether it can meet that 2025 timeline.

“Nobody’s waving the white flag here until the last hour of the last day,” he said.

Following the first launch this year, the company is eyeing three more next year and five the following — eventually building its cadence to one flight per month. Asked how soon the firm expects to be competing directly with SpaceX, Beck said Rocket Lab is focused on helping meet demand in the market, not on a particular competitor.

“We service the opportunities that are in front of us, sort of run our own race,” he said. “I don’t wake up every morning going, ‘How am I going to compete with Elon?’ We see this opportunity here and we’re going to go after it and do a good job.”

The path forward

As Rocket Lab prepares for Neutron’s first flight, progress at its Virginia launch site in the coming weeks will be crucial.

Shaun D’Mello, a company vice president overseeing Neutron, told reporters that the rocket’s hardware is starting to be delivered by boat to a barge at the facility, and the company expects to have the full rocket on site by November. Much of the vehicle’s assembly is happening at a Rocket Lab facility in Middle River, Maryland, so the final mating at the site should be straightforward, he said.

“Our second stage is already on a boat on its way,” he said. “So, we’ll soon see a lot of hardware coming together here. A lot of the key integration will take place up there, so it’ll be a just-in-time arrival of hardware at Wallops before we put it on the pad.”

Once the rocket is on site and integrated, the firm will start static-fire testing, where engineers will validate the engine’s performance and ensure the system is ready for launch.

In parallel with hardware arrivals, Rocket Lab’s focus right now is in two key areas: readiness for Neutron’s ascent mission and the performance of its propulsion system.

While the rocket will eventually be fully reusable, the company is aiming for a controlled “splash-down” landing after its first flight. D’Mello said the firm’s experience with its high-cadence Electron rocket — which flies smaller satellites from Wallops Island and the Mahia Peninsula in New Zealand — has prepared it for the launch itself, but the rocket’s landing is a new challenge.

“Most of the engineering team right now is spending time landing the rocket and not launching because we know how to get to orbit,” he said.

On the propulsion system, dubbed Archimedes, D’Mello said the company isn’t seeing any showstoppers but noted that this phase of testing is difficult.

“It’s a physics problem at the end of the day, rocket engines, so we’re holding the course and seeing some great performance,” he said. “Other than the fact that rockets are hard, there’s nothing out of the ordinary. ... We’re not too particularly concerned about any specific aspect. It all just needs to come together.”

Room to grow

As Rocket Lab expands Neutron and Electron operations in the coming years, D’Mello and Beck said they’re looking forward to seeing the facility at Wallops Island grow with them.

Officials representing the spaceport and NASA as well as state and local politicians said this week that Neutron’s success is a key part of their strategy for making Wallops a bigger player in commercial and military launch.

Speaking Thursday during the Rocket Lab event, Virginia Governor Glenn Youngkin said the state’s vision is for the launch facility to one day be “the most vibrant space industry complex in America.”

“In order to lead, it requires us to have an ecosystem that can fulfill not just the current requirements but the future requirements,” Youngkin said.

Rocket Lab officials said that while they considered more traditional medium and heavy-lift facilities like Cape Canaveral Space Force Base in Florida, space is tight at those venues. Expanding their operations at Wallops offered a chance for the company to be a priority tenant and have a better chance at meeting its high-cadence launch targets.

“The Cape is an amazing place,” Beck said. “But it’s busy. It’s only getting busier.”

The facilities at Wallops are sufficient for Rocket Lab’s current operations, but as the company looks to the future, with an eye toward supporting human spaceflight, it will require more infrastructure, including deep-water ports.

“Right now, we have the minimum viable equipment that we need to get on and start ramping up,” he said. “As that ramps and we’re successful and the state’s successful, then we’ll all continue to invest in that infrastructure.”

Pending range approval, the mission is slated to lift off Aug. 12 from Cape Canaveral Space Force Station in Florida, carrying Space Force satellites destined for geosynchronous orbit, about 22,000 miles above Earth.

The milestone is a long time coming for ULA, the Defense Department’s longtime launch provider, which started developing its Vulcan rocket in 2014. The vehicle — which is replacing the company’s legacy workhorse, Atlas V — flew two demonstration missions last year and achieved its required military certification from the Space Force in March.

ULA and SpaceX are the only companies with rockets certified to fly National Security Space Launch (NSSL) missions. For years, prior to the certification of SpaceX’s Falcon 9 rocket, ULA was the sole provider of military launch services. In 2020, SpaceX won a 40% share of those missions slated to fly between fiscal years 2022 and 2027 with ULA maintaining the remaining 60%.

In April, the Space Force awarded SpaceX, ULA and national security launch newcomer Blue Origin a combined $13.5 billion for missions that will fly between fiscal years 2027 and 2032. Under the deal, SpaceX will receive $5.9 billion to fly 28 missions, ULA $5.3 billion to launch 19 and Blue Origin $2.3 billion to conduct seven.

The Space Force mission set to fly on the rocket next week, USSF-106, was originally set to launch in 2022. After being rescheduled as a Vulcan mission in 2023, the satellites have been grounded for two years as the rocket faced delays.

The mission includes an experimental spacecraft built by L3Harris called Navigation Technology Satellite-3, or NTS-3, that will demonstrate capabilities to augment the Space Force’s GPS constellation or support a future program. That effort, led by the Air Force Research Laboratory in partnership with Space Systems Command, will test technologies like steerable beams to provide regional coverage, a reprogrammable payload that can receive upgrades in orbit and protections against signal jamming.

Once NTS-3 is in orbit, AFRL will experiment with those capabilities over a one-year period, exploring how new satellite configurations could strengthen the Space Force’s positioning, navigation and timing, or PNT, capabilities.

That could mean integrating new capabilities onto the Space Force’s GPS IIIF satellites, which are built by Lockheed Martin, and approaching initial production. The service may also funnel relevant technology into other positioning, navigation and timing programs, including an effort called Resilient GPS, which is meant to augment the larger constellation with a fleet of small, lightweight, lower-cost satellites.

NTS-3 is the Air Force Research Laboratory’s first major PNT demonstration in nearly a half century. The last NTS satellite flew in 1977 and proved out capabilities that ended up being integral to the GPS program.

CEO Tony Gingiss, who’s been in the role since last December, told Defense News that much of his initial focus has been on managing the firm’s growth.

The company, which was founded in 2001 and acquired by Boeing in 2018, has delivered 14 commercial and government satellites to orbit. It’s now positioned to produce about that same number of spacecraft this year alone and expects to double its deliveries next year, Gingiss said in a recent interview.

The firm also has around 70 or 80 vehicles in its backlog to deliver over the next four years.

“Really, my focus is on the team, the infrastructure, the processes that we have, the tools that we have ... to be able to meet that demand,” he said. “And that’s a commitment to today, but also to continue to fill the pipe with new customers and new business because there’s clearly a lot of exciting things going on in the industry that we want to be prepared to continue to respond to and win work to keep filling our pipe.”

Millennium is a key provider for several Space Force programs, including 12 satellites for Space Systems Command’s Missile Track Custody program and eight vehicles for the Space Development Agency’s Foo Fighter effort — both programs focused on space-based missile warning and tracking.

For both programs, the company is scheduled to deliver initial satellites in fiscal 2027.

To accommodate those spacecraft and other Defense Department and NASA programs, the company is focused on two levers that Gingiss said are key to scaling capabilities: efficiency and capacity.

On the efficiency front, the firm and its parent company Boeing have been working to introduce more automation and build in commonality between product lines that allow for more capacity and output without a larger footprint.

The company is also streamlining processes so that its various operations — from manufacturing to analysis — take less time to conduct.

While those approaches helped make room for more work, the Foo Fighter contract Millennium received last year required further expansion to its vehicle integration capabilities. Rather than build a new facility, the company announced in April it would use 18,000 square feet of existing space adjacent to its current small site factory that resides within Boeing’s facility in El Segundo, California. That amounts to about a 30% increase to Millennium’s overall manufacturing footprint and will come online later this year.

“When you turn both of those knobs of efficiency and capacity simultaneously, ultimately what you do is increase capacity, but you decrease the dollar per capacity unit that you’re spending because you’re getting more value with every hour spent,” Gingiss said.

As Gingiss and his team prepare for that planned growth, there’s also the possibility for unplanned production increases — especially as DOD pursues programs like Golden Dome and invests in tactically responsive space capabilities that could require significant production ramp-ups or requests for additional satellites on short notice.

Millennium has some additionally capacity built into its inventory to meet “unforecasted demand” for certain parts that are common across its product lines, Gingiss said. While those don’t equate to a full vehicle, they provide a baseline level of responsiveness to customer needs.

At the vehicle level, Gingiss said the company works closely with its customer — in this case SDA or SSC — to understand where they might want flexibility to buy more of a particular vehicle or what future capabilities or variants they might want.

“I think a lot of that is really staying in lockstep with the customer on what their strategy is and getting that clear demand signal from them about what would those demands look like,” he said.