New York National Security Forum 2026: State of Play Brief

The New York National Security Forum (NYNSF) is an annual, invitation-only convening that puts defense-tech founders in the same room as the acquisition executives, appropriators, senior government officials and investors who move capability from prototype to program of record. NYNSF 2026 convenes:

• Investors from defense-focused venture funds, growth investors, family offices and strategic investors, representing more than $1 billion in dedicated defense and dual-use venture capital, together with multi-strategy investors managing more than $10 billion and the Department of War’s (DoW) Office of Strategic Capital
• Founders and industry executives, most of them building in AI and digital operations, autonomous systems, counter-UAS, ISR and advanced manufacturing; roughly two-thirds of the founder companies already hold a federal contract
• Acquisition officials and senior government officials, drawn from Navy PAE offices and the DoN Rapid Capabilities Office, Army Applications Laboratory, Marine Corps CD&I and the Marine Innovation Unit, SOCOM, DIU, OSC, NATO ACT and the SBA

The Forum exists to overcome the structural problems that keep the U.S. military from acquiring defense technology faster than its adversaries. This brief sets out those problems in the order the Forum will take them up. Day 1 takes the demand side: what the warfighter needs, how fast that need changes, and why what gets built so rarely reaches the unit that asked for it. Day 2 takes the capital side: why plentiful private money converts into so little acquired capability, and what investors, acquirers and founders can each do about it. Between the two sits the argument that connects them: operational validation is the missing pricing mechanism.

Each section carries a tag naming the session in which it will be argued, so the brief can be read as a companion to the program. Sessions are held under the Chatham House Rule; the tags reference the public program only.

What the founders build and what the investors are looking for converge on a narrow set of areas. The companies cluster in AI and digital operations, autonomous systems, counter-UAS and terminal defense, ISR and sensing, and advanced manufacturing. The investors name hardware and manufacturing, autonomy and robotics, space, AI and software, and energy and power, weighted toward pre-seed through Series A with a smaller cohort at Series B and growth.

Most of the founder companies are past first revenue. The room therefore sits on the far side of the first-contract problem and squarely inside the conversion problem this brief describes. The commands those companies most want to reach are the Navy, SOCOM, the Army and INDOPACOM; the Navy, the Army and SOCOM also supply much of the government attendance.

01Part I · The Demand Side (Day 1)

01 The Adversary Rewrites the Requirement Every Six Weeks

In the room: D1-06, Warfighter Demand Signal, Now, Mon 13:50

Requirements and battlefield needs move faster than the acquisition system. In Ukraine, an 80% solution delivered this quarter beats a perfect one delivered in 18 months, because by the time the perfect one arrives, the requirement it was built against has already changed. Systems should be scored on iteration velocity, the time from battlefield observation to a fielded fix that holds on retest, rather than on compliance with a specification frozen at contract signature.[1]

Defense technology should be procured as a service. Different theaters make different demands of a system. A drone radio operating near 1.8 GHz may work in the U.S. but fail in Europe, where that range sits inside LTE Band 3, the continent’s crowded cellular pairing; U.S. allocations near 1.8 GHz fall on different pairings.[2] A special forces unit may need different capabilities than a cavalry unit. An adversary may render a system obsolete within two weeks of combat.

Buying capability the way software is bought, delivered and continuously updated rather than sold once as a fixed product, addresses all three: the contract ties the vendor to outcomes and obligates it to iterate as the warfighter’s needs change. The model has precedent. In June 2024 the Coast Guard awarded Shield AI a $198.1 million indefinite-delivery contract for contractor-owned, contractor-operated V-BAT surveillance services, and the approach has been argued for directly as "defense as a service."[3] It also shifts risk onto data rights, wartime surge capacity and the legal limits on contractors operating at the front, and contracts should be written with those costs in view.

02 Misaligned Requirements and Broken Feedback Loops

In the room: D1-08, From Prototype to Fielding Capability, Mon 15:20; D1-06, Warfighter Demand Signal, Mon 13:50; D2-03, From the Hill to the Board Room, Tue 11:05

End users are often not the buyers. The soldiers who operate a system typically have little influence over its procurement, while the acquisition officials who select it are often removed from the specific operational needs of a unit in a specific theater. This creates a structural tension: warfighters prioritize capability and fit for their environment, while acquisition officials must prioritize standardization to enable interoperability, simplify logistics and support mass production. The result is a persistent dilemma between scale and usefulness. Procuring 100,000 drones accomplishes little if an airborne unit cannot bring them on airborne operations.

A scaling strategy applied to a problem that first requires iteration destroys the conditions iteration depends on.[4] The solution is to build feedback loops between warfighters, acquirers and entrepreneurs, or in other terms, between end users, buyers and vendors. In practice, that means inviting industry into military exercises where warfighters can test novel platforms and give feedback directly. It also means carrying the experience of warfighters up into the echelons where acquisition decisions are made, so that procurement cycles are built around solving a problem statement rather than fielding a capability.

When acquisition programs fail, the lessons are rarely captured. From 2012 to 2021, only 16% of DoW SBIR/STTR-funded firms, 748 of the 4,703 analyzed, reached a Phase III award.[5] What happens to the other 84% is largely unrecorded. There is unlikely to be a systematic mechanism that tracks why a given program did not transition, whether similar SBIR-funded firms are aware of each other’s parallel work, or whether failed approaches are documented so they are not repeated.[6] The technical data and software from a failed project could help another firm avoid the same dead end, or pick up where the original company could not finish. Instead, under the current DFARS clause the Government must protect that data for 20 years from the date of award, whether the program transitioned or failed outright, subject to exceptions and to terms negotiated after award.[7]

The Secretary of War has ordered authority over drone procurement pushed down to warfighters, yet in practice it still is not reaching them. In the 10 July 2025 "Unleashing U.S. Military Drone Dominance" memorandum, Secretary Hegseth wrote: "I am delegating authorities to procure and operate drones from the bureaucracy to our warfighters."[8] Units can use Operations & Maintenance (O&M) funds to buy commercial defense technology only when the system unit cost is not more than $350,000 (or $500,000 where the Secretary determines it necessary for a combatant commander in a named contingency operation); anything above that is an investment item that must be bought with procurement funds, which a brigade does not control.[9] Research, Development, Test and Evaluation (RDT&E) funding covers development and testing, but it is appropriated into high-level accounts at service headquarters. A U.S. Army brigade holds no direct RDT&E obligation authority and must work through a higher-echelon research organization to use it.[10]

Decentralized acquisition without standards multiplies the problem. Different vendors may solve the same problem without building systems that work together. A commander is not more combat effective watching five separate screens for five drone systems acquired through five different fast-acquisition vehicles. Newly acquired systems need interoperability standards for software and hardware to plug into the kill chain.[11] The analogy is the European Union’s requirement that all smartphones use USB-C chargers: a unit should not need to carry five different battery chargers into battle for five different drones. The standards exist but are not enforced. The Modular Open Systems Approach, codified in law, requires major defense acquisition programs that reached Milestone A or B after 1 January 2019 to use a modular open system approach "to the maximum extent practicable." The Army Science Board’s Robotic and Autonomous Systems-of-Systems Architecture study recommends that the Army define a modular open system architecture for robotic and autonomous platforms. The Army’s Joint Small UAS Capability Development Document calls for a Joint Reference Architecture to sustain joint technical standards for sUAS interoperability, and DoD Instruction 8330.01 sets interoperability requirements for Department information technology within a scope that excludes test, training and non-operational systems. Of the three, only the Instruction is binding, and none is operationalized in drone procurement. Warfighters of the 173rd Airborne Brigade report connection, boot-time and battery problems with separately sourced controller and airframe combinations, so that the assembled system fails to meet its advertised specifications.[12]

03 The Exquisite vs Attritable Mismatch

In the room: D1-06, Warfighter Demand Signal, Mon 13:50; D1-07, Distributed Spectrum, Mon 15:00

Tactical engagement won, economic battle lost. Each time a PAC-3 MSE interceptor, budgeted at $3.9 million a unit in the Army’s FY2026 procurement justification and reported near $5.3 million in FY2027 request documents, destroys a Shahed-136 that CSIS puts at $20,000 to $50,000, the intercepting side loses the economic exchange. The ratio runs from roughly 77 to 1 to roughly 265 to 1: for the price of one interceptor, the attacker can field scores of drones. Ukraine has run the exchange the other way. Its unmanned surface vessels (USVs) cost roughly $200,000 to $250,000 each and have destroyed or disabled Russian warships worth tens of millions of dollars apiece, some more than $100 million. Aggregate totals remain uncertain.[13]

Attritable systems should be priced on cost per effect, not unit cost. F-35s, M1 Abrams and Patriot batteries are exquisite systems, priced by unit cost and service life because they are built to last decades. Attritable systems are different. Designed for a single mission and meant to be expended in mass, they should be priced by cost per effect. The mismatch shows up when an exquisite system (a Patriot interceptor) is used against an attritable threat (a Shahed). Attritable threats should be countered with attritable systems, priced and procured on the same cost-per-effect logic.[14]

FIGURE 01 · COST PER EFFECT, WITH THE UNCERTAINTY SHOWN

One PAC-3 MSE interceptor costs as much as 77 to 265 Shahed-136 drones. Cost per effect divides unit cost by the probability of a kill; the cheap-interceptor figures and both kill probabilities are illustrative, and the PAC-3 MSE range runs from the Army’s FY2026 budget to FY2027 request reporting.

Source: U.S. Army FY2026 Missile Procurement justification book; FY2027 request reporting; CSIS Missile Defense Project, Shahed-131/-136 profile, 2 June 2026; Hollenbeck et al., CSIS, 19 February 2025.

04 The Small Unmanned Aircraft System (sUAS) Vulnerability

In the room: D1-06, Warfighter Demand Signal, Mon 13:50; D1-07, Distributed Spectrum, Mon 15:00

The U.S. military is ill-equipped to fight a war in which thousand-dollar sUAS can knock out vehicles costing tens of millions. The U.S. Army’s 173rd Airborne Brigade has been building its own first-person-view (FPV) drones in an organic drone lab, at about $1,000 a unit, because neither the commercial drone market nor the Army had delivered what it needs.[15] During Combined Resolve 26-07, held at Hohenfels, Germany, from 9 April to 10 May 2026, Ukrainian drone operators from the 412th Separate Unmanned Systems Brigade repeatedly defeated the 3rd Armored Brigade Combat Team, 1st Cavalry Division, in the exercise’s initial iterations, before U.S. units improved their dispersion, camouflage and electronic warfare.[16] At Exercise Hedgehog 25 in Estonia in May 2025, a Ukrainian team of roughly ten, acting as the adversary force, mock-destroyed 17 armored vehicles and ran some thirty further simulated strikes in about half a day, costing two NATO battalions their combat effectiveness.[17] The DoW launched the offensive Drone Dominance program in December 2025, aiming to buy some 340,000 one-way-attack drones over two years, and JIATF-401 expanded a separate marketplace for counter-UAS systems on 3 August 2026, but an important gap between funding and fielding remains.[18]

05 Lagging Arms Manufacturing

In the room: D1-04, Scaling Production for the Warfighter, Mon 11:30; D1-03, Securing the Supply Chain, Mon 11:15

Demand for U.S. ammunition exceeds what the industrial base can supply. Rheinmetall’s multinational production network, with sites in Spain and South Africa as well as Germany, aims to produce up to 1.1 million 155mm artillery shells a year by 2027, rising to around 1.5 million by 2030. That 2027 target alone would exceed the latest publicly reported U.S. rate. The U.S. Army was producing 36,000 rounds a month as of March 2026, short of the 1.2 million annual (100,000 monthly) target it had set for October 2025, according to the Department of War Inspector General’s evaluation of 155mm production capacity.[19] U.S. Patriot production likewise cannot keep pace with wartime demand.

Lockheed delivered 620 PAC-3 MSE interceptors in 2025, about 1.7 missiles a day; coalition forces fired at least 1,700 Patriots in the first five weeks of the Iran war, close to three years of production at that rate.[20] CSIS estimates that Patriot stockpiles fell from roughly 2,330 before the war to 1,030 at the April 2026 ceasefire, and then to between 759 and 827 by late July 2026, about a third of the prewar inventory.[21] These interceptors are fundamental to ballistic and cruise missile defense.

The munitions shortfall for a protracted war with China predates Operation Epic Fury. Navy procurement of the Long-Range Anti-Ship Missile (LRASM) was projected at 48 a year in FY2020 and 91 in FY2024, rising to 149 a year by FY2028, and Navy jets would run out of LRASMs within days in a fight with China.[22] CSIS wargaming suggests U.S. forces could use hundreds of some missile types, including the LRASM at roughly $3 million each, in just the first week of a Taiwan conflict.[23] On the Tomahawk, U.S. forces expended more than 1,000 missiles during the Iran war, against a 10-year average procurement rate of 86 missiles a year. Production is not projected to replace what was fired at Iran until late 2030.[24]

The FY2027 budget addresses the shortfall. Its request for the fourteen munitions tracked under the Munitions Acceleration Council, twelve legacy systems plus two emerging capabilities, totals $52.885 billion, up from $10.828 billion enacted in FY2026, a 388 percent increase.[25]

FIGURE 02 · WEEKS TO EXPEND, YEARS TO REPLACE

Panel A applies pre-war production and procurement rates; CSIS projects Tomahawk expenditure replaced by late 2030 as production ramps. Panel B shows that about a third of the Patriot drawdown came after the April ceasefire.

Source: Cancian and Park, CSIS, 27 July 2026 and 27 May 2026; Lockheed Martin, 6 January 2026; Amoah, Bazilian and Matisek, FPRI, 18 May 2026.

02Part II · The Bridge: Validation as the Pricing Mechanism

06 What "Good" Looks Like Remains Unclear

In the room: D1-08, From Prototype to Fielding Capability, Mon 15:20; D2-02, Capital Formation in Defense Tech, Tue 10:00

The most combat-tested defense industrial base in the world attracts roughly one-hundredth of one percent of global venture funding. PitchBook recorded $57.2 million of equity across 28 deals in Ukrainian defense tech in 2025, against an estimated 2025 market output of $6.8 billion and production capacity near $35 billion.[26] Artem Moroz, Head of Investments at Brave1, estimates that the same firm, at the same stage of revenue and technological readiness, would cost at least 10 times more in the United States and 3 times more in Europe.[27] Two things explain the gap.

First, demand comes almost entirely from a single buyer, the Ministry of Defence, and a one-customer market is worth less than a diversified one.[28] Second, the rest is structure rather than substance: only a few firms have properly registered or transferable IP rights, and many have no legal standing at all, operating as groups of private entrepreneurs.[29]

FIGURE 03 · UKRAINE’S DEFENSE TECH: FUNDING AGAINST MARKET

Equity invested and market size measure different things and are shown on separate scales. Market output is what the industry sold in 2025; production capacity is what it could make.

Source: Javaheri, “Ukraine’s Defense Tech Ecosystem,” PitchBook, 1 July 2026; KSE Institute, “The Ukrainian defense technology market,” March 2026; Crunchbase News, 7 January 2026; KPMG Venture Pulse Q4 2025, 19 January 2026.

Operational validation is the missing pricing mechanism. Proving a concept in battlefield-like conditions should happen earlier in the development cycle, embedding end-user feedback into quick iteration rather than a single late-stage test. Pharmaceuticals require clinical trials. Vehicles require crash tests. Defense technology requires the equivalent: a rigorous, repeatable checkpoint between demonstration and fielding. A validation result, earned in battlefield conditions and certified by an evaluator with no stake in the outcome, functions as a due-diligence accelerant.

For that checkpoint to carry weight, the rating has to reflect what the evidence actually covers. A vendor’s own bench data and an observed trial in operational conditions should not earn the same result, and the result should say which one it rests on.

Ukraine has already built the state version. Brave1, established in April 2023 by the Ministries of Economy, Defence, Digital Transformation and Strategic Industries with the General Staff and the National Security and Defence Council, bundles grant funding with structured testing, certification and direct procurement matchmaking between developers and military end users, and counted 2,500 companies as of September 2026.[30] Brave1 Market lets combat units earn e-Points for verified enemy-target hits, recorded in DELTA and confirmed by Brave1, and spend them on codified drones, ground robotic systems and electronic-warfare equipment; the Defence Procurement Agency then contracts, pays and manages logistics.[31] Combat performance becomes the price signal: a drone that keeps working gets reordered. This is how success can be assessed.

07 Lessons Learnt in Ukraine Stay in Ukraine

In the room: D1-08, From Prototype to Fielding Capability, Mon 15:20

Ukraine holds the deepest combat-tested knowledge in drone warfare, and most of it stays inside the country. Arms exports were virtually absent until early 2026. There was no outright legal ban, but approval runs through the State Service for Export Control, a seventeen-member interagency commission and the Cabinet, and in practice exports simply did not happen.[32] As of March 2026, Wild Hornets, maker of the Sting interceptor, could not sell abroad. "Exports of drones are not permitted at this time," its spokesman said, while the company reported inquiries from Middle Eastern and EU representatives and President Zelensky said eleven nations were interested in Ukrainian interceptors.[33] Kyiv has since opened controlled exports in stages, and licenses go only to selected companies; Cabinet Resolution No. 875 set a 30-day track for applications from eligible partner states, while the older route could take 90 days or, via direct Cabinet designation, up to six months.[34] The friction runs both ways: Airlogix, a Ukrainian drone maker building in the United States through a joint venture with Auterion, reports that licensing its U.S.-developed technology for use in Ukraine can take four months, against an iteration cycle measured in weeks.[35] Some firms with the resources to structure around the controls can get past them.

The British firm Skycutter topped Gauntlet I of the Pentagon’s Drone Dominance Program, scoring 99.3 out of 100 among eleven winners, on a bid built in partnership with the Ukrainian firm SkyFall.[36] Swarmer, a drone software company billed as the first Ukrainian defense company to list on Nasdaq, was incorporated in Delaware on 15 May 2023 and completed its Nasdaq listing in March 2026; in September 2026 it agreed to acquire the Ukrainian ground-robotics maker Ratel for up to $224 million, most of it contingent on earnouts.[37] Most Ukrainian firms do not have such privileged access to the U.S. market.

Neros Technologies is proof of what direct access to Ukraine’s battlefield knowledge is worth. Its founders built thirty Archer prototypes and took them to Ukraine in September 2023 for field testing and feedback, and tested every later version with frontline units. That firsthand knowledge built a company that in July 2026 won a U.S. Army indefinite-delivery contract with a $500 million ceiling, and produced an FPV drone well regarded by the 173rd Airborne Brigade.[38]

03Part III · The Capital Side (Day 2)

08 The Investment/Acquisition Gap

In the room: D2-01, The Future of Capital in Defense Tech, Tue 09:30; D1-02, The State of Defense Acquisition, Mon 10:15; D2-04, Private and Public Partnerships, Tue 11:55

Capital for defense tech is plentiful. Few of those dollars convert into acquired solutions for the U.S. military. Silicon Valley Defense Group (SVDG) describes 2025 as a year of seismic shifts across policy, capital and contracting. SVDG’s NatSec100 firms, the top 100 venture- and PE-backed defense tech companies, received $86.3 billion from private capital in calendar 2025, or $39.6 billion excluding OpenAI, which entered the list in the 2026 edition.[39] Federal obligations to that same cohort totalled $4.3 billion in FY2025, up 22 percent on FY2024.[40] Investors need returns, and defense tech firms need to survive the valley of death between prototype and procurement. Both require federal acquisition spending to increasingly match the scale of pre-procurement investment.[41]

FIGURE 04 · THE INVESTMENT AND ACQUISITION GAP

Private capital is calendar 2025; federal obligations are fiscal 2025, so the ratio is indicative rather than like-for-like. Both private-capital totals are SVDG’s published figures.

Source: Silicon Valley Defense Group, 2026 NatSec100 Report, 26 May 2026.

The conversion problem is also present in federal R&D programs. From 2012 to 2021, sixteen percent of firms funded through the DoW’s Small Business Innovation Research (SBIR) program, 748 of 4,703, received a Phase III award, the acquisition or advanced research stage. Department obligations through Other Transaction Agreements (OTAs) for prototyping and production grew from $1.8 billion in FY2016 to over $18 billion in FY2024, of which some $16 billion was prototype work and $2 billion production. A production OTA can follow a competitively awarded prototype without further competition once the prototype has been successfully completed; it is not automatic, and the Department may use a FAR contract instead. Nevertheless, according to the Government Accountability Office, the Department cannot systematically track the extent to which it transitions prototype projects funded through OTAs into standard production efforts.[42]

09 Risk Aversion in the Approval Chain

In the room: D2-03, From the Hill to the Board Room, New Acquisition Reform, Tue 11:05

Slower acquisition increases the risk of wastage. IVAS, the Littoral Combat Ship and the Constellation-class frigate each show how a slow, multi-year process can still end badly, in different ways. GAO reports that none of the Army’s nearly 10,000 IVAS units will be fielded; that the Navy’s LCS operating-and-support estimate rose from $38 billion to more than $60 billion after $28 billion had been spent building 32 ships; and that the Constellation program carries some $310 million in identified cost growth, with its lead ship about three years late.[43] Faster acquisition carries its own risk of failure. But for the attritable systems warfighters increasingly need, the cost of staying slow will likely exceed the cost of occasional failures from moving fast.

Individuals in approval chains default to "no" because denial carries no personal risk. A Harvard Kennedy School Belfer Center study describes a warrant officer explaining that the word "jamming" triggers an automatic denial of a request, whereas "transmitting" does not, "despite there being no functional difference in the underlying capability." Innovation cannot occur without failure, and it is preferable for things to go wrong in testing rather than on the battlefield.[44]

10 Venture Structure vs Capability Building

In the room: D2-02, Capital Formation in Defense Tech, Tue 10:00; D2-01, The Future of Capital in Defense Tech, Tue 09:30

Venture capital structure itself pushes against long-term capability building. VC funds typically operate on a ten-year partnership term, deploying capital during an investment period of roughly the first three to five years, and need a credible realization plan for LPs well before the term expires.[45] That pressure shapes what founders build. 2025 defense tech exits, mostly acquisitions, hit a record $54.4 billion, up from $18.2 billion in 2024, according to PitchBook, and "have a clear exit thesis" is now standard advice to defense tech founders.[46] A company engineered to be acquired in five years optimizes for what an acquirer will pay for, not for what a warfighter needs sustained and supported over a program’s actual lifecycle, which regularly runs a decade or more.

04Part IV · What Is Moving

The problems set out above are the standing condition. They are not the whole picture. Four changes since mid-2025 bear directly on the arguments in this brief, and each is close enough to the Forum’s sessions to be worth testing against the people in the room.

11 The Requirements Process Has Been Rebuilt

In the room: D1-02, The State of Defense Acquisition, Mon 10:15; D2-03, From the Hill to the Board Room, Tue 11:05

JCIDS is gone. A memorandum signed on 20 August 2025 by Secretary Hegseth and Deputy Secretary Feinberg directed the Vice Chairman to begin the immediate disestablishment of the Joint Capabilities Integration and Development System and ordered the Joint Requirements Oversight Council to stop validating component-level requirement documents.[47] CJCSI 5123.01J, issued 15 January 2026, replaced JCIDS in its entirety with the Joint Force Requirements Process. Service and component requirement validation is now the responsibility of the services themselves.[48] The reoriented JROC identifies and annually ranks Joint Force Key Operational Problems; a Requirements and Resourcing Alignment Board, co-chaired by the Vice Chairman and the Deputy Secretary, selects topics from the top-ranked problems each budget cycle and feeds its recommendations into the budget build; and a Mission Engineering and Integration Activity was directed to work with industry to refine requirements through rapid experimentation.[49]

What it means for this room. The document path that carried a capability from idea to validated joint requirement has been removed, and with it one of the standard explanations for why a working prototype stalls. Twenty-two years of JCIDS practice, dating to June 2003, has been unwound in under a year. Whether the services build something faster than what they inherited is unresolved, and it is the question worth putting to the acquisition and reform panels rather than answering here.

12 Federal Capital Has Become an Instrument of Industrial Policy

In the room: D2-01, The Future of Capital in Defense Tech, Tue 09:30; D2-04, Private and Public Partnerships, Tue 11:55

The Office of Strategic Capital now lends at a scale that changes what private capital is deciding about. Congress established OSC in the FY2024 NDAA with 31 covered technology categories, since expanded to 34, and FY2024 appropriations then supported up to $984 million in direct loans. The One Big Beautiful Bill Act appropriated $500 million in credit subsidy for critical minerals and related projects, against a statutory ceiling of up to $100 billion in loan principal. OSC executed its first direct loan in 2025, $150 million to MP Materials for heavy rare earth separation at Mountain Pass.[50] In July 2026 it announced a conditional commitment of up to $820 million to Performance Drone Works to establish high-volume domestic manufacturing of critical drone components. The FY2027 request seeks $20.2 billion for the credit program. That is credit-subsidy budget authority, not loan volume: the justification book models it as supporting roughly $200 billion in direct-loan capacity.[51]

FIGURE 05 · OFFICE OF STRATEGIC CAPITAL: SUBSIDY, CEILING AND CAPACITY

Four different kinds of number. Credit subsidy is appropriated cash; the ceiling is statutory loan principal; the request is budget authority; capacity is a modeled projection at an assumed subsidy rate.

Source: Public Law 119-21, section 20004(d), 4 July 2025; CRS, IF13215, 8 September 2026; Department of War, FY2027 Defense Strategic Capital Credit Program justification, April 2026; DoW releases of 10 August 2025 and 31 July 2026.

The instrument is credit, not equity, and that matters to how it lands. A loan against manufacturing equipment reaches a hardware company at a point in its life where venture structure serves it least, which is the mismatch set out in section 10, Venture Structure vs Capability Building. OSC also lends to third-party investment funds, and the SBIC Critical Technology Initiative run jointly with the Small Business Administration routes that capital through fund managers rather than around them.[52] For an investor weighing a hardware position, the question is no longer whether public capital exists but how it prices alongside private money in the same round.

13 The Budget Is Moving in the Same Direction

In the room: D2-03, From the Hill to the Board Room, Tue 11:05

The FY2027 request is the largest in American history. The total request is $1.5 trillion: $1.1 trillion in base discretionary Department of War authority plus $350 billion in mandatory reconciliation funding. Both armed services committees authorized roughly $1.14 trillion within their jurisdiction; CRS puts the House bill at $1,141.7 billion, matching the request, and the Senate bill at $1,140.2 billion. Neither addressed the $350 billion reconciliation tranche, which remained unresolved at the time of writing.[53] The munitions line described in section 05, Lagging Arms Manufacturing, $52.885 billion for the fourteen munitions tracked under the Munitions Acceleration Council against $10.828 billion enacted in FY2026, sits inside that request.[54]

Authorization is not obligation. The gap this brief describes is between capital raised and capital obligated, and a larger topline closes it only if the money reaches production contracts rather than accumulating in accounts the acquisition system cannot spend at speed. As this brief went out, neither the FY2027 NDAA nor full-year defense appropriations had been enacted: the House had passed H.R. 8800, the Senate bill remained at a motion to proceed, and a continuing resolution, Public Law 119-103, funds most programs at FY2026 levels through 11 December.[55] The operative figures should be checked against the enacted bills before anyone builds a plan on them.

14 The Validated-Marketplace Model Is Being Imported

In the room: D1-08, From Prototype to Fielding Capability, Mon 15:20; D1-06, Warfighter Demand Signal, Mon 13:50

The counter-UAS marketplace is the closest U.S. analogue to what section 06, What “Good” Looks Like Remains Unclear, describes. Joint Interagency Task Force 401 expanded its C-UAS marketplace on 3 August 2026, built with Kaizen Laboratories under a $15 million prototype Other Transaction Agreement that Kaizen reports was awarded on 8 May 2026. Authorized users search validated technologies by operational requirement, review performance data before committing, and purchase through a streamlined process integrated with existing acquisition pathways. The platform also feeds the Army’s Foreign Military Sales Fast Lane, where submitted products are evaluated and pre-cleared for export.[56] A validated performance record attached to a purchasable catalogue entry is a price signal in the sense this brief has used the term.

DIU’s Project GI takes the iteration argument directly from Ukraine. Launched on 2 June 2025 with a $20 million pool across three design reference missions, it took submissions on a rolling basis rather than to a fixed deadline, seeking high-readiness small and medium unmanned systems and keeping companies in contact with operators rather than contracting officers. DIU’s Blue UAS lead put the ambition plainly: "We want to start trying to solve the problem in 24 hours."[57] Measured against the iteration-velocity standard proposed in section 01, The Adversary Rewrites the Requirement Every Six Weeks, this is the first U.S. program to adopt the clock rather than the vocabulary.

Ukraine has opened its side of the loop. Brave1’s Test in Ukraine platform, launched in July 2025, invites foreign developers to trial systems with combat-tested units and returns structured performance reports, and BraveTech EU, announced on 11 July 2025 with up to €50 million in EU funding matched by €50 million from Ukraine, funds startups testing on the same terms.[58] Access is now a matter of programs rather than personal relationships, which lowers the cost of the direct-access advantage described in section 07, Lessons Learnt in Ukraine Stay in Ukraine, without removing the export frictions that surround it.

None of this is finished. The requirements rebuild is a year old, the largest OSC commitments are conditional, the FY2027 bill is unresolved, and the marketplaces are early. Taken together they suggest the system is moving toward the diagnosis in this brief rather than away from it, which makes the next two days a question of execution rather than persuasion.

05About This Brief

Palisade Bergschrund (Pal-Berg) is a cross-border solutions firm operating across the United States, Europe, Ukraine and allied markets. We help manufacturers, procurers, investors and governments navigate the most demanding challenges of the modern emerging and defense technology market.

The Pal-Berg team was built specifically to meet the needs of that market — integrating direct operational, military, diplomatic, technical and capital-markets experience to deliver the solutions leading innovators, investors and governments require in a shifting marketplace, with people on the ground across New York, California, London, Brussels and Kyiv.

The firm’s Research, Data and Analytics practice covers defense technology capital flows, procurement and policy movement, and the performance of fielded systems in active theaters. It assesses technologies against the evidence behind them and publishes recurring briefs alongside longer analytical work.

This brief was prepared for the New York National Security Forum 2026, hosted by the Sparrow Point Foundation. It draws on open sources, the Forum’s own attendee data, and the lead author’s field research in Ukraine. Luc Hillion is a co-author of Closing the Loop, the Harvard Kennedy School Belfer Center study cited throughout this brief. Views expressed are those of the authors.

06Authors

Jesse Nuese, Managing Partner, Research & Technology, Palisade Bergschrund

Jesse Nuese leads research and technology at Palisade Bergschrund and co-founded the New York National Security Forum. He is a doctoral researcher in Defence Studies at King’s College London, where his work examines private security and military innovation in Ukraine, Taiwan and Israel.

His career spans cyber threat intelligence at Recorded Future, international strategy work for Boeing Defense, Space & Security, and business development leadership at Cracken AI. He served as an infantryman and paratrooper with the 82nd Airborne Division, including a 2014 combat deployment to Afghanistan, and holds degrees from Columbia University, the Fletcher School at Tufts, and NYU.

Luc Hillion, Emerging Technology Researcher, Palisade Bergschrund

Luc Hillion is an Emerging Technology Researcher at Palisade Bergschrund. He co-authored two studies for the Harvard Kennedy School’s Belfer Center for Science and International Affairs, Closing the Loop: Improving Innovation Feedback Between Soldiers, Startups, and Strategy (2026) and Russian Threats to NATO’s Eastern Flank (2026), and conducted research on European affairs for Belfer Center Senior Fellow Daniela Schwarzer.

He previously worked at the OECD Nuclear Energy Agency, the UN Office of Counter-Terrorism and the Samuel B. Griffith Foundation for Chinese Military Studies. He holds a Master in Public Policy from the Harvard Kennedy School and a dual BA from Sciences Po and Columbia University.

07Notes

1. Drew DeJarnette, “DALO Dispatch: The Half-Life of Relevance,” Palisade Bergschrund, 27 August 2026, https://pal-berg.com/research/dalo-dispatch-the-half-life-of-relevance.
2. Belfer Center for Science and International Affairs, Harvard Kennedy School, Closing the Loop: Improving Innovation Feedback Between Soldiers, Startups, and Strategy, 29 June 2026. Band plans: 3GPP/ETSI TS 136 106 (LTE Band 3, 1710–1785 / 1805–1880 MHz); Federal Communications Commission, Advanced Wireless Services (AWS-1 and AWS-3 in the United States).
3. U.S. Coast Guard, contract award notice, 26 June 2024; Shield AI, press release, 1 July 2024; Jake Chapman, “Moving Toward Defense as a Service,” War on the Rocks, 29 November 2024.
4. Belfer Center, Closing the Loop.
5. Amanda Bresler and Alex Bresler, “Assessing the Effectiveness of Defense-Sponsored Innovation Programs as a Means of Accelerating the Adoption of Innovation Forcewide,” SYM-AM-23-101, Acquisition Research Program, Naval Postgraduate School, May 2023, Table 3. The dataset covers SBIR and STTR awards.
6. U.S. Government Accountability Office, Small Business Innovation Research: DOD’s Program Has Developed Some Technologies that Support Military Users, but Lacks Comprehensive Data on Transition Outcomes, GAO-14-748T, 23 July 2014.
7. DFARS 227.7104-2, Rights in SBIR or STTR data; DFARS 252.227-7018 (August 2025). Older contracts follow the clause version they incorporated.
8. Secretary of War, “Unleashing U.S. Military Drone Dominance,” memorandum, 10 July 2025.
9. Under Secretary of Defense (Comptroller), Financial Policy Memorandum 23-02, update to DoD 7000.14-R, Volume 2A, Chapter 1, 22 March 2023; Consolidated Appropriations Act, 2026, Division A, Title VIII, section 8039.
10. Congressional Research Service, Defense Primer: Research, Development, Test, and Evaluation (IF10553); U.S. Government Accountability Office, Small Business Research Programs: Clearer Guidance Could Improve Award Data to More Effectively Measure Outcomes, GAO-25-107942, 23 September 2025.
11. Belfer Center, Closing the Loop; 10 U.S.C. §§ 4401–4403; Army Science Board, Robotic and Autonomous Systems-of-Systems Architecture study, 2016; DoD Instruction 8330.01, 27 September 2022.
12. Belfer Center, Closing the Loop.
13. Department of the Army, FY2026 Missile Procurement justification book (PAC-3 MSE procurement unit cost $3.871 million); FY2027 request reporting; CSIS Missile Defense Project, Shahed-131/-136 profile, 2 June 2026; Hollenbeck et al., CSIS, 19 February 2025; TIME and BBC Verify reporting on Ukrainian USV unit prices (about $200,000 and an advertised $250,000); Jesse Nuese, “Intelligent Mass,” Palisade Bergschrund, 11 August 2026, https://pal-berg.com/research/intelligent-mass. Target values are analyst and media estimates.
14. DeJarnette, “The Half-Life of Relevance.”
15. U.S. Army, DVIDS, 8 May 2025; Belfer Center, Closing the Loop.
16. 7th Army Training Command, “Combined Resolve 26-07,” DVIDS, 7 May 2026; Isabel Coles and Nikita Nikolaienko, “Ukrainian Drone Operators Humbled a U.S. Army Brigade,” Wall Street Journal, 12 August 2026.
17. Jillian Kay Melchior, Wall Street Journal, 13 February 2026, on Exercise Hedgehog 25 (Siil 2025), Estonia, 5–23 May 2025; Estonian Defence Forces.
18. U.S. Department of War, Drone Dominance announcement, 2 December 2025; U.S. Department of War, “Department of War Enhances C-UAS Marketplace to Expand Access to Validated Technologies,” 3 August 2026.
19. Department of War Office of Inspector General, Evaluation of the DoW’s Capability and Capacity to Produce 155-Millimeter Artillery Ammunition, DOWIG-2026-095, 13 July 2026; Rheinmetall press releases, 13 January 2025 and 14 September 2026; Ellie Cook, “Germany Overtakes US in Ammunition Production Capacity,” Newsweek, 28 April 2026, relaying a claim by Rheinmetall chief executive Armin Papperger.
20. Lockheed Martin, 6 January 2026 (2025 PAC-3 MSE deliveries); Macdonald Amoah, Morgan Bazilian and Jahara Matisek, “Scaling Patriot Production: The Industrial Base Crisis Explained,” Foreign Policy Research Institute, 18 May 2026.
21. Mark F. Cancian and Chris H. Park, “Renewed Iran War Would Test Diminished Interceptor Inventories,” Center for Strategic and International Studies, 27 July 2026. Figures are CSIS estimates derived from budget documents.
22. Megan Eckstein, “Supplier bottlenecks threaten US Navy effort to grow arms stockpiles,” Defense News, 6 February 2024.
23. Seth G. Jones, “Is the United States Prepared for a War with China?” Center for Strategic and International Studies, 12 May 2026.
24. Mark F. Cancian and Chris H. Park, “Rebuilding U.S. Missile Inventory: A Multiyear Project,” Center for Strategic and International Studies, 27 May 2026.
25. Department of War, FY2027 Budget Request Overview Book, April 2026, Figure 3.4. The 388 percent increase is calculated from the unrounded totals.
26. Javaheri, “Ukraine’s Defense Tech Ecosystem: From Wartime Experimentation to Industrial Scale,” PitchBook, 1 July 2026; KSE Institute, “The Ukrainian defense technology market,” March 2026. The $6.8 billion is the KSE Institute estimate of 2025 market output across UAV, UGV and EW, not a measure of funding.
27. Daniel Kosoy, “Why Global Investors Are Pouring Millions Into Ukraine’s Combat-Tested Defense Tech,” UNITED24 Media, 15 December 2025.
28. Luke Smith, “Ukraine 2025 defence tech investment topped $57.2M, but the ‘funded market’ is $6.8B says PitchBook,” Resilience Media, 8 July 2026.
29. Deloitte Ukraine, “M&A in Ukraine’s defence sector: battlefield innovation meets investment reality,” 21 November 2025; interview with Dominique Piotet, managing partner, Ukraine Phoenix Tech Fund.
30. Brave1, About (accessed September 2026); Ministry of Economy of Ukraine, 27 April 2023. Official counts differ: the Ministry of Digital Transformation reported more than 3,000 companies in April 2026.
31. Ministry of Digital Transformation of Ukraine, release on Brave1 Market and e-Points, 10 July 2025.
32. Oleksandr Matviienko, “The Red Tape of Ukraine’s Semi-Open Arms Exports,” Lawfare, 16 April 2026.
33. Howard Altman, “Ukrainian Companies Prohibited From Exporting Shahed Interceptor Drones,” The War Zone, 13 March 2026.
34. Centre for Eastern Studies (OSW), “Exports under special supervision: Ukraine liberalises arms exports,” 28 May 2026, https://www.osw.waw.pl/en/publikacje/analyses/2026-05-28/exports-under-special-supervision-ukraine-liberalises-arms-exports; Cabinet of Ministers of Ukraine, Resolution No. 875; Matviienko, Lawfare.
35. Patrick Tucker, “US investors warm to Ukrainian defense startups, but export laws slow cooperation,” Defense One, 11 May 2026.
36. U.S. Department of War, Drone Dominance Program Gauntlet I results, 3 February 2026; Shephard, 12 March 2026.
37. Swarmer Inc., Form 424B4, U.S. Securities and Exchange Commission, 17 March 2026; Swarmer Inc., Form 8-K and Exhibit 99.1, 10 September 2026.
38. U.S. Department of War, Contracts for 14 July 2026; interview with Soren Monroe-Anderson, The Defender, 20 October 2025; Belfer Center, Closing the Loop.
39. Silicon Valley Defense Group, 2026 NatSec100 Report and release, 26 May 2026. SVDG publishes both totals; private capital is calendar 2025 and federal obligations fiscal 2025.
40. Silicon Valley Defense Group, 2026 NatSec100 Report, May 2026.
41. Center for Strategic and International Studies, “SBIR and STTR Reauthorization and the Future of Small Business Innovation,” updated 18 March 2026; AFWERX, “Phase III,” 2026; 15 U.S.C. § 638.
42. U.S. Government Accountability Office, Other Transaction Agreements: Improved Contracting Data Would Help DOD Assess Effectiveness, GAO-25-107546, 3 September 2025; 10 U.S.C. § 4022(f).
43. U.S. Government Accountability Office, GAO-26-108457 (July 2026) on IVAS; GAO-22-105387 on the Littoral Combat Ship; GAO-24-106546 on the Constellation-class frigate.
44. Belfer Center, Closing the Loop, p. 22.
45. National Venture Capital Association, model legal documents; Institutional Limited Partners Association, ILPA Principles 3.0.
46. “Defense tech startups had their best funding year ever in 2025,” Defense News, 20 January 2026, citing PitchBook data; Erik Kannike, “Beyond the hype: how to actually invest into defence startups, 6 lessons for success,” Medium.
47. Office of the Secretary of Defense, “Reforming the Joint Requirements Process to Accelerate Fielding of Warfighting Capabilities,” memorandum signed 20 August 2025 by Secretary Hegseth and Deputy Secretary Feinberg; reissued by the Office of the Secretary of War, 7 November 2025.
48. Chairman of the Joint Chiefs of Staff Instruction 5123.01J, Charter of the Joint Requirements Oversight Council and the Joint Force Requirements Process, 15 January 2026.
49. Office of the Secretary of War memorandum, 7 November 2025; National Defense Magazine, “Unpacking the Pentagon’s New Requirements Memo”; Covington, “JCIDS, Rewired: What DoD’s New Requirements Memo Means.”
50. 10 U.S.C. § 149; Public Law 118-31, section 903; Public Law 118-47; Public Law 119-21, section 20004(d); Congressional Research Service, Office of Strategic Capital: Overview and Considerations, IF13215, 8 September 2026; U.S. Department of Defense, “Office of Strategic Capital Announces First Loan Through DoD Agreement With MP Materials,” 10 August 2025.
51. Department of War, FY2027 Defense Strategic Capital Credit Program justification, April 2026; U.S. Department of War, Performance Drone Works conditional commitment release, 31 July 2026; Jon Harper, “Pentagon requests more than $20B for strategic capital loan program in 2027,” DefenseScoop, 6 April 2026.
52. CRS, IF13215.
53. Congressional Research Service, IN12703, 9 July 2026; H. Rept. 119-698; Senate Armed Services Committee report, 15 June 2026.
54. FY2027 Budget Request Overview Book, Figure 3.4.
55. Congress.gov, H.R. 8800, S. 4784 and H.R. 6500; Congressional Research Service, FY2027 appropriations status table, 18 September 2026. Status as of 22 September 2026.
56. U.S. Department of War, C-UAS Marketplace release, 3 August 2026; Kaizen Laboratories, release, 8 May 2026 (the OTA value is company-reported); U.S. Army, FMS Fast Lane, army.mil, 14 August 2026.
57. Defense Innovation Unit, “Project G.I.,” 2 June 2025; Courtney Albon, “Pentagon prize challenge seeks ‘ready-now’ uncrewed systems,” Defense News, 3 June 2025; Patrick Tucker, Defense One, 2 June 2025, quoting DIU Blue UAS lead Trent Emeneker.
58. Ministry of Digital Transformation of Ukraine, launch of the Test in Ukraine platform at LANDEURO, Wiesbaden, 17 July 2025; European Commission, BraveTech EU announcement, 11 July 2025.