Mission Autonomy · Uncrewed Aircraft · Simulation
Shield AI
Mission-autonomy software and vertical-takeoff aircraft for contested environments
PRIVATE
No public ticker
No public ticker. Valuation marks come from disclosed funding rounds and secondary transactions; see the verified metrics below.
Shield AI is privately held. The $12.7 billion figure is the historical post-money mark attached to the financing package closed by June 2026, not a live market value. The $591.8 million Form D amount covers the exempt equity offering identified in that filing and should not be read as the entire $2 billion capital package. Absolute revenue, margins, cash, share count and backlog remain undisclosed.
Shield AI builds mission-autonomy software, uncrewed aircraft, electro-optical sensing software and simulation systems. Hivemind is the central software layer. It is designed to let an aircraft or other uncrewed system perceive conditions, select actions and execute a mission under human direction without continuous communications or access to the Global Positioning System. Shield AI sells the software for integration on customer and partner platforms while also using it on the company’s V-BAT and developmental X-BAT aircraft.
This mixed model matters to procurement. Software can move across aircraft through the government-owned Autonomy Government Reference Architecture, while owned hardware gives Shield AI operational data and a route to aircraft and services revenue. Aechelon simulation and the company’s Vision Systems products extend the development loop from synthetic training through sensing and field operations. The principal customers are U.S. military services, the U.S. Coast Guard and allied governments seeking intelligence, surveillance and reconnaissance, collaborative combat capability and runway-independent aircraft.
Platform-agnostic software that assumes functions normally performed by a pilot or vehicle operator. Hivemind includes mission behaviors, autonomy middleware, command tools and integration tooling for systems that must operate when communications or satellite navigation are degraded. The U.S. Air Force placed Shield AI in its six-vendor Collaborative Combat Aircraft mission-autonomy pool and awarded the company one of three opening production options. Hivemind has flown on Anduril’s YFQ-44A, the General Atomics MQ-20 Avenger, the Kratos MQM-178 Firejet and other test platforms.
Software products serving two distinct workflows. Shield AI describes Hivemind Enterprise as a suite for building, testing and deploying autonomy, spanning middleware, mission control, autonomy catalogs and development tools. Benchmark uses training data, digital grading rubrics and AI-assisted analysis to assess aircrew performance, readiness and safety. These products support software and analytics revenue independent of Shield AI aircraft, but pricing, customer counts and segment revenue are not publicly disclosed.
Group 3 ducted-fan vertical-takeoff and landing uncrewed aircraft for expeditionary and maritime intelligence, surveillance, reconnaissance and targeting. Shield AI advertises endurance above 12 hours with an electro-optical and infrared payload, a 40-pound maximum payload and launch from a 15-foot-by-15-foot area. The aircraft supports the U.S. Coast Guard under a five-year contractor-owned, contractor-operated services vehicle with a $198.1 million ceiling. V-BAT has also operated in Ukraine, aboard U.S. Navy and Coast Guard vessels and with international customers. Shield AI does not publish a standard unit price.
Jet-powered vertical-takeoff and landing aircraft unveiled in October 2025 for runway-independent strike, counter-air, electronic warfare and intelligence missions. Shield AI lists design specifications exceeding 2,000 nautical miles of maximum range and 50,000 feet of ceiling. These are developmental targets rather than field-validated fleet performance. In August 2026 the Department of the Navy Rapid Capabilities Office, Portfolio Acquisition Executive for Aviation and Defense Innovation Unit identified X-BAT in a $50 million Runway Independent Maritime Expeditionary Strike investment. The public announcement did not disclose Shield AI’s contract vehicle, funded share or production quantity.
High-fidelity visual simulation, physics-based sensor modeling and synthetic-environment technologies acquired in June 2026. Aechelon continues to serve existing customers while working with Hivemind. Its systems support training, mission rehearsal, autonomy development and the Department of Defense Joint Simulation Environment.
Passive optical detection and wide-area search software built around technology acquired with Sentient Vision Systems in 2024. The ViDAR product uses electro-optical or infrared imagery and machine learning to detect and classify targets over land and sea. On V-BAT, it expands the aircraft from a narrow camera field of view into a wider maritime and terrestrial search system.
Shield AI was founded in San Diego in 2015 by Brandon Tseng, Ryan Tseng and autonomy engineer Andrew Reiter. Early work produced the Nova small uncrewed aircraft system for indoor reconnaissance. The company then shifted toward a broader mission-autonomy stack designed to operate at the edge when communications and satellite navigation are unavailable.
Acquisitions supplied capabilities that would have taken years to build independently. Shield AI acquired Heron Systems in July 2021 for reinforcement-learning and air-combat autonomy. It added Martin UAV and the V-BAT aircraft later in 2021. Sentient Vision Systems joined in 2024, bringing ViDAR passive wide-area detection. Aechelon Technology closed on 22 June 2026, adding high-fidelity visual and sensor simulation. Transaction values were not publicly disclosed for these acquisitions.
Leadership changed in 2025. Gary Steele, previously chief executive of Splunk and Proofpoint, became Shield AI’s chief executive and a director on 13 May 2025. Ryan Tseng moved from chief executive to President, Co-Founder and Chief Strategy Officer. Brandon Tseng remains President and Co-Founder.
The Air Force selected Shield AI for Collaborative Combat Aircraft mission-autonomy work in February 2026 and awarded it an opening production option in June. That award keeps Hivemind in a funded competition with Anduril and RTX Collins rather than establishing Shield AI as the final autonomy supplier. Also in June, Shield AI confirmed the close of a $2 billion financing package at a $12.7 billion post-money valuation and completed the Aechelon acquisition.
Shield AI does not publish audited financial statements or absolute revenue. In a May 2026 company update, management said FY2026 orders and revenue doubled year over year. That establishes the direction and approximate rate of growth but not the revenue base, customer mix, gross margin, cash burn or program profitability.
Four lanes drive the business. Hivemind creates software-license and integration revenue across Shield AI and third-party platforms. V-BAT generates aircraft, support and contractor-operated services revenue, including funded orders under the Coast Guard vehicle. Vision Systems and Aechelon add sensors, simulation, training and synthetic-environment work. X-BAT remains a development program whose future revenue depends on flight performance and the conversion of Navy and allied interest into funded procurement.
The Coast Guard indefinite-delivery, indefinite-quantity contract provides a five-year ordering path, not $198.1 million of guaranteed revenue. The CCA contract also preserves competition rather than guaranteeing Shield AI the primary software role. International expansion can diversify demand, but export approvals, local-content requirements and government procurement schedules determine when announced partnerships become recognized revenue.
Shield AI does not publish a complete current board roster. Company announcements confirm that Gary Steele joined the board with his appointment as chief executive, Advent International Chairman David Mussafer was to join in connection with the 2026 Series G and JPMorganChase executive Todd Combs received observer status. Shield AI’s Form D filings identify certain directors and related persons for exempt-offering purposes, but a Form D is not a proxy statement and should not be treated as a complete real-time governance roster.
| Filed | Document | Registry / Authority | Link |
|---|---|---|---|
| 2026-05-01 | Form D (exempt offering) | SEC EDGAR | View → |
| 2024-01-08 | Form D/A | SEC EDGAR | View → |
| Ongoing | Historical Form D and Form D/A notices | SEC EDGAR | Company filings → |
| Date | Customer / Agency | Program / Vehicle | Disclosed Value | Funding & Status |
|---|---|---|---|---|
| 13 Aug 2026 | Department of the Navy Rapid Capabilities Office / PAE Aviation / DIU | RIMES · X-BAT runway-independent maritime strike | $50 M program investment | Program-backed Shield AI’s funded share, vehicle and quantity undisclosed |
| 17 Jun 2026 | U.S. Air Force | CCA Increment 1 mission autonomy · six-year pool and opening production option | Undisclosed | Competitive phase One of 3 opening options; primary-provider selection planned summer 2027 |
| 19 May 2026 | Office of the Under Secretary of War for Research and Engineering | LUCAS · Hivemind integration for coordinated one-way attack systems | Undisclosed | Selected Operational demonstration announced for fall 2026 |
| 26 Jun 2024 | U.S. Coast Guard | V-BAT contractor-owned, contractor-operated maritime UAS services · IDIQ | $198.1 M ceiling | Active through 2029 Approximately $56.8 M in combined current award amounts shown by USAspending at the 19 Aug 2026 review |
| 11 Feb 2026 | Taiwan National Chung-Shan Institute of Science and Technology | Hivemind integration, demonstration and deployment on Taiwan-developed systems | Undisclosed | Contract signed Includes software-development-kit and turnkey integration work |
| 23 Jun 2026 | Polish Navy / local industrial partners | V-BAT maritime operations support | Undisclosed | Announced Quantity, vehicle and funded value not publicly disclosed |
N/A - no Section 16 reporting. Shield AI has no registered class of equity securities, so no Forms 3, 4 or 5 exist. The capital history below is the closest available substitute and is the primary governance signal for a private issuer.
| Date | Round | Lead / Notable Investors | Amount | Post-Money |
|---|---|---|---|---|
| Closed by Jun 2026 | Series G + preferred financing | Advent International, JPMorganChase, Blackstone | $1.5 B equity + $500 M preferred | $12.7 B |
| 6 Mar 2025 | Series F-1 | L3Harris, Hanwha Asset Management, Andreessen Horowitz, USIT and others | $240 M equity | $5.3 B |
| Expanded 29 Dec 2023 | Series F + debt | USIT, Riot Ventures, ARK Invest, Hercules Capital and others | $300 M equity + $200 M debt | $2.7 B equity-round valuation |
| 9 Jun 2022 | Series E | Snowpoint Ventures, Riot Ventures, Disruptive, Homebrew and others | $90 M equity + $75 M debt | $2.3 B |
| 25 Aug 2021 | Series D | Disruptive, Point72 Ventures and others | $210 M equity closed | More than $1 B |
Exact ownership percentages, voting rights and founder control are not publicly disclosed. The 2026 financing announcement says Advent International Chairman David Mussafer would join the board and JPMorganChase executive Todd Combs would become a board observer. Blackstone’s $500 million preferred investment was accompanied by a separate $250 million delayed-draw commitment. That commitment is future financing capacity, not capital already raised and not part of the $12.7 billion post-money equity valuation.
Shield AI has reached the point where its software is funded inside the Air Force’s most consequential autonomy competition while V-BAT provides an operational aircraft and services base. The next test is conversion. Hivemind must win or retain meaningful CCA work, X-BAT must progress from engine and subsystem tests into credible flight demonstrations and international agreements must become funded orders.
Shield AI completes both CCA competitive phases, retains a funded role even if another vendor becomes the primary provider, expands Coast Guard and allied V-BAT orders and uses Aechelon to shorten simulation-to-flight development. X-BAT reaches vertical-flight milestones but remains in prototype and evaluation work rather than large-scale production.
Hivemind becomes the Air Force’s primary CCA Increment 1 autonomy supplier or wins licenses across multiple airframes. LUCAS demonstrates coordinated one-way attack operations, Taiwan and European partnerships produce recurring software work and the Navy converts RIMES testing into a funded X-BAT development and production path.
Shield AI loses the primary CCA role, V-BAT orders convert slowly, X-BAT slips on vertical-flight or payload integration and the company absorbs high aircraft-development costs without a program of record. Aechelon and Sentient integration could also divide engineering attention while private-company reporting leaves investors and customers with limited visibility into cash consumption and program margins.
Watch the second CCA competitive award, the planned summer 2027 primary-provider selection, X-BAT vertical flight and weapons-integration milestones, funded RIMES documents, Coast Guard task orders against the $198.1 million ceiling, the fall 2026 LUCAS demonstration, repeat V-BAT orders and evidence that Aechelon simulation is reducing test time or certification risk.
Hivemind’s strongest strategic feature is portability. It has flown aircraft from Shield AI, Anduril, General Atomics and Kratos. A larger platform set gives the company more integration data and prevents the business from depending entirely on one airframe. The same government-owned interfaces that open those platforms also make replacement easier, so openness expands the market while limiting lock-in.
V-BAT gives Shield AI an operational aircraft, maritime references and data from communications-contested environments. That is more defensible than an autonomy demonstration confined to simulation. The advantage depends on data rights, customer permissions and whether lessons transfer from a Group 3 intelligence aircraft to larger combat platforms.
Aechelon can connect visual simulation, sensor models, Hivemind training and live-flight results. If integration works, Shield AI can test more behaviors before consuming scarce aircraft and range time. The moat will be measured in faster integration, safer testing and accepted capability, not the acquisition itself.
Shield AI has a credible technical and procurement position, not an assured monopoly. Anduril has comparable software ambition and greater portfolio breadth. Traditional primes retain deeper certification and sustainment systems. The decisive evidence will be CCA performance, repeat software licenses, V-BAT task-order conversion and X-BAT flight results.
CCA competition: Shield AI holds an opening production option but has not won the primary Increment 1 role. The Air Force can also license software from any of the six pool members during the six-year vehicle. A loss in summer 2027 would weaken the largest visible software opportunity without necessarily removing Shield AI from the pool.
Technical validation: Collaborative autonomy must work across multiple aircraft, dynamic threats and degraded communications while preserving human command authority. Successful demonstrations do not by themselves prove fleet-scale reliability, cybersecurity, weapons safety or certification.
Aircraft-development capital: X-BAT adds propulsion, thermal, structural, flight-control, payload and manufacturing risks to a company already scaling software and V-BAT. The $50 million RIMES announcement should not be mistaken for a production order or sufficient funding for the full program.
Contract conversion and concentration: The Coast Guard’s $198.1 million figure is a ceiling, CCA values are undisclosed and many international agreements omit quantities. A small number of government decisions can therefore change growth faster than the public contract totals suggest.
Integration and transparency: Heron, Martin UAV, Sentient and Aechelon must operate as one product system rather than separate acquisitions. Private ownership leaves margins, cash burn, backlog, debt terms, customer concentration and voting control outside the public record.
Export and policy exposure: Allied sales depend on export licensing, technology-security rules and local industrial arrangements. Policy governing autonomous weapons, human release authority and artificial-intelligence assurance can change program requirements after development begins.
Material strategic dependence remains concentrated in U.S. Air Force CCA outcomes, Coast Guard V-BAT task orders and a limited set of allied programs. Hivemind can diversify across platforms, but Shield AI does not disclose enough revenue or backlog data to quantify that diversification.