A Navy drone that has executed hundreds of targeting missions in Ukraine while GPS was being jammed every single flight just entered the competitive pool for up to $800 million in U.S. Navy surveillance contracts. Shield AI's V-BAT was selected on April 20 by Naval Air Systems Command to compete for persistent ISR task orders under a contractor-owned, contractor-operated services framework. The contract runs through August 2031, and it is not a technology demonstration. It is a program of record. The Navy already knows what it is getting. It has been getting it from other customers for three years.
The COCO model matters more than the dollar figure. Under this arrangement, Shield AI owns the aircraft, operates them, maintains them, and trains the crews. The Navy pays per-mission or per-hour, not per-platform. This strips away the acquisition risk, the depot maintenance burden, and the crew pipeline problem that have constrained unmanned ISR for decades. It is why the Coast Guard awarded Shield AI a separate $198 million COCO contract in 2024 and why that precedent is being replicated at scale. Four other vendors compete in this pool: AeroVironment, Insitu, Textron, and one other. Textron brings the Aerosonde Mk. 4.7, which has logged hundreds of thousands of operational hours and is already integrated into at least 11 Navy vessels. That is not a small thing. AeroVironment and Insitu have their own installed bases. Shield AI has the V-BAT and a specific operational story that the others do not.
The V-BAT is a Group 3 UAS: 161 pounds gross weight, 12.5 feet long, 9.6-foot wingspan, ducted-fan VTOL powered by JP-5 heavy fuel. It carries up to 40 pounds of payload, operates for more than 12 hours, and needs only a 12-by-12-foot footprint to launch and recover. Those specifications are published and unclassified. What matters is what has been done with them. Shield AI's operational claim is granular: more than 100,000 pounds of narcotics interdicted in the Caribbean and Pacific. Hundreds of targeting operations in Ukraine where GPS denial and electronic warfare are constant conditions. The V-BAT participated in UNITAS 2025 aboard USS Cooperstown. The Royal Netherlands Navy declared the V-BAT system operational in March 2026 and is acquiring 12 systems for eight vessels. The Japan Maritime Self-Defense Force operates it. Frontex, the European border agency, operates it. These are not concept demonstrations. These are operational deployments by actual military customers with real accountability.
What created the conditions for this pool entry in April 2026 is structural and recent. The U.S. Navy has struggled with organic ISR capacity on smaller vessels and in contested environments where manned aviation is too expensive or too vulnerable. The Aerosonde franchise (Textron) filled that gap starting years ago, but Textron is not designed for rapid scaling under COCO terms. AeroVironment owns a different market segment. The combination of V-BAT's operational maturity, Shield AI's operational autonomy software stack, and the COCO contracting model created an opening. Shield AI's Series G round closed in March 2026 at a $12.7 billion valuation, a 140% increase year-over-year, with capital co-led by Advent International and JPMorgan Chase. The company is projecting more than 80% revenue growth by end of 2026, which translates to at least $540 million in annual run rate. The Navy contract pool does not generate that revenue alone, but it legitimizes the business model at scale and demonstrates the market exists beyond prototyping.
Textron wins if the Navy weights operational maturity and installed-base integration heavily in task order competitions. Textron's Aerosonde is not just a drone; it is a data architecture already embedded in 11 ships. The institutional lock-in is real. AeroVironment and Insitu have their own advantages in certain mission sets and regional commands. Shield AI's position is the opposite: it has no installed base to defend and everything to gain from proving that the V-BAT can deliver equivalent or superior performance on cost-per-effect terms while requiring no legacy integration overhead. The COCO model actually favors this dynamic. The Navy does not need to re-certify platforms across the fleet; it simply awards task orders to the lowest-cost, highest-capability vendor for each mission. That is a competitive advantage for a company whose business model is built on operating hardware rather than selling it.
Here is the actual story: Shield AI has moved from a promising commercial drone company with military users into the position of an active competitor for a major Navy services contract. That is not because the Pentagon suddenly decided to de-risk small companies or because autonomous software is finally ready. It is because V-BAT has three years of operational track record, the COCO model transfers risk away from the Navy, and Shield AI's capital position ($2 billion raised in the last year) means it can actually absorb the working capital demands of sustained operations. The company is not asking the Navy to bet on the technology. The Navy is being asked to pick a vendor for ISR services, and one of the options has already done this work in Ukraine under GPS jamming. That is not a marginal difference. The question now is whether Shield AI's software autonomy stack — the fusion of advanced autonomy with proven hardware that Jetstream Venture Fund cited in its April 27 investor commentary — actually produces better targeting decisions or faster mission execution than a mature Aerosonde fleet already wired into carrier strike group operations. The COCO model does not guarantee that Shield AI wins task orders. It guarantees that the competition happens on operational merit, not acquisition risk, and that is the condition that favors the company with the better platform and the lower cost structure.
Watch three specific milestones. First, the initial task order awards — expected within 90 days of the September 1, 2026, contract start date. The Navy will reveal how heavily it weights existing integration versus new capability. Second, operational performance data from the first 12 months: mean time between failures, mission success rates by environment (contested or denied), and cost-per-flight-hour relative to the Aerosonde baseline. Shield AI's operational claims are auditable; the V-BAT's performance in the Caribbean interdiction mission and in Ukraine will become institutional knowledge within Navy ISR command. Third, the acquisition rate for additional V-BAT systems by foreign military services. The Royal Netherlands Navy order for 12 systems signals demand, but the real indicator is whether the Israel Defense Force, the UK Ministry of Defence, or South Korea commits to operational inventory. That would tell you whether the COCO model is replicable at scale internationally and whether Shield AI's capital position (and the Blackstone preferred equity deal) is securing long-term manufacturing capacity. All three of these are measurable within 18 months.