Bone AI has raised $12 million to expedite work on a centralized platform for autonomous air, ground and maritime systems, a development that places the Seoul- and Palo Alto–based startup in competition with entrenched Asian defense manufacturers. The company is betting that “physical AI” — the melding of autonomy software with purpose-built hardware and scaled production — will pave a path toward more rapid deployments of unmanned systems to defense and government buyers at scale across allied markets.
A strategy based on physical AI and vertical integration
Instead of treating autonomy as a software layer slapped on top of third-party platforms, Bone AI is building a vertically integrated stack from next-gen simulation and onboard AI down to embedded systems, sensor fusion, and production-grade hardware design. In this effort, the co-founder — who was also one of the co-founders of MarqVision — describes his company’s mission as being primarily about building up a domestic supply chain for physical AI in Korea and then exporting that to the U.S., Europe and partner countries.
That approach builds on South Korea’s widely acknowledged strengths in cars, shipbuilding, semiconductors and precision parts — advantages that might translate into reliable, relatively inexpensive unmanned systems. Korean defense leaders’ combined backlogs are estimated at just under $69 billion in media reports as well, indicating that demand is not confined to Korea alone. But there’s actually a shortage of such companies in the country to feed this defense-tech startup ecosystem, meaning that newcomers can scale rapidly if they can work with local suppliers who have previously proven themselves viable. Bone AI’s “buy versus build” playbook looks to ingest specialized hardware capabilities, and accelerate time-to-field.
Early traction with government buyers and programs
Though Bone AI has a portfolio in mind that will span UAVs, UGVs and USVs, it is launching with defense-oriented rotorcraft for logistics resupply, wildfire detection and counter-drone missions. The startup claims to have already landed a seven-figure B2G contract and made $3 million in its first year. And it’s been chosen as part of a South Korean government-endorsed end-to-end logistics program that will put autonomous air and ground vehicles using Bone’s autonomy stack at scale.
Operationally, the emphasis is on practicality: payload-agnostic platforms with modular bays and strong datalinks; essentially a group of drones currently metaphorically held together with duct tape.
In brief, these are systems which exploit gradual degradation autonomy in contested environments. Lessons learned from recent wars have demonstrated the impact of swarming, rapid attrition and electronic warfare on unmanned performance. Bone AI is designing for that, prioritizing robust navigation, multi-vehicle coordination and counter-UAS detection as essentials not add-ons.
Riding a global realignment in defense procurement
Europe’s rearmament has redrawn the maps of procurement, making South Korea a significant supplier of artillery and vehicles to members of NATO, according to SIPRI and public tenders. This has consolidated relations through the EU–South Korea Security and Defence Partnership; large orders in Poland and other markets have demonstrated that Korean industry can deliver at pace. Despite this industrial momentum, startups have been absent — a vacuum that Bone AI hopes to fill by marrying Korea’s manufacturing base with Silicon Valley autonomy talent.
The startup will still have to contend with long procurement cycles, export controls and stringent safety cases. But the upside is big: standardized interfaces and software-defined capabilities enable the certification of one core autonomy across a variety of platforms and geographies. It can also minimize integration friction with existing command-and-control and ISR workflows by building to open architectures widely employed by other allied forces, she said, including NATO STANAG profiles and modular open systems approaches.
Taking on the defense primes’ new look and models
Bone AI is entering a landscape transformed by venture-backed defense firms, which deliver quickly, iterate in software and vertically integrate. In the U.S., Anduril has been the poster child for sensor and autonomous systems integration, in Europe defense AI specialist Helsing has become a high-profile start-up for a software-led modernization. Investor estimates would give those companies private multibillion-dollar valuations, highlighting the scale of the opportunity for software-defined defense.
The name of the game for the potential success of Bone AI is whether it can deliver cross-domain autonomy and hardware at a level cheaper than Western counterparts that won’t break down. If it can convert Korea’s supply chain into a repeatable pipeline for the raw airframes, power systems and mission payloads on its list — along with allies comfortable enough to buy those in volume — it could offer an appealing package of price, performance and delivery speed for potential customers, especially for “attritable” platforms designed to be fielded in large numbers.
What the $12M funding fuels next for Bone AI
The new funds will be directed to autonomous R&D, flight testing, certification and manufacturing ramp.
On the horizon: greater endurance of logistics drones, closer fusion of EO/IR and RF payloads for counter-UAS, and demonstrations that bring air and ground robots under a common control layer. The one to monitor is new military contracts for repeat governmental orders on operational trials, not just pilots — that would be the clearest sign yet that Bone AI’s physical AI thesis is turning into fielded capability.