G2 Systems & Technologies

**Product and the operational problem.** G2 Systems & Technologies is developing a cognitive and edge-computing layer for systems that must make decisions close to the point of action. Its publicly disclosed product program, Sentinel, is an autonomous perimeter-security system intended to coordinate drones and cameras for monitoring, detection, and response. That addresses a concrete gap in conventional security architecture: a fixed camera network can observe a large area but cannot quickly reposition, while a drone can extend coverage but requires a communications link, an operator, and a separate workflow for interpreting its feed. G2’s stated objective is to turn those separate assets into a coordinated system that can process sensor data locally, trigger autonomous workflows, and retain human oversight for consequential decisions. The use case is especially relevant where backhaul is intermittent, response time matters, and sending every video frame to a remote cloud is impractical. The company describes its target domain as defense, homeland security, and mission-critical environments, while its edge-cognition positioning also leaves room for civilian critical-infrastructure and industrial deployments.

**Core technology and how it works.** Public descriptions place G2 at the intersection of artificial intelligence, hardware acceleration, and real-time data processing rather than presenting it as a general-purpose cloud AI vendor. In the Sentinel proof of concept, G2 is leading system-level platform development: it is responsible for the software stack that handles high-load data processing, monitoring, and autonomous response workflows across unmanned systems. The disclosed implementation integrates GSI Technology’s Gemini-II Associative Processing Unit, whose compute-in-memory architecture processes data near where it resides and is intended to reduce data movement for low-power inference. This is a partner component, not evidence that G2 manufactures the processor. GSI reported preliminary time-to-first-token results of about three seconds for a multimodal edge workload using video and text at approximately 30 watts, and said the chip was selected through a competitive evaluation for the Sentinel work; these are GSI-reported benchmarks and should not be treated as an independent Sentinel performance result. The architectural logic is nevertheless clear: local perception and decision support can reduce latency, bandwidth demand, and dependence on a continuously available command link, while a human remains in the approval loop for sensitive actions.

**Market, customers, and go-to-market.** G2 is pursuing a B2G and defense-industrial route into customers that buy systems rather than standalone AI APIs. Startup Nation Finder classifies it in Aerospace, Defense & HLS and describes Sentinel as coordinating drones and cameras for perimeter security; IVC lists homeland security and critical infrastructure as target markets. The likely initial buyer set therefore includes defense and homeland-security agencies, infrastructure operators with exposed sites, and prime contractors or unmanned-system manufacturers that need an autonomy layer for their platforms. The company’s most concrete public commercial channel is the GSI Technology relationship, where G2 contributes the system platform and GSI supplies the edge-computing subsystem. CTech reported that G2 planned to expand operations toward the United States and Europe, including subsidiaries to market its AI-based defense products. The public record does not name an end customer, recurring contract, production deployment, or revenue figure. The government-backed Sentinel evaluation is better understood as a qualification and reference-building motion: if successful, it could provide a system demonstration and a route to design wins, but it is not itself proof of scaled procurement.

**Traction, funding, and third-party validation.** G2 was founded in January 2025 and raised a seed round reported at $6 million by CTech and $6.2 million by Startup Nation Finder, with Stratos Ventures identified as the investor. The company then became visible through the January 2026 GSI announcement of a government-funded proof of concept backed by the U.S. Department of Defense, described in that release as the Department of War, and a foreign government agency. The parties said total governmental funding for the evaluation was on the order of millions of dollars; GSI expected to receive roughly $1 million for software optimization and integration, while its later earnings-call transcript said G2 was receiving additional funds to develop the drone platform. GSI also stated that G2 had conducted a competitive evaluation before selecting Gemini-II and that a successful government demonstration could lead to a design win, but explicitly cautioned that a proof of concept might not result in a commercial contract or recurring revenue. As of this record’s crawl date, public sources reviewed do not confirm the outcome of that later demonstration, a production award, or a named government customer. The combination of institutional seed capital, a listed Israeli company registration, an active official website, and a government-supported integration project is meaningful early validation, but still a pre-scale evidence base.

**Founders and team background.** The founding team gives G2 unusually direct exposure to both Israeli defense R&D and commercial technology building. CEO and CTO Ryan Gity is described by Gvahim as a reserve colonel and a technology and defense-innovation leader with more than two decades at Israel’s Directorate of Defense Research and Development, where he held senior roles in autonomy, artificial intelligence, and strategic planning. CTech reports that one of his final positions was heading the Autonomy and AI Department and that he later served as secretary of the Nagel Committee, which examined Israel’s defense budget for the following decade. Dror Feldheim, G2’s strategy co-founder, previously co-founded Trax in 2010 with Joel Bar-El; Trax developed business intelligence from smartphone images and became a prominent Israeli computer-vision company. Amir Geva is listed as a co-founder in IVC and Startup Nation Finder. CTech identifies him as an Israel Aerospace Industries executive vice president and former U.S. Department of Defense R&D attaché. That background is strategically valuable but also creates a governance diligence issue: IAI said Geva obtained permission to participate only as a passive shareholder and not an operating executive, while an investor deck reportedly listed him as president; Geva said the title was an editing mistake. Public sources disagree on headcount, ranging from 1–10 to 25, so organizational depth is not yet confirmable.

**Competitive dynamics and edge.** G2 competes in a layered market rather than against one identical product. NVIDIA’s Jetson and IGX platforms, Qualcomm’s edge-AI systems, and Israeli accelerator company Hailo offer compute foundations for low-latency inference. Anduril’s Lattice provides a more mature software layer for fusing sensors and coordinating autonomous defense assets, while Palantir’s AIP and mission software compete for the command-and-decision layer. Israeli primes including Elbit Systems, Rafael Advanced Defense Systems, and IAI can build comparable mission systems in-house and bring established procurement relationships, integration capacity, and certifications. G2’s plausible edge is its system-level focus on bringing an AI cognitive layer to constrained autonomous assets, with a design that can combine heterogeneous drones, cameras, and edge compute while keeping operators in control. Ryan Gity’s defense-R&D background may shorten the path from operational requirement to demonstrator, and the GSI relationship gives the startup a concrete hardware-acceleration partner. Neither advantage is yet a moat: incumbents can bundle software with platforms, accelerator vendors can move up the stack, and G2 must prove that its orchestration layer improves mission outcomes rather than merely adding another integration surface.

**Defense, security, and resilience relevance.** G2 qualifies as a direct dual-use company because the disclosed core system is designed for defense and homeland security while the same technical architecture can serve civilian critical-infrastructure protection. A locally processed camera-and-drone network could support airport, seaport, energy, water, communications, logistics, or industrial perimeters where fast anomaly detection and limited connectivity matter. In defense, the transfer path is even more direct: distributed surveillance, border monitoring, base protection, convoy overwatch, and unmanned-system coordination all benefit from low-power inference and reduced dependence on a remote data center. The human-oversight design is important because it frames autonomy as decision support and workflow coordination rather than an unbounded weapon-control claim. Strategic value comes from strengthening an allied supply of edge intelligence for systems operating in contested or bandwidth-limited conditions, potentially reducing reliance on centralized infrastructure and enabling faster iteration across existing platforms. The calibration is equally important. G2 has not publicly disclosed a fielded defense system, combat validation, classified program, safety certification, cybersecurity accreditation, or autonomous engagement capability. Sentinel is a government-backed evaluation, not proof of deployment. Its strategic relevance is therefore high, but its current demonstrated capability remains early and bounded.

**Stage, trajectory, and diligence risks.** G2 is classified as early: it is a 2025-founded Seed company with an R&D-stage product, limited public operating history, no disclosed revenue, and no publicly named production customer. The trajectory is attractive if Sentinel converts from a funded demonstration into a repeatable platform adopted by drone makers, defense primes, and critical-infrastructure operators. It could become an enabling layer across multiple unmanned systems rather than a single-purpose perimeter product, but that expansion depends on proving portability, reliability, and maintainability outside a controlled evaluation. Key diligence points are: (1) obtain independent Sentinel test results, including detection quality, false alarms, latency, degraded-link behavior, and operator workload; (2) clarify ownership and licensing of software, data, and any jointly developed IP with GSI and government sponsors; (3) verify the status of the promised government demonstration and whether it produced a contract or only a technical milestone; (4) understand how G2 will support certification, secure updates, export controls, and classified or sensitive deployments; (5) resolve the IAI conflict-of-interest history and the founders’ current operational roles; and (6) reconcile employee-count and funding discrepancies. The company has a credible strategic wedge and a strong founding profile, but its valuation, scalability, and field readiness remain unverified.