Smack Tech Raises $32M for Defense AI

Smack Technologies' approach is rooted in the operational experience of its founders, Andrew Markoff and Clint Alanis, who bring over 20 years of combined combat experience as MARSOC veterans. Their work is a direct response to the observation that U.S. military decision-making processes have not kept pace with the speed of modern conflict. CEO Andrew Markoff even had a stint in Business Operations and Strategy at Palantir, a major player in defense data analysis. The company is developing domain-specific AI models using deep reinforcement learning, a departure from the large language models common in the commercial sector. This technique trains AI agents in proprietary synthetic warfare environments, allowing them to learn optimal strategies by simulating countless conflict scenarios. Their goal is to create physics-based reasoning models that can make complex time-space calculations under pressure. Smack's technology is split into two main product suites: "Omega," a command-level stack for campaign planning, and "Alpha," a platform with lightweight agents designed to run on tactical edge hardware. This dual approach addresses the need for high-level strategy and in-the-field decision-making, even with degraded communications. The company has already secured contracts with the Joint Fires Network (JFN) and the Marine Corps Warfighting Lab (MCWL). The $32M funding round was co-led by Geodesic Capital and Costanoa Ventures, with participation from notable investors like Point72 Ventures and Bloomberg Beta. This investment highlights a significant trend of venture capital flowing into defense technology, a sector that saw a record $49.1 billion in VC deals in 2025. This surge is driven by geopolitical instability and the demonstrated effectiveness of technologies like AI and drones in modern conflicts. The push for "decision-dominance" necessitates a robust and secure hardware foundation, creating a complex "build vs. buy" dilemma for the Department of Defense (DoD) and its contractors. To ensure a supply of trusted, U.S.-made semiconductors, the DoD has awarded multi-billion dollar contracts to foundries like GlobalFoundries and Intel. Intel's RAMP-C program, for instance, is specifically designed to create a U.S.-based commercial foundry ecosystem for critical DoD systems. For edge applications like Smack's "Alpha" platform, the hardware choice is critical and often involves a mix of GPUs, FPGAs, and potentially custom ASICs. FPGAs (Field-Programmable Gate Arrays) are particularly valuable in defense for their reconfigurability, allowing hardware to be adapted to new threats in the field. Companies like NVIDIA are dominant in the AI training space, but the need for low-power, rugged, and secure chips at the tactical edge creates opportunities for specialized hardware solutions. The GTM motion for companies selling into this space involves navigating a complex procurement process that the DoD is actively trying to streamline. Startups like Smack often gain traction by securing contracts with innovation-focused military branches like the Marine Corps Warfighting Lab, which is actively seeking troops with AI and drone expertise to experiment with new technologies. Success here can lead to larger, program-of-record contracts as the technology proves its value. The competitive landscape includes established defense tech giants like Palantir and heavily-funded startups such as Anduril Industries, which has raised billions to build autonomous systems. Smack aims to differentiate itself through its founders' deep operational expertise and its focus on deep reinforcement learning models specifically for military decision-making, rather than repurposing general-purpose AI.