The next phase of defense AI may be less about putting autonomy into individual drones and vehicles and more about coordinating decisions across an entire force. Smack Technologies has raised $61 million in Series B funding to expand its AI platforms from command-level planning to communications- and compute-constrained environments at the tactical edge.
Smack Technologies is betting that the biggest opportunity in defense AI sits one level above autonomous platforms: the decision-making layer that coordinates people, sensors and machines across a contested battlefield.
The Austin-based defense technology company announced a $61 million Series B round led by Costanoa Ventures and First In, bringing its total funding to more than $90 million. Point72 Ventures, Geodesic Capital, Nomi Capital, Felicis, Sapphire Ventures, Scribble Ventures, Fortitude Ventures, Bloomberg Beta and Palumni VC also participated.
The new capital will fund development of Alpha, Smack’s tactical AI platform, including proprietary hardware intended to bring its decision-making models directly to the edge. The company also plans to expand its models across military domains and warfighting functions and hire additional AI researchers and engineers.
Smack describes its broader approach as “Decision Dominance” and “Intelligent Autonomy.” The distinction is important. Rather than focusing only on whether a drone can navigate autonomously or whether one operator can control multiple vehicles, Smack is attempting to apply AI to decisions across different levels of military operations.
Its existing Omega platform operates at the command level, while Alpha is designed for frontline units and autonomous systems. Smack says Alpha will combine lightweight models with proprietary hardware so that AI reasoning can continue when communications are degraded or computing resources are limited.
That architecture reflects a growing concern in military AI: cloud-connected systems can be powerful, but a battlefield is not a reliable cloud environment.
Communications can be disrupted. Bandwidth can disappear. Units can become isolated. Sensors can generate enormous amounts of data while commanders have limited time to interpret it. An AI system that requires continuous access to centralized computing may therefore be less useful precisely when operational conditions become most difficult.
Smack’s approach is to distribute some of the reasoning capability.
The company says its models are trained using deep reinforcement learning in proprietary synthetic environments intended to represent military tactics, physics, resource constraints and operational outcomes. Its product materials describe Omega as a command-level AI stack and Alpha as an edge-level system built around lightweight models and proprietary hardware.
That puts Smack in a growing defense-AI market that includes substantially larger companies.
Palantir, for example, markets AIP for Defense as an AI platform that can operate across classified networks and tactical-edge environments. Its TITAN program combines sensors, networks and AI-enabled automation to deliver decision-quality information to soldiers.
Anduril has built its defense technology strategy around autonomy, AI, sensor fusion and networking, with a focus on deploying capabilities quickly to military users. NASA recently highlighted Anduril’s approach to integrating AI, autonomy, computer vision and networking into defense systems.
The distinction Smack is attempting to establish is the model itself.
The company argues that commercial large language models are not sufficient for battlefield decision-making because military operations involve physics, resource constraints, adversarial behavior and mission-specific objectives that are poorly represented by conventional language-model training.
That does not mean general-purpose models disappear from defense AI. Smack’s own architecture uses commercially available LLMs at its human-machine interface, according to company materials, while its proprietary reasoning models handle more specialized decision-making.
This hybrid approach could become increasingly relevant as defense organizations try to balance commercial AI progress with mission-specific requirements.
McKinsey estimates that the U.S. Department of War has more than 600 AI efforts spanning areas from administration and engineering to operational applications, but argues that the department has yet to realize AI’s full potential at scale. The consulting firm identifies the transition from individual pilots to broader organizational transformation as a central challenge.
For military technology teams, that creates a problem familiar to enterprise AI buyers: the model is only one component.
A useful defense AI system also needs data pipelines, interfaces, deployment infrastructure, security controls and the ability to operate within the user’s existing command-and-control environment. In Smack’s case, moving Alpha to proprietary hardware adds another layer: the company will need to demonstrate that its edge systems can deliver useful reasoning under stringent size, weight, power and connectivity constraints.
The funding comes as Smack expands its government relationships. Reuters reported that the company secured two seven-figure prototype contracts in July with the Joint Fires Network and Marine Corps Warfighting Laboratory, while its workforce had grown from 19 employees to 51 since April.
Those contracts matter because defense AI companies ultimately face a different commercialization cycle from conventional enterprise software. A successful prototype must move through testing, procurement and operational deployment before it becomes a meaningful production business.
Smack’s strategy also reflects a broader shift in how military organizations are thinking about autonomy.
NATO is developing AI-enabled systems that can coordinate drones, sensors and other capabilities while keeping humans responsible for high-consequence targeting decisions. Recent reporting on NATO’s Eastern Flank Deterrence Initiative shows how the alliance is exploring AI-assisted autonomous systems while retaining human control over lethal decisions.
That human-machine boundary will remain one of the defining issues for defense AI.
Smack’s proposition is not that humans disappear from the decision process. Instead, the company argues that AI should absorb computationally intensive analysis and routine coordination, allowing human operators and commanders to concentrate on decisions requiring judgment, ethics and strategic responsibility.
The technical challenge is considerable.
AI models deployed at the tactical edge must contend with limited compute, intermittent connectivity, uncertain sensor inputs and rapidly changing conditions. They also need to produce outputs that military users can trust and understand.
Smack is now using its new capital to test whether its architecture can solve those problems at scale.
If it succeeds, the company’s competitive proposition will extend beyond another defense chatbot or autonomous vehicle. It will be an attempt to build an AI reasoning layer connecting command-level planning with decisions made by people and machines at the edge.
That is a much larger ambition—and one that could help define the next generation of defense AI infrastructure.
Market Landscape
Defense AI is evolving across several overlapping layers:
- Command-and-control AI: Systems that analyze information and assist with planning and operational decisions.
- Tactical-edge AI: Lightweight models capable of operating without continuous cloud connectivity.
- Autonomous systems: Drones, vehicles and other platforms capable of sensing, navigating or acting with reduced human intervention.
- Sensor fusion: AI systems combining information from multiple sources into operationally useful data.
- AI infrastructure: Hardware, networking and software required to run models in classified, disconnected and resource-constrained environments.
The competitive landscape includes Palantir, Anduril and established defense contractors, as well as a growing group of specialized AI startups.
Smack’s differentiation is its emphasis on domain-specific reasoning models trained through reinforcement learning and synthetic warfare environments, combined with a two-level architecture spanning Omega at command level and Alpha at the edge.
The larger market challenge is moving from demonstrations to dependable operational systems. McKinsey’s 2026 analysis argues that U.S. defense has accumulated hundreds of AI initiatives but still needs to translate experimentation into scaled impact.
Top Insights
- Smack Technologies raised $61 million to develop Alpha, extending its defense AI from command-level planning into communications-constrained tactical environments.
- Its Omega and Alpha platforms are designed to connect campaign-level reasoning with frontline units, autonomous systems and distributed decision-making.
- Smack uses domain-specific models and deep reinforcement learning rather than relying exclusively on commercial large language models for military reasoning.
- Palantir and Anduril demonstrate the competitive pressure, with both companies developing AI, autonomy and edge technologies for defense organizations.
- The funding highlights a broader defense AI shift from platform-specific autonomy toward coordinated human-machine decision-making across military echelons.
Power Tomorrow’s Intelligence — Build It with TechEdgeAI
