Watchtower Labs Raises $35M for Defense AI

Bellevue defense startup Watchtower Labs launched from stealth today with $35 million in seed funding. The firm builds AI software for defense supply planning. It helps operational teams make faster, data-backed choices under pressure. Shield Capital and Conviction co-led the seed round. Founders Fund and In-Q-Tel joined the investment as well.

Watchtower began inside Shield Capital. CEO Akhil Iyer co-founded the firm. He served as a Marine Corps officer in infantry and special operations. After leaving the service, Iyer built counter-drone tools. He also led defense startup investments at Shield.

CTO Souren Aghajanyan and Head of Engineering John Morman joined as co-founders. Both led engineering teams at Microsoft. They previously built large distributed systems and fraud detection platforms under Lou Manousos at RiskIQ.

Technical Factsheet: Watchtower Labs

Why Pure LLMs Fail at Supply Chain Logistics

Most discussions around AI focus on search or text extraction. In defense supply chains, pure language models fall short. Moving fuel, munitions, and spare parts is a math problem. It has strict physical limits and transit constraints. It is not a text prediction task.

“You can’t LLM your way to a really complicated math problem, yet at least. It needs to be set up in the right way, and you need to do some real analysis behind it.”

— Akhil Iyer, CEO of Watchtower Labs

Watchtower combines two distinct software layers to solve this gap:

1. Agentic Natural Language UI: Planners ask operational questions in plain English. For example, a planner can ask: “If port capacity drops by 40% over two weeks, what fuel reserves remain in Sector B?”
2. Deterministic Simulation Engine: The software converts the query into clear math constraints. It tests hundreds of timelines against available planes, ships, and trucks. The model maps out bottlenecks and shows the trade-offs behind each path.

Most defense tools show where gear sits right now. Watchtower looks ahead. It builds models that forecast delivery dates and show backup routes when supply lines break. Crucially, human commanders still make the final call.

Active Deployments: The DIU Joint Sustainment Program

Watchtower already holds live defense contracts. In January, the firm won a prototype deal through the Defense Innovation Unit (DIU). The project is the Joint Sustainment Decision Tool (JSDT) program. Watchtower won the deal with Air Space Intelligence.

The DIU program serves three major defense partners:

• U.S. Indo-Pacific Command (INDOPACOM): Lead operational partner, dealing with wide sea routes across the Pacific.
• U.S. Northern Command (NORTHCOM): Handles homeland defense readiness.
• Defense Logistics Agency (DLA): Coordinates military supply depots across the globe.

These tools track use rates and forecast shortfalls weeks ahead. Key supplies include guided missiles, jet fuel, repair parts, and medical goods.

Pacific Northwest Defense Tech Corridor Growth

Watchtower’s launch reflects rapid tech growth in the Puget Sound region. Seattle has a deep pool of cloud and distributed systems talent. Several defense companies now run large hubs in the area:

• Overland AI: A University of Washington spinout building autonomous ground vehicles. It raised $100M this year. It also won a $20M Marine Corps contract for its Ibex vehicle.
• Shield AI: Building its X-BAT autonomous jet plant in Des Moines, Wash. It plans for 4,000 regional staff.
• Anduril Industries: Employs over 560 engineers in Seattle and Bellevue, with goals to reach 1,000.
• Echodyne: Operates a 40,000-square-foot radar plant in Woodinville.
• Cloud Providers: Microsoft and Amazon secured Pentagon deals in May to run AI models on classified networks.

Watchtower employs 30 people across Bellevue, San Francisco, Hawaii, and Washington, D.C. The company is moving to a larger Bellevue office. It plans to double its team, hiring systems engineers, simulation specialists, and machine learning developers. Roles require U.S. citizenship.

Architecture Lessons for Enterprise Engineers

For software teams, Watchtower shows an effective pattern for applied AI:

• Use language models for interface translation, not raw calculation.
• Keep business logic, inventory rules, and route solvers strictly deterministic.
• Expose assumptions and trade-offs so human operators stay in control.

Managing configuration files or API schemas for distributed services? Format and check your code with TeckScaler’s free client-side JSON to YAML Converter.