Defense tech startups raised $30 billion in 2025 — and Silicon Valley's ethics debate hasn't slowed them
In February 2022, a commercial satellite company streamed live footage of Russian tanks massing on Ukraine's border. Starlink terminals kept Ukrainian communications running after ground infrastructure was destroyed. Commercial drones designed for agriculture and filmmaking were repurposed for reconnaissance and strike missions. The war made visible what defense analysts had argued for years: the technology gap between consumer-grade and military-grade hardware had collapsed. Silicon Valley noticed.
US defense tech startups raised approximately $30 billion in venture capital in 2025, up from around $3 billion in 2019, according to data from PitchBook and Defense News. The sector has become one of the most actively funded categories in venture, sitting alongside AI, climate tech, and biotech. The growth is concentrated in a handful of large rounds and a long tail of seed and Series A investments in autonomous systems, software-defined radio, AI-driven logistics, and electronic warfare.
The companies leading the category
Anduril Industries is the clearest example of the category's ambitions. Founded in 2017 by Palmer Luckey — best known as the creator of the Oculus Rift — Anduril has raised over $4.5 billion across multiple rounds and is now valued at over $28 billion. Its flagship product is Lattice, a software platform that fuses sensor data from drones, ground sensors, cameras, and satellites into a common operational picture. Lattice powers Anduril's portfolio of autonomous systems: the Ghost and Ghost-X drones, the Fury autonomous combat aircraft, the Roadrunner counter-drone missile, and the Pulsar electronic warfare system.
Anduril's most significant contract is with the Australian government under the AUKUS agreement: a $4 billion deal to develop autonomous undersea vehicles for the Australian Defence Force. This is the kind of long-duration, large-value government contract that justifies the company's valuation — defense procurement moves slowly, but the contracts are enormous and relationships are sticky.
Shield AI, founded by former Navy SEAL Brandon Tseng and his brother Ryan, has built Hivemind — an autonomous AI pilot that can fly fighter jets and drones without GPS or communications. Hivemind has been demonstrated on the F-16 and the V-BAT drone. Shield AI raised $500 million in 2023 at a $2.7 billion valuation. Joby Aviation, Archer Aviation, and Wisk are nominally commercial air mobility companies, but their autonomy and manufacturing technology is explicitly dual-use. Skydio, whose autonomous drones have been widely adopted by US law enforcement, was blocked from selling to Chinese entities in 2022 under export controls and pivoted more aggressively toward defense.
Where the venture capital is coming from
Andreessen Horowitz launched its American Dynamism fund in 2022 explicitly targeting defense, aerospace, and manufacturing — sectors the firm had historically avoided. The fund's thesis is that national security and commercial success are aligned for the US in this moment. Founders Fund, General Catalyst, Peter Thiel's Palantir (itself a public defense tech company), 8VC, and several sovereign wealth funds from allied nations have all made significant commitments to the category.
The Department of Defense's Defense Innovation Unit (DIU), established in 2015 and significantly expanded under successive administrations, acts as a bridge between startups and military procurement. DIU can issue Other Transaction Agreements — contracts that bypass traditional Federal Acquisition Regulations and allow faster, more flexible procurement. This has made it easier for startups to get initial government revenue without waiting years for a traditional DoD contract.
The ethics debate that hasn't gone away
In 2018, Google employees staged an internal rebellion against Project Maven — a Pentagon contract to apply Google's AI to analyzing drone footage. Google eventually declined to renew the contract and published AI ethics principles that excluded weapons applications. The episode became a defining moment in tech industry discourse about the ethics of defense work.
Seven years later, the consensus has fractured. Some founders and investors remain categorically opposed to defense work: refusing to build autonomous weapons, refusing to take DoD contracts, and explicitly marketing that position to recruit employees who share it. Others have moved in the opposite direction, arguing that the alternative to Silicon Valley building defense tech is that authoritarian states build it instead. "Democracies need sharp swords," as Palmer Luckey has put it.
Anduril has leaned into this position. The company explicitly recruits on the premise that working there is patriotic and morally defensible. Palantir's CEO Alex Karp has made similar arguments publicly. The counterargument from critics focuses on the specific technologies involved — autonomous targeting systems, surveillance infrastructure, tools for border enforcement — rather than the abstract principle of defense spending.
The internal tech industry debate has largely moved from "should companies work in defense?" to "which specific applications are acceptable?" Few people object to logistics software or satellite communications; more people object to autonomous kill decisions, predictive policing, and mass surveillance systems. The line between these applications is contested, and the companies building the platforms generally argue that the end-use decisions belong to the military and government customers, not to the technology vendor.
The financial reality of defense tech
Defense contracts are slow to close and slow to scale. A startup that wins an initial DIU contract for $1 million in year one may spend three years proving the technology before getting a program of record worth $50 million, and another two years before full-rate production at $500 million. The timeline from founding to significant government revenue is typically seven to ten years. This is incompatible with traditional venture capital timelines of three to five years to exit.
The category has adapted. Defense tech investors are increasingly structured as patient capital: 10-15 year fund lifespans rather than seven. Many investors are explicitly betting on IPOs rather than acquisitions — Palantir and Joby are public comps, and several private defense tech companies are expected to go public in the next two to three years.
The total addressable market argument is real: the US defense budget exceeds $900 billion annually, and allied defense spending adds trillions more. Even a fraction of a percent of procurement shifting toward startup vendors represents an enormous market. Whether that shift happens at the pace venture capital requires is the bet investors are making.