2
0
A security dilemma arises when one state’s efforts to increase its capabilities prompt rivals to do the same, thereby fueling an arms race. Traditionally, this was driven by government-funded research that led to proven technologies and procurement. Since the late 20th century, private-sector investment and wealth concentration have shifted the process, with venture capital now creating security dilemmas through speculation before real capabilities exist. Speculation adds new market-driven mechanisms to the traditional logic of reciprocal insecurity.
The 1960s McNamara doctrine forced defence companies to absorb cost overruns, making profits less certain. Budget cuts and financial crises exposed market risks, pushing companies to diversify into civilian sectors and focus on share prices. This financialization turned defence from a state-led enterprise into a competitive market, where investor confidence is parallel to strategic goals.
Venture-Capital and Speculative Investment
Financialization kept the old pattern: funding led to capability, then procurement. Now, VC-driven defence markets raise private capital for unproven, speculative weapons, not just proven ones.
Shield-AI recently unveiled the X-BAT fighter jet, powered by their proprietary AI “Hivemind”. The plane, which has not flown, can theoretically deploy vertically without a runway and land in small or contested areas where runways are impractical. Such deployment would be a significant advance in military technology. However, the X-BAT has been in development for only 18 months and will not be deployable before 2028, according to Shield-AI. This deployment date is only a prediction. The company has actually created capabilities before, and followed a similar structure to traditional procurement dynamics. In 2024, the U.S. Coast Guard awarded a contract to Shield AI valued at approximately $200 million USD for its unmanned aircraft systems (UAS), which are similar to drones but can be vertically deployed. The contract, which runs until 2029, does not have a specified quantity or delivery time. The UAS was only operationally tested this July, and actually deployed this August.
Palantir’s valuation has soared on AI-defence narratives, winning up to $10 billion in contracts, including $480 million for a prototype. These deals hinge on anticipated, not proven, AI capabilities—much of Palantir’s value remains speculative.
Companies hype capabilities to attract investment and contracts, resulting in speculative deals. In the speculation-driven security dilemma, rivals react to announcements—real or not—by funding their own programs before actual capabilities exist.
Speculation isn’t new to security dilemmas, but today’s version expands uncertainty: states now fear not just rivals’ current capabilities, but what might emerge soon. Without tested frameworks for novel tech like AI, states invest quickly out of insecurity.
The Pentagon’s “Replicator Initiative” illustrates how speculation about future technologies is driving state behaviour. The Replicator Initiative-1, now in its second phase, was an effort by the U.S. government to speed up the development of autonomous weapons systems. The goal was to deliver thousands of autonomous weapons – mainly autonomous drones – within an 18- to 24-month period. The initiative explicitly emphasizes the role of private-sector tech companies not only as sources of innovation but also as capable of scaling. Of the over 800 companies involved in Replicator-1, 75% of them are “non-traditional defence contractors.” The initiative is a response to China’s rapid expansion of its armed forces, amid uncertainty and anxiety about what technologies may emerge in the future. Unlike traditional arms races driven by observable deployments, investment in AI technologies is driven by projected capability gaps.
A Fear of Falling Behind
NATO’s “Emerging Disruptive Technologies” similarly stresses speed and private investment, with a $1 billion venture fund to keep up in the AI arms race.
Fear of falling behind drives rapid, less-verified tech acquisitions—many see acting on hype as safer than being left vulnerable.
For tech companies hoping to court some of the massive resources states allocate, exaggeration has few downsides. In a security dilemma dynamic, the cost of a failed investment is low. As states continue to invest massively in emerging technologies, the hype builds, attracting further investment. In a speculative-security dilemma, hype becomes a political asset that legitimizes large expenditures with limited oversight. Militaries are adopting a Silicon Valley ethos: better an imperfect technology now than a perfect one too late. Nevertheless, the speculation-driven security dilemma is not just about market forces, it is also fueled by rumours in Western defence circles that war with China is inevitable. A Chinese invasion of Taiwan in 2027 has been widely speculated about. U.S. military intelligence claims Xi Jinping told the military to prepare by 2027, with invasion expected by 2027 or 2035. At a 2024 conference, Xi denied this timeline. Still, belief in looming war remains strong. The U.S. is sending more resources to the Pacific and arms to Taiwan, including $15 billion to the Pacific-Indo Command. Belief in an imminent war hastens investment in new technologies, driven by the fear that delay will leave the country vulnerable. Chinese tech developments, like unmanned drones shown at the 2025 military parade, further justify this speculative investment. States and militaries are treating speculative projections as if they were known threats.
The risks of this security dilemma are similar to those of traditional ones. In both dilemmas states remain unaware of their rivals’ capabilities and are inherently concerned about their security. This insecurity leads to an arms race and an increased likelihood of conflict, even if neither side wants war. However, the mechanisms for risk assessment differ. Instead of focusing on real capabilities, the speculation-driven dilemma centers on perceptions of future capabilities that may or may not materialize. Because capabilities are imagined rather than observed, escalation is based on hype. This fuels over-investment, rushed procurement, and the deployment of untested systems. This carries real risks of harm, especially as many companies – such as Palantir – propose using their systems for target identification and when decision-makers act on uncertain knowledge and treat speculation as intelligence. Valuation and hype begin to bleed into strategic assessment. Together, these forces entrench a feedback loop between markets and militaries. High valuations further incentivize state investment: hype becomes a driver of national security. By treating market speculation as a proxy for adversary capability, states risk accelerating a cycle of escalation detached from material reality – while raising the likelihood of an armed conflict unlike any before it.
Edited by Jude Archer
The argument defended in this article is solely that of the author and does not reflect the position of the McGill Journal of Political Science, the Political Science Students’ Association, or the McGill Department of Political Science.
Featured image by James Smith, obtained through Wikimedia Commons.
