Booming Military Spending on AI is a Windfall for Tech—and a Blow to Democracy

The imperative to win a global AI race—and an AI-powered war—is also supercharging the tech industry’s influence over governments, markets and people, write Brian J. Chen, Tina M. Park, and Alex Pasternack. This piece is part of “Ideologies of Control: A Series on Tech Power and Democratic Crisis,” in collaboration with Data & Society. Read more about the series here.

“We rebuilt the military during my first term and we have great things happening with our military. We also essentially approved a budget which is in the [vicinity], you'll like to hear this, of $1 trillion. One trillion, and nobody’s seen anything like it.” —US President Donald Trump, April 7, 2025

Speaking to the press alongside Israeli Prime Minister Benjamin Netanyahu, President Donald Trump announced his intent to propose a $1 trillion budget for the Department of Defense (DoD). About a month later, the president’s budget blueprint to Congress made it official. This stunning figure, if approved, would make it the largest defense budget in US history. Leading AI companies are well-positioned to reap the historic rewards.

Like the Trump administration’s “AI-first strategy” to push AI tools into every federal agency, senior Defense Department officials are focused on dramatically increasing AI’s role amid rising competition with China. But there is an important difference in these approaches. For non-defense purposes, the administration views AI as an opportunity to impose austerity, firing federal workers and eliminating “waste” in social spending. By contrast, the Pentagon’s pursuit of AI is not an exercise in fiscal discipline: now, as ever, cash seems to flow endlessly for national security spending.

Traditionally, increases to the US defense budget have enjoyed bipartisan support in Congress. Lawmakers have worked in recent years with both the first Trump and Biden administrations to authorize massive increases to the Pentagon budget. The majority of allocated funds are disbursed through private contractors. In fiscal year 2022 alone, the DoD incurred a total of $415 billion in contract obligations, roughly half of its total budget authority. For decades, these funds were doled out to a small constellation of prime contractors who provided everything from researching and making weapons and other systems and materials required by the DoD, to providing logistical support like supply chain management and transportation, to providing data analysis and expertise.

Technology companies have long held close relationships with the US military and the DoD. Companies such as IBM, Microsoft, and Oracle are de facto defense contractors, providing the Pentagon with advanced data storage, processing, and computing capabilities. Recently, defense tech startups like Palantir and ScaleAI have been competing against legacy contractors such as Raytheon and Lockheed Martin. Dominant consumer technology companies like Google and Meta have generally maintained more distant relationships with the Pentagon, with internal policies banning the use of their AI models for military use or surveillance purposes.

Their hesitancy changed after Trump’s return to office, as several companies walked back the restrictions. Within six weeks of Trump’s election day victory, major tech companies, including OpenAI, Anthropic, and Amazon, had announced partnerships with defense tech companies like Palantir and Anduril to jointly bid for defense contracts. The products of those early joint ventures are beginning to surface. Last month, Meta and Anduril announced a collaboration to build technology for US soldiers, starting with an augmented reality headset first prototyped by Microsoft.

The “tides have turned” in tech towards the defense business, Meta’s chief technology officer Andrew Bosworth said at a recent conference. Cold War military spending helped create Silicon Valley, and “there’s really a long history here that we are kind of hoping to return to,” he said, “but it is not even day one.”

Today, advances in AI technology are playing a major role in national security and warfare. Maven, a Pentagon project conducted in collaboration with companies including Google, Palantir, Amazon, and Microsoft, uses machine learning and data fusion to analyze information, recognize objects, and identify targets at rapid speed. The Israeli military’s AI-based Lavender program identified targets for assassination in Gaza, reportedly without checking if the nearly 37,000 Palestinians identified as suspected militants were, in fact, militants. The Ukrainian military’s recent attack against a Russian outpost using a barrage of AI-enabled ground and aerial drones marks a new chapter in battlefield warfare: robots fighting robots. The drones were remotely controlled, but the next generation of advanced weaponry can carry out attacks autonomously.

The increasing role of consumer-oriented technology companies in warfare has implications for everyone—not just those who are targeted by AI weapons. As “dual-use technologies,” AI tools have both civilian and military applications, with data collection and product testing in one domain being leveraged to make improvements in the other. At the heart of several of the latest tech consortia being sold to the DoD are commercial foundation models, like ChatGPT (OpenAI), Claude (Anthropic), and Llama (Meta). Originally developed for commercial, civilian purposes using public and private data, they are now being repurposed for military use.

Civilians who use commercial digital tools—just about all of us—do not always know that they are contributing to the improvement and proliferation of military technology. They’re also largely unaware that the technology is repackaged for use against civilians. For example, in 2012, Google experimented with showing images of street names and traffic signs from Street View imagery in CAPTCHAs, the common “are you human” test to detect bots, in order to build better AI training datasets with accurate labels for images. These labeled datasets are integral to the advancement of computer vision technology, or AI that accurately identifies objects in images.

As a technology company improves its computer vision algorithms, it can apply these algorithms to military AI-based surveillance tools. Once tested in military operations, these tools are often sold back to domestic law enforcement agencies and other customers seeking automated surveillance capabilities, with minimal transparency and accountability. As their businesses become further entwined with national security, tech companies are better able to shield themselves from oversight.

The grand strategy of Trump’s foreign policy remains difficult to discern. What is clear is his administration’s reliance on advanced technology to backstop American global power. As tech companies get in on the booming defense action, the ramifications will be felt far beyond the battlefield. Because US national security relies on a marketplace of private vendors, the ones who win large government contracts wield outsized geopolitical influence. As AI companies make themselves essential to national defense, they will be able to maneuver a vast security apparatus, repurposing public power to meet their bottom lines. At a historical moment of global unease, new heights of military spending stand to be the latest and most extravagant source of Silicon Valley’s power, driving risks to open societies and individual freedom.