Tech v. Tech? (2026) | K illustrator Photo / Shutterstock Editorial License
Military and surveillance tech giant Palantir recently published an unsettling and dystopian manifesto penned by billionaire CEO Alex Karp. Karp’s 22-point document extolls American hegemony and claims that democracies need “hard power” to survive; it also treats one thing as a foregone conclusion—the militarization of artificial intelligence. “Our adversaries will not pause to indulge in theatrical debates about the merits of developing technologies with critical military and national security applications,” the manifesto pronounces. “They will proceed.”
Palantir has untold billions to gain from the ongoing and escalating entanglement of military, surveillance, and tech industries. It is also far from the only leading tech firm supporting the use of frontier technology for weapons and war. Starting in early 2024, companies including OpenAI, Meta, Anthropic, and Google have all shifted their official company policies to normalize cooperation with the US military.
This fusion of big tech with state interests marks a notable departure from the neoliberal free market ideals which previously dominated Silicon Valley. For Nick Srnicek, in his book Silicon Empires: The Fight for the Future of AI (Polity, 2026), explaining why this shift takes place necessitates thinking across several registers—from the technical to the sectoral to the geopolitical. Srnicek’s goal is to illustrate this ambitious framework while staying grounded in dynamics of profit and power enabled by state-capital alliances.
Silicon Empires opens by laying out the economic and technical dynamics of the AI sector, and Srnicek makes clear the concrete strategies and logics of “corporate capture” pursued by different firms within the AI ecosystem. For example, some firms, such as Amazon, pursue an “infrastructure strategy,” investing in the physical underpinnings of AI and other digital activity, extracting cloud rents from as wide a user base as possible. Higher up the AI stack, companies like OpenAI pursue a “frontier strategy,” developing the most advanced general purpose AI models to pull in a mass user base, while companies like Google follow a “conglomerate strategy” of building and providing AI tools adapted to specific industries.
But Silicon Empires is most ambitious and compelling when it expands this mapping of the sector into a broader political-economic frame. In the US, Srnicek depicts an emergent form of techno-nationalism, departing from the previous neoliberal era “Silicon Valley consensus” of free trade and free-flowing data. Now, Srnicek points out, the American “silicon empire” is distinctly right wing, Sinophobic, and committed to a post-neoliberal state interventionism.
This is, in part, a response to tech firms’ shifting infrastructural needs: AI buildout requires extensive access to land, water, critical minerals, and energy for data centers. This is only obtained with the help of a state willing to secure global supply chains, remove regulations, shower corporations with tax breaks, and override local and other political sites of opposition; Donald Trump’s administration has proven a willing partner.
Srnicek turns, too, to the story of AI in China, noting its similarity to the US model of state and capital’s alliance but highlighting China’s distinct developmental strategy. The Chinese government, for example, successfully realigned tech giants with state goals, as well as reconfigured the hierarchy of firms in the domestic ecosystem.
This was most visible beginning in late 2020, when the Chinese state targeted the tech sector with anti-trust fines, denied mergers, and blocked the record-setting $34 billion IPO of Ant Group. This crackdown signaled not just regulatory power over the tech sector but the Chinese government’s distinct willingness to undermine financial accumulation to control what they call the “disorderly expansion of capital”—a clear dig at the bubble-prone dynamics of US financial markets.
Srnicek explains that the key distinction of the Chinese AI trajectory is its focus on diffusion into different industries, rather than a pursuit of frontier breakthroughs. While this could be viewed as a political choice to favor a particular form of capital accumulation, Srnicek emphasizes the material constraints China faces—such as in compute capacity or in access to the advanced chips necessary for frontier development—thus framing its focus on AI diffusion as a product of necessity rather than ideology.
Srnicek aptly threads the needle between ideological and material explanations for the shape of current techno-development. The author also helpfully elucidates policy choices that seem incoherent. One such example is the contradictory US policy around export controls of advanced chips to China. The second Trump administration reversed course on its own ban on Nvidia’s H20 chip exports to China, as well as ending a major Biden-era policy that withheld the more advanced H200 chip. This represents a drastic pivot away from the tactic of strangling Chinese AI development toward leveraging Nvidia’s dominance to both promote and monetize reliance on US hardware.
Rather than tying this incoherence to that of Trump’s erratic character, Srnicek places the policy in the context of power contestation within the US tech sector. Companies focused on expanding US-controlled infrastructure across the globe—like the chipmaking giant Nvidia—advocate for an environment of more open trade. Meanwhile, startups like OpenAI have increasingly relied for their profitability on a nationalist vision of technological development, and as a result want to restrict foreign access to that same infrastructure. The ambiguous trade policy towards China is a result of these competing elite interests.
Yet whatever their differences when it comes to trade positions, each tech faction needs to win over the power of the state. Thus, for Srnicek, the way these elite alliances play out will determine the form the “silicon empire” takes.
Srnicek points to other arenas besides trade policy in which the fissures within the US silicon empire manifest. One is the project of securing the physical resources, like energy and critical minerals, needed for the AI build out. Here we can see the same factional division between infrastructural companies, like Nvidia and Amazon, and the startup AI leaders, like OpenAI. While both factions support state intervention to secure resources, the former advocates a diplomatic approach that refrains from stoking geopolitical tensions and closing markets, while parts of the latter are either indifferent to or actively supportive of Trump’s bellicose approach to international relations.
This context allows us to interpret the sort of rhetoric found in Karp’s militant manifesto. Karp’s perspective is neither anomalous nor universal among tech capitalists. Palantir clearly has direct commercial interest in military conflagration—but Srnicek’s analysis allows us to see that it is also part of a broader startup faction of tech capital who are driven by their material interests to support nationalistic confrontation on the global stage.
Srnicek’s framework can unearth the material mechanisms behind dangerous narratives like Karp’s while also presenting a path towards unraveling them. His book provides the tools to see the US and China not only as two technological empires in a winner-take-all battle but as manifestations of tangible, and mutually distinct, political-economic forces.
Silicon Empires does not claim to chart any grand narrative of AI development. Rather, it seems to do quite the opposite, suggesting that what may seem like a consolidation of new oligarchy is really a more fragile and transient compact between tech capital and the state. From this perspective, the political question around who controls the future of AI is not settled, and our current moment resembles more of an interregnum than an endpoint.
