Excerpt adapted from Silicon Empires: The Fight for the Future of AI by Nick Srnicek. Published by arrangement with Polity Books. Copyright © 2026 Nick Srnicek.

---

At the opening of 2024, the world’s four leading AI developers—Anthropic, Google, Meta, and OpenAI—were united in ruling out military use of their cutting-edge AI tools. Within 12 months, that shared commitment unraveled completely.

OpenAI first quietly walked back its blanket ban on “military and warfare” AI applications in January, and reports soon emerged confirming the firm was already working on “a number of projects” alongside the U.S. Pentagon. That November, the same week Donald Trump won re-election as U.S. president, Meta announced the U.S. and its key allied nations would be cleared to use its Llama model for defense work. A few days later, Anthropic followed suit, confirming it would permit military access to its AI systems and had entered a formal partnership with major defense contractor Palantir. Before 2024 closed, OpenAI announced its own collaboration with defense startup Anduril. Finally, in February 2025, Google revised its public AI ethics principles to clear the way for developing and deploying weapons and other technologies that can cause harm to people. In just 12 months, widespread anxiety over the existential risks of artificial general intelligence all but vanished, and military integration of AI had been fully normalized.

One driver of this shift is the staggering cost of building and scaling modern large AI models. Research on general-purpose technologies (the other meaning of the acronym GPT) has long highlighted the defense sector’s unique role in solving early adoption barriers for transformative new technology. “GPTs develop faster when there’s a large, demanding, income-generating application sector,” economist David J. Teece wrote in 2018, pointing to “the US Defense Department’s purchases of early transistors and microprocessors” as a defining example. Flexible defense budgets, long-term contracting commitments, and loosely defined success metrics make the military an unusually desirable client for unproven new technologies. For cash-strapped AI startups in particular, which need large, patient capital infusions to keep advancing, a turn toward military funding was perhaps always inevitable. But this dynamic alone cannot explain how fast the shift occurred, nor why every top U.S. AI lab moved in lockstep to reverse their earlier positions.

The past few years have dramatically reshaped the landscape of global capitalist competition: what was once a system guided by neoliberal free market ideals has become saturated by great power geopolitical rivalry. To understand this shift, we have to examine the evolving relationship between states and their largest technology firms. These state-capital alliances have long been core to modern imperialism—Vladimir Lenin famously defined the imperialism of his era as a merger between monopoly capital and great power states—and this dynamic remained influential across the 20th century. In recent decades, it took the shape of a broad cross-elite consensus over digital technology’s role in driving innovation, economic growth, and national power.

Over recent years, however, this shared harmony of elite interests has fallen apart. A series of overlapping shifts, which gained particular momentum in the 2010s, dismantled the old order, leaving behind fragments of potential new arrangements in both the United States and China.

The Silicon Valley Consensus

Up until roughly the mid-2010s, U.S. politics and industry operated under what we can call the Silicon Valley Consensus: a broad agreement shared by both political and tech elites on technology’s global role, the conditions needed for tech to flourish, the American values it purportedly embodied, and the requirements for capital accumulation in the tech sector. For both groups, the global free flow of communication, capital, data, and technology advanced their overlapping interests.

The Silicon Valley Consensus appealed to both elites because it centered on technology’s ability to build an American-led world of borderless commerce and open data exchange. While the tech sector initially held more utopian impulses than the state’s hard-headed geopolitical realism, both sides saw their separate goals advanced by the same policy framework.

In practice, this meant the tech sector was given near-total autonomy to operate, with regulation either glaringly absent or intentionally structured to facilitate corporate expansion. Deregulation was a core feature of the broader neoliberal era, but it was especially pronounced for tech firms, which regularly blurred existing regulatory categories and “disrupted” long-standing rules. The absence of meaningful federal privacy legislation, and the lack of action to clarify worker protections for gig economy employees, are clear examples of policymakers’ willingness to let digital firms set their own terms. Under President Bill Clinton, the Framework for Global Electronic Commerce laid out policy priorities that, as international studies scholar Henry Farrell put it, succeeded in “discouraging policymakers from seeking to tax or regulate” the digital economy, instead leaning on voluntary, industry-led self-regulation. The core belief here—that any government regulation would simply hinder innovation and the expansion of U.S. technological power—remains widely held today.

What regulation did pass was almost entirely designed to enable digital firm growth. Section 230 of the 1996 Communications Decency Act, for example, codified limited legal liability for tech firms for content posted on their platforms—a stark departure from the accountability rules applied to traditional publishers. The result is that “courts applying this standard have dismissed a variety of claims against service providers including defamation, negligence, housing discrimination, and cyberstalking claims.” Section 230 also gave tech firms near-total control over content moderation, letting them unilaterally decide what content is appropriate for their platforms.

In foreign policy, the Silicon Valley Consensus meant the U.S. state pressured other countries to abandon any policies that harmed U.S. digital firms’ profits. When France proposed a 3% tax on large platform giants, the U.S. threatened 100% retaliatory tariffs in response, even though France ultimately moved forward with the policy anyway.

Elites also worked to reshape international trade rules to align with tech firm priorities. Many recent trade deals—including the U.S.-Mexico-Canada Agreement and the U.S.-Japan Digital Trade Agreement—embed clauses that mirror Section 230’s liability protections. Using the same outdated 1990s legal language around “interactive computer services,” these clauses aim to spread the same legal protections U.S. digital firms enjoy at home to other jurisdictions, dramatically expanding their shield from accountability, litigation, and other legal challenges. While many countries resist these terms, during Trump’s first term these deals were hailed as the “gold standard” for all future digital trade agreements.

Crucially, the Silicon Valley Consensus around economic and digital globalization also extended to U.S. policy toward China during this period. Not every segment of American capital supported opening economic relations with China, and this created core tensions between different factions of U.S. capitalism through the 2000s. Domestic manufacturers feared competition from cheap Chinese goods, organized labor opposed job outsourcing to China, and the national security wing of the U.S. state remained wary of China’s growing geopolitical power. But when it came to setting U.S. China policy, these groups lost out to a coalition of financial capital, platform capital, and large manufacturers that viewed China as a source of cheap low-wage labor.

By the late 1990s, U.S. policy shifted sharply toward opening to China, with the explicit goal of integrating it into the global capitalist system. In 1998, Clinton became the first U.S. president to visit China after the 1989 Tiananmen Square crackdown—a clear signal that economic interests had overtaken human rights concerns as a policy priority. Soon after, the two countries signed a landmark trade deal and the U.S. supported China’s entry into the World Trade Organization.

Importantly, the segments of American capital that supported these moves also acted as a stabilizing buffer during geopolitical crises between the two powers. As political economist Ho-fung Hung writes in Clash of Empires, these factions helped facilitate “quick, reconciliatory resolutions to crises such as the Taiwan missile crisis in 1996, the US bombing of the Chinese embassy in Belgrade in 1999, and the clash between a US spy plane and a Chinese fighter jet over the South China Sea in 2001.” U.S. capital, in other words, played a key role in preserving peace and deepening economic integration between the two powers.

For leading U.S. tech companies, openness to China was a core strategic priority. Firms like Apple came to rely heavily on low-wage Chinese labor for manufacturing. As China grew into a middle-income economy, its massive consumer market became an even bigger draw, leading many companies to comply willingly with the Chinese government’s market access requirements.

Ties between U.S. platform capital and China even extended to personal connections, with many top tech elites actively courting favor with Chinese leadership. In a striking example, Mark Zuckerberg learned Mandarin, publicly stated he was reading Xi Jinping’s collection of speeches, and during a meeting with Xi reportedly asked the leader to give his unborn daughter an honorary Chinese name (Xi declined). Platform capital’s support for Chinese integration into the global order was a defining pillar of the Silicon Valley Consensus.

The consensus that held through the 2000s and most of the 2010s was therefore defined by a broad harmony of interests between platform capital and the U.S. state. Both sides agreed on the importance of globalization built on capitalist markets and enabled by digital technology, and both shared a commitment to minimal regulation paired with maximum innovation, which were seen as naturally reinforcing.

Throughout this period, economic interests were ascendant. National security concerns were sidelined, with most elites assuming growing economic integration would automatically resolve geopolitical frictions. Even the post-9/11 War on Terror, despite its bellicose rhetoric, was never treated as an existential threat that required disrupting global economic integration. The primacy of economic interests and the growing political power of the tech sector gave elites the driver’s seat, setting the direction for U.S. society and the global economy.

The Consensus Undone

Today, the relationship between big tech platforms and nation-states is unrecognizable. Geopolitical concerns now dominate every policymaker’s thinking, and increasingly shape the investment decisions of capitalists. A key driver of this transformation is the emergence of a new faction of Silicon Valley capital, which has split the American tech elite along competing ideological lines.

On one side, a powerful bloc of established Big Tech firms still benefits from globalized digital capitalism. These firms have historically been socially liberal and economically neoliberal—just as willing to embrace a sanitized, corporate form of identity politics as they are to use state power to advance their global expansion. They remain committed to a globalist vision today, not least because their power and business models depend on a neoliberal world order that allows free flow of goods, services, and data across borders. At the same time, even these Big Tech firms have shifted dramatically over the past decade: they increasingly rely on the state to defend their interests abroad, have become deeply integrated into the national security state, and have moved decisively to push back against the liberal and left-leaning views of their rank-and-file workers.

On the other side, a newly prominent tech right has grown rapidly in influence. Unlike the relatively unified, orthodox neoliberalism of established Big Tech, this grouping encompasses a range of competing radical ideologies—from libertarian network-state utopias to techno-monarchism to openly eugenicist worldviews. This faction has always existed in Silicon Valley, but it was largely sidelined for decades. While most Big Tech has historically been firmly aligned with the Democratic Party, this new faction has been openly pro-Trump from the start. Peter Thiel was the first major tech elite to back Trump’s 2016 campaign, and Anduril founder Palmer Luckey claims he was fired from Meta for donating to Trump (Meta denies this allegation). While the tech right has benefited from Trump’s political rise, its new prominence also reflects broader global shifts: the global rise of the far right, the growing crackdown on tech worker activism, and the broad post-neoliberal policy shift advanced by both the Trump and Biden administrations.

The growth of great power competition has also opened huge new streams of political attention, venture capital funding, and congressional budget allocations for this faction. Old neoliberal ideas of connecting the world through corporate infrastructure have been replaced by a worldview that frames great power conflict as irreconcilable. While the more neoliberal Big Tech bloc either profits heavily from global operations or relies heavily on exports, the new tech right often counts the U.S. government as its primary client, and pushes for security-focused reshoring of manufacturing.

The rise of this competing center of tech power means the era of consensus is over. Instead, a hegemonic struggle is now underway to define how American tech and the American state will work together. Two core points of contention shape this fight: what shared goals tech and political elites will pursue, and what vision of the international order this alliance will advance.

On the first question, the new tech right calls for a reshaping of the state-tech alliance: instead of centering on the liberal expansion of digital markets, the alliance should focus on securing the U.S. from geopolitical rivals. The most prominent examples of this new approach are emerging defense tech startups that aim to displace long-entrenched incumbent defense contractors. The stakes could not be higher: the Pentagon spent over $14 trillion in the first two decades of the 21st century, with roughly one-third of that spending going to just the five largest incumbent contractors.

Against these incumbents, the new military-industrial complex promises a far more agile network of startups that deliver faster, more adaptive innovation. Venture capitalists have broadly lined up behind this new techno-nationalist agenda. Given the industry’s high concentration, signals from leading firms like Andreessen Horowitz and General Catalyst have been particularly influential. New defense-focused investment funds are launching at a rapid clip, and a growing, vocal cohort of VCs now openly argues for a large-scale turn toward defense industrialization. There are already early signs of a cultural shift among rank-and-file tech workers as well: what was once a reliably liberal space now has a growing number of workers rejecting past protests against military contracts, dismissing those protests as symptoms of anti-American sentiment.

Even as established Big Tech retains its globalist vision, it has also increasingly moved to ingratiate itself more closely with national security agencies around the world. Amazon in particular has become de facto essential infrastructure for the U.S. military: it signed a cloud computing deal with the CIA and other U.S. intelligence agencies in 2013, a follow-on deal in 2020, a $10 billion contract with the NSA in 2021, and a new contract with the U.S. Army in 2024. But every major cloud provider has built specialized infrastructure for security agencies and even traditional defense giants.

Beyond the flow of contract money, there is also a growing flow of personnel between the two sectors, with a well-established revolving door between tech firms and the military creating ever-tighter integrated networks. Amazon, Microsoft, and Google have long participated in this practice, and in 2024 OpenAI joined the trend when it appointed former NSA head Paul M. Nakasone to its board of directors. And as global conflicts have multiplied, Big Tech firms have increasingly become active participants: from securing Ukrainian government data to providing infrastructure for Israel’s war in Gaza. The 21st-century national security state is increasingly built on infrastructure owned and operated by Big Tech, tying these digital giants into a new form of military-industrial complex.

This shift brings a second major change to the tech-state relationship: a transformation in how elites envision the global international order. Alongside growing support for techno-nationalism, national security efforts to stymy China’s rise have moved to the top of the U.S. policy agenda in recent years. Deploying a range of tools—tariffs, investment screening, export controls, and more—the U.S. has definitively broken with the neoliberal era of globalization and free trade.

Rhetoric about the threat of Chinese competition has also been weaponized by many tech companies to resist domestic regulation. And leading AI startups have increasingly pushed a narrative that frames U.S.-China competition as a zero-sum struggle. Sam Altman shifted from emphasizing the need for dialogue and engagement with China on AI, to arguing in The Washington Post that the future of AI should be controlled by a “US-led coalition of like-minded countries.”

Similarly, Anthropic CEO Dario Amodei shifted from warning about the dangers of a U.S.-China AI arms race, to insisting that the world is divided between democratic and authoritarian countries, and that democracies must win the AI race. By contrast, large hyperscaler AI firms and semiconductor companies have largely avoided stoking geopolitical tensions. Consistent with their global footprint and preference for a neoliberal world order, they routinely advocate for a global system built around the free flow of goods, services, and data.

All of these shifts add up to a major breakdown of the old Silicon Valley Consensus. Where the U.S. state and platform capital once shared a unified goal of building a globalized world of free trade and open technology, we are now seeing this order fracture along techno-nationalist lines. The vaguely socially liberal culture of old Silicon Valley is being attacked and replaced by an increasingly militant right wing that aligns with the state through government contracts and a bipolar, us-versus-them vision of global politics. Our current era is defined by competing hegemonic visions: on one side, the old order of neoliberal globalization, and on the other, a Manichean, conflict-driven vision of the global order. We are in desperate need of progressive alternatives, and the future of this transformation is far from settled.