The global race to develop and control artificial intelligence has entered a highly volatile, politically charged phase. As the technology transitions rapidly from a specialized research novelty to the primary engine of modern economic productivity and national security, governments around the world are no longer content to serve as passive regulators. Across Europe, the Middle East, and Asia, states are actively launching “Sovereign AI” initiatives, investing billions of dollars of public capital to build their own national data centers, cleanroom infrastructure, and state-owned foundation models.
This rising trend of state-directed technology development has triggered a powerful, highly visible counter-reaction from one of America’s most influential business leaders. In a blunt, uncompromising opinion piece published on Bloomberg Opinion, Michael Bloomberg delivered a stark warning to international policymakers. He argued that building, funding, and operating state-owned, government-controlled artificial intelligence platforms is a fundamentally dangerous idea that poses an existential threat to democratic values, free enterprise, and national security.
The warning from the former New York City mayor and billionaire businessman arrives right as the political consensus around technology regulation is fracturing. While some progressive lawmakers and nationalist leaders promote public ownership as a necessary tool to bypass the monopoly power of Silicon Valley conglomerates, Bloomberg asserts that putting the state in control of the primary information engine of the 21st century is a recipe for economic inefficiency, political censorship, and the rapid rise of an inescapable surveillance state. To protect the future of human freedom and economic dynamism, he argues, the development of artificial intelligence must remain firmly anchored in the competitive, private marketplace.
The New Frontier of State Capitalism: The Rise of Sovereign AI
The transition toward state-backed technology is one of the defining structural trends of the current global economy. Over the past year, several major governments launched ambitious programs to build out their own sovereign computing power. At the AI Action Summit, France announced plans to invest billions of public capital to fund a government-owned AI model and build supporting high-performance data centers. At the same time, the European Union has actively promoted “InvestAI,” a massive project designed to construct public AI factories across the bloc to reduce its reliance on American technology.
This trend is not confined to Europe. In the United States, prominent political figures have floated ideas of making the American public a “partner” or direct shareholder in leading AI companies, using federal lands and state-backed energy grids to support their operations in exchange for a share of the corporate equity.
Furthermore, the U.S. military has embarked on its own aggressive path toward state-owned technology. Following a dramatic contract collapse with Anthropic over ethical limits on military applications, the Pentagon initiated a major strategic shift, constructing its own secure, government-owned AI environments in federal data centers to train its next-generation defense models.
This rapid rise of state capitalism in the technology sector is a deeply concerning development. He argues that by allowing governments to build, own, and operate these massive computing systems, society is inviting an unprecedented level of political interference, economic waste, and algorithmic bias into the very plumbing of the digital age, completely undermining the open-market principles that historically drove the Western world’s technological and economic leadership.
The Economic Argument: Why State Capital Cannot Out-Innovate Silicon Valley
The primary economic critique outlined in Bloomberg’s warning centers on the absolute inefficiency of state-run technology programs. For more than a century, the history of industrial policy has proved that governments are notoriously poor at picking winners, managing high-tech operations, and driving rapid, market-disrupting innovation.
Private venture capital and public equity markets are infinitely better and faster at allocating capital, managing risk, and driving technological breakthroughs than government bureaucracies.
The software and semiconductor industries operate at a breakneck, highly unpredictable pace.
A cutting-edge artificial intelligence model developed today can easily become completely obsolete in a matter of months as newer, more efficient architectures enter the market.
A government agency, bound by rigid procurement rules, annual budget cycles, and intense political scrutiny, simply cannot move fast enough to iterate on these models, ensuring that state-owned AI projects will inevitably lag behind the fast-moving commercial frontier.
The Inevitable Pitfall of Bloated, Obsolete State Projects
When a private venture capital firm invests $100 million in a failed AI startup, the loss is absorbed by private investors who understood the risks.
When a government invests billions of taxpayers’ dollars into building a state-owned foundation model that eventually fails to find market traction or becomes technologically obsolete, the public is left holding the bag.
Bloomberg warns that state-owned technology initiatives historically turn into bloated, bureaucratically paralyzed money pits.
Because politicians are highly reluctant to admit failure and cut their losses during an election cycle, they will continue to funnel precious public funds into underperforming projects long after they have lost their commercial viability.
This waste of public capital deprives other critical sectors—such as primary education, public healthcare, and physical infrastructure—of the funding they need to support everyday citizens, hurting national productivity.
Distorting the Market and Crowding Out Private Startups
Using taxpayer funds to build and operate state-owned AI models also creates a highly dangerous, market-distorting crowding-out effect. When a state-owned entity enters the market, it does not have to operate under the same financial discipline as a private company.
It does not need to achieve profitability; it has access to cheap, subsidized capital, and it can rely on preferential government procurement contracts to secure its market share.
This unfair competition makes it incredibly difficult for independent, local technology startups to survive.
Private venture capital firms will naturally refuse to invest in a local tech ecosystem if they know that their startups must compete directly with a government-backed, tax-exempt national champion.
By attempting to build a sovereign AI monopoly, governments risk destroying their domestic startup pipelines, leaving their tech sectors stagnant, uncompetitive, and entirely dependent on state subsidies.
The Authoritarian Threat: Propaganda and Algorithmic Censorship
While the economic arguments against state-owned AI are compelling, the most terrifying risks of the trend are political and ethical. The power of modern large language models lies in their ability to serve as the primary gateway to human knowledge, answering questions, writing essays, summarizing historical events, and shaping how users perceive the world.
If a government owns and operates the primary foundation model used by its citizens, it gains the absolute, unchecked power to control the flow of information.
The state can quietly tweak the underlying algorithms to generate state-approved propaganda, censor political opponents, and manipulate public opinion during critical election cycles, completely undermining the foundations of democratic public discourse.
The Weaponization of the Algorithm for Political Control
This risk of political manipulation is not a hypothetical concern. In several authoritarian regimes, state-run media and internet regulators already use advanced AI algorithms to filter search results, block sensitive political keywords, and track the online activities of dissident groups in real time.
If a democratic government builds its own sovereign AI platform, it places a highly dangerous weapon in the hands of whoever wins the next election.
A populist leader or an authoritarian-leaning administration could easily use their executive authority to redirect the state-owned AI to generate biased information, attack their political rivals, and discredit independent news organizations.
By separating the state from the technology, the private market serves as a vital, highly resilient shield for free speech, ensuring that a diverse array of independent, competing platforms exists to challenge any single, centralized narrative.
Building the Foundation for a Mass-Surveillance State
The consolidation of advanced artificial intelligence under direct state control also provides the perfect, highly dangerous infrastructure for the rapid rise of an inescapable mass-surveillance state.
Modern AI models are incredibly skilled at processing, indexing, and finding patterns across vast, unstructured databases, including facial recognition streams, mobile location registries, financial transactions, and social media communications.
If a government owns and operates both the physical data centers and the underlying AI models, it can easily integrate these separate databases into a single, centralized brain.
This integration would allow the state to monitor, profile, and grade its citizens’ behavior in real time, creating a highly restrictive surveillance network reminiscent of China’s controversial social credit system.
To protect personal privacy and civil liberties, the physical computing infrastructure and the algorithmic brains must remain separated from the state’s law enforcement and intelligence apparatus.
The Pentagon Precedent: The Ethical and Operational Clash
The dangerous tension between national security requirements and ethical artificial intelligence principles became exceptionally clear during the U.S. military’s recent, highly publicized “AI Exodus.” The conflict highlights why keeping advanced computing systems entirely under government control without independent, private-sector oversight increases the risk of deploying unsafe, autonomous technologies.
In March, negotiations between the Pentagon and leading AI developer Anthropic collapsed entirely over fundamental disagreements regarding the military’s desired level of access to the company’s frontier models.
Anthropic’s leadership team insisted on binding contractual safeguards that would prohibit two specific, highly controversial applications: the mass surveillance of American citizens and the deployment of autonomous weapon systems capable of firing without direct human intervention.
The Anthropic Contract Collapse and the Military AI Alternatives
The Pentagon’s Chief Digital and AI Officer, Cameron Stanley, confirmed that the military refused to accept these ethical limitations, choosing instead to launch a massive, in-house program to build its own secure, government-owned AI models in federal data centers.
The military wanted absolute operational freedom to integrate AI into its command-and-control systems, and officials were unwilling to let a private company dictate their ethical boundaries during times of national security crisis.
This military AI exodus is a classic warning sign of the dangers of state-controlled technology.
By building its own proprietary models in closed, government-owned data centers, the military has successfully bypassed the ethical guardrails and independent oversight designed by private-sector researchers.
Without these independent checks, the risk of deploying unsafe, un-audited autonomous weapons that can make life-or-death decisions without human intervention increases exponentially, highlighting the critical need for private-sector cooperation and public transparency in even the most sensitive national security applications.
Reimagining the Policy: The Path to Balanced AI Governance
To protect our democratic values and maintain our technological edge, global leaders must reject the dangerous siren song of state-owned artificial intelligence.
Instead of trying to buy up the factories of compute or build their own national models, governments should focus their energy and resources on their traditional, highly effective roles as regulators, coordinators, and early adopters.
The path to balanced AI governance requires several key policy pillars:
- Establishing Clear, Enforceable Safety Standards: Rather than building models, the state should focus on writing clear, predictable, and legally binding safety regulations that private companies must satisfy before launching their systems to the public, similar to how the FAA regulates commercial aviation.
- Promoting Open-Source Research: Governments can support technological diversity and prevent corporate monopolies by funding open-source research and providing public grants to academic institutions, ensuring that smaller developers have access to the resources needed to innovate.
- Using Tax Incentives to Drive Infrastructure: Instead of spending billions of dollars to build state-owned data centers, governments should use targeted tax credits, streamlined permitting, and land grants to encourage private utility and tech companies to build secure, green-powered infrastructure domestically.
- Acting as a Sophisticated Buyer: The public sector can drive immense efficiency and improve its public services by acting as a sophisticated buyer of private-sector technologies, purchasing specialized AI tools to automate administrative tasks, optimize public transport, and improve healthcare delivery without taking on the financial risks of technology development.
By establishing these clear boundaries, the international community can ensure that the digital revolution remains safe, open, and highly competitive, protecting human freedom and democratic values while continuing to drive economic growth and prosperity across the globe.
The global technology transition has arrived at a critical, defining crossroads. The choices made by policymakers over the coming years regarding the governance and ownership of artificial intelligence will shape the future of human society for decades to come.
By rejecting the dangerous, highly inefficient model of government-owned AI and keeping the technology anchored in the competitive, private market, we can ensure that this historic technological revolution continues to serve as a powerful tool for human empowerment, rather than a dangerous instrument of state control.
