Nvidia recently made the largest private investment in history: a staggering $100 billion into OpenAI. However, this substantial expenditure is not aimed at empowering individuals or fostering innovation, as Sam Altman suggests; rather, this form of vertical integration primarily concerns financial gain, control, and dominance. It represents the latest move in Big Tech’s decades-long effort to capture every segment of the digital economy—from processing units to cloud services to the applications people use. A small number of multi-trillion-dollar corporations now constitute an AI oligopoly, posing significant threats to both market competition and national security.
These companies are constructing an AI economy where they possess the infrastructure, the technology, and its diverse applications, effectively denying fair opportunities to others. Nvidia, recognized as the world’s most valuable company, has historically dominated the design market for graphics processing units (GPUs), which are essential for AI operations.
Amazon, Microsoft, and Google collectively control two-thirds of the cloud computing sector, where these chips are utilized and AI models are developed. Each of these three dominant “hyperscalers” ranks among the top five most valuable companies globally. Fundamentally, cloud services function similarly to public utilities like electricity and water; computing is a standardized service, generated remotely (in data centers) and delivered via a network (the internet). Nevertheless, unlike other utilities, hyperscalers operate without regulation, enabling them to favor certain customers over others. For most developers, this results in significant vendor lock-in and dependence, a situation that became common and accepted even before the current AI surge.
Initially, OpenAI and Anthropic appeared poised to challenge the established Big Tech players. Instead, they became intertwined with these silicon giants. What once seemed like healthy market rivalry has transformed into a cycle of Big Tech ownership, with emerging companies being absorbed before they can become genuine competitors. Microsoft and now Nvidia are OpenAI’s primary investors. Amazon and Google are among Anthropic’s leading owners. Furthermore, these major corporations have also acquired or invested in numerous other AI startups.
Tech companies frequently label these arrangements as “partnerships,” but regulatory bodies should not accept this designation. These actions clearly represent cross-ownership, industry consolidation, and sectoral dominance. It’s a recurring strategy Big Tech has employed for decades: . Google bought . Amazon utilized its marketplace data to . Microsoft pioneered this approach in the 1990s by . Apple used its . The AI strategy is no different, and it involves the very same companies that have already successfully deployed these tactics.
When an entire technological ecosystem is controlled by just a few companies, competition ceases, and collaboration morphs into collusion. The Federal Trade Commission has observed cloud providers prioritizing scarce GPUs for companies in which they hold investments over independent startups. For a startup, these identical few companies can simultaneously act as suppliers, investors, customers, and competitors. This creates unavoidable and intolerable conflicts of interest that undermine the competitive dynamics necessary for healthy markets.
History offers insights into the ultimate outcome of such scenarios and . A century ago, railway companies acquired coal mines and granted preferential shipping rates to their own cargo; subsequently, Congress compelled these rail conglomerates to divest their coal holdings. Later, telecommunications companies were mandated to allow competitors to interconnect their networks instead of walling them off. Banks were structurally separated from commercial activities to avert conflicts of interest. In the past, legislators intervened when private entities monopolized critical infrastructure, and it is long overdue for similar action in the digital economy. The rise of Big AI makes this necessity undeniably clear.
The crucial first step toward a more balanced market is to address entities that are vertically integrated, preventing platforms from competing with their own clients. Chip manufacturing must be distinct from cloud services, and cloud providers must operate independently of AI model developers. These models should compete solely on their merits, not based on their ties to a trillion-dollar sponsor. Regulators should unequivocally reject Nvidia’s investment in OpenAI. Additionally, Congress should enact legislation to dismantle other components of the Big AI ecosystem—including reversing investments and other arrangements that resemble acquisitions designed to sidestep regulatory oversight—before it solidifies even greater control over the digital economy.
Undoubtedly, Big Tech will contend that regulation would impede innovation. This assertion should be disregarded. The true innovation being stifled today is the innovation that will never materialize, suppressed by vertical integration and acquisitions masquerading as partnerships. Should lawmakers fail to act, the future of AI will be shaped not by open competition or groundbreaking ideas, but by the same small group of firms that already dominate sectors where AI promises significant economic benefit: e-commerce, search, and productivity software.
Americans previously dismantled railroad monopolies, curbed the power of banks, and mandated open networks for telecommunications. The necessary tools and clear precedents exist. What is lacking is the political will to break apart Big AI before it undermines society.
