The global artificial intelligence boom is being driven not by a diffuse wave of innovation but by a relatively small group of extraordinarily powerful companies whose influence now stretches across supply chains, capital markets and national economic planning. In a recent bulletin published by the Bank for International Settlements, economists Jon Frost, Kumar Rishabh and Vatsala Shreeti argue that a handful of “global AI giants” are rapidly becoming systemically important to the world economy, with the scale of their investments and the breadth of their operations giving them unprecedented influence over how the technology evolves.
According to the study, artificial intelligence is not a single industry but a layered ecosystem that spans specialised chips, cloud infrastructure, data systems, foundational models and user-facing applications. Building and maintaining each of these layers requires immense upfront capital, from semiconductor fabrication plants to hyperscale data centres, while the economics of the sector reward companies that can operate across multiple stages simultaneously. This combination of high fixed costs and cross-layer integration has naturally favoured very large firms capable of mobilising resources on a global scale.
The bulletin identifies the United States and China as home to most of the dominant firms, with additional critical players located in Chinese Taipei, South Korea and the Netherlands. Many of these companies are already among the most valuable corporations in the world, and their presence across several segments of the AI supply chain allows them to bundle services, internalise costs and move into adjacent markets faster than smaller competitors. The result is a reinforcing cycle in which scale enables expansion, and expansion further entrenches scale.
Researchers found that the largest AI firms now account for an increasingly significant share of market capitalisation within their home jurisdictions. By the end of 2025, such companies represented roughly 30 to 40 percent of total equity market value in several advanced economies. Their weight is not confined to stock markets alone. The same firms are responsible for a rising share of capital expenditure, research spending and industrial revenues, indicating that AI investment is no longer a niche technological bet but a major driver of macroeconomic activity.
The report suggests that these corporations are now in a position to shape aggregate investment trends and influence the direction of innovation itself. Massive spending on research and development, coupled with aggressive construction of data centres and computing infrastructure, has allowed them to set the pace of technological change. When these firms increase investment, suppliers, startups and even governments often adjust their own strategies in response, amplifying the economic ripple effects.
At the heart of this transformation is the structure of the AI supply chain. The lowest layer consists of computing power, especially advanced microprocessors and high-bandwidth memory capable of handling the intense calculations required to train and run modern AI models. Above that sits infrastructure such as cloud platforms and data centres, followed by data ecosystems that feed models with vast multimodal datasets. Foundation models form the next stage, acting as adaptable engines for a wide range of uses, while applications—from automated assistants to industrial analytics—occupy the top layer closest to consumers and businesses.
Historically, companies tended to specialise in one or two of these segments. The BIS research shows that today’s leaders increasingly operate in three or four layers at once, with some spanning the entire chain. Over the past two decades, major firms have steadily expanded outward from their original lines of business, acquiring startups, investing in complementary technologies and forging partnerships to secure access to computing resources and proprietary data. This expansion has enabled them to control both inputs and outputs, strengthening competitive advantages that go beyond any single market.
Deal-making patterns underline this shift. Large AI companies now account for a substantial proportion of investment transactions across AI-related sectors, including model development and application deployment. Since around 2017, their acquisitions and venture financing have tilted heavily toward AI-specific markets rather than unrelated industries, signalling a strategic consolidation of focus around the technology’s long-term potential.
The economic footprint of these firms is already visible in national accounts. In some economies, their capital expenditure alone constitutes a sizeable share of total private investment. Estimates cited in the bulletin suggest that AI-related spending reached levels equivalent to several percentage points of gross domestic product in major economies, highlighting how digital infrastructure is beginning to rival traditional industrial investment as a driver of growth.
Yet the rise of these giants also raises concerns about market concentration and innovation dynamics. The report notes that emerging technologies often follow a life cycle marked by intense early competition followed by consolidation as weaker firms exit. While vertical integration can improve efficiency and accelerate breakthroughs, it may also create barriers to entry if dominant players gain tight control over essential inputs such as computing power or proprietary datasets.
This possibility has prompted policymakers worldwide to explore what the bulletin describes as “sovereign AI” strategies aimed at reducing dependence on foreign technology providers. Governments are weighing whether to build domestic capabilities to ensure resilience or rely on global supply chains to avoid duplicating enormous fixed costs. The trade-off is complex: national self-sufficiency may enhance security but risks fragmenting innovation and inflating expenses, while deeper integration could increase vulnerability to geopolitical tensions.
For smaller economies, the research suggests a pragmatic alternative to competing head-on with established leaders. Rather than attempting to replicate full-stack AI ecosystems, these countries may find advantages in specialising in narrower segments such as domain-specific data, customised applications or niche computational tools. Such targeted participation could allow them to benefit from the AI boom without incurring the prohibitive costs associated with building every layer of the technology stack.
The study also underscores the importance of regulatory frameworks that preserve competition and interoperability. Ensuring fair access to data, promoting multi-cloud environments and maintaining vigilant competition policies are seen as essential to preventing excessive concentration while still encouraging innovation. Financial regulators, meanwhile, are being urged to monitor systemic dependencies that could arise if critical infrastructure or services become too closely tied to a small number of firms.
Ultimately, the BIS bulletin frames artificial intelligence not merely as a technological revolution but as a structural shift in the global economy. The companies building AI systems are becoming gatekeepers of productivity gains, investment flows and even geopolitical leverage. Their decisions about where to build data centres, which markets to enter and how to price access to computational resources may increasingly influence economic outcomes far beyond the technology sector itself.
As nations race to harness AI’s transformative potential, the report concludes that the central challenge will be balancing innovation with resilience. The world must decide whether the future of artificial intelligence will remain concentrated in a few dominant hubs or evolve into a more distributed ecosystem that spreads its benefits—and its risks—more broadly across the global economy.
