Who Benefits Most from AI by 2028? The Divide Is Already Visible

No evidence can identify the winners of 2028–2029 with certainty. What can now be seen is the mechanism of advantage: AI benefits are accumulating around organizations and places that combine computing capacity, investment, usable data, skilled people and the ability to integrate tools into real work.

The important update is that this is no longer simply a contest between the United States and China. Production remains highly concentrated, but adoption is spreading into middle-income economies and smaller technology hubs. For creators, freelancers and digital businesses, location still matters, yet access to audiences, affordable models, local knowledge and a repeatable workflow may matter more than proximity to a frontier laboratory.

The first winners own infrastructure and distribution

The most concentrated gains are likely to remain with companies that build models, chips, cloud systems and data centers, along with investors and highly skilled employees who hold equity in them. These actors can charge for access to scarce infrastructure while also using AI inside their own operations.

The geographic concentration is measurable. The 2026 Stanford AI Index reports that the United States hosted 5,427 data centers and attracted $285.9 billion in private AI investment in 2025, compared with $12.4 billion in China. It also says industry produced more than 90% of notable frontier models that year. Those figures describe control over production, however—not an automatic guarantee that every American worker, creator or small company receives an equal dividend.

Distribution is a second source of power. Businesses that already own customer relationships, payment systems, professional software or large audiences can insert AI into an existing revenue stream. A technically impressive model without customers may create less captured value than an ordinary tool embedded in a widely used marketplace or production process.

Adoption capacity separates users from beneficiaries

Access to a chatbot is not the same as capturing economic value. A person or company benefits when the tool is connected to reliable data, trusted processes, capable workers and a product that customers will buy. That distinction favors organizations able to redesign work rather than merely generate isolated text, images or code.

The divide is already visible inside relatively wealthy economies. An OECD analysis of emerging AI divides found that 13.5% of EU enterprises with at least ten employees used AI in 2024, up from 8% in 2023. Across the OECD area, 39% of large firms used AI, compared with 12% of small firms. The report also found faster uptake in capital regions, knowledge-intensive services and places with stronger innovation systems.

This makes the likely 2028 split more granular than “rich countries win.” Large firms may pull away from small businesses in the same city; technology-oriented regions may outperform other parts of the same country; and workers with similar occupations may face different outcomes depending on whether their employer invests in training and workflow redesign.

Middle-income countries are not confined to the losing side

Lower-cost models, cloud access and tools that run on ordinary devices create a route around some infrastructure barriers. Countries do not necessarily need to train a frontier model domestically to apply AI in commerce, education, agriculture, health or media. They do, however, need dependable connectivity, relevant data, digital skills and affordable access to computing resources.

The World Bank’s AI foundations findings show both sides of this picture. More than 40% of global ChatGPT traffic came from middle-income countries in mid-2025, led by Brazil, India, Indonesia and Viet Nam. Yet high-income countries held 77% of global co-location data-center capacity as of June 2025, while low-income countries held less than 0.1%. The Bank identifies connectivity, compute, locally relevant context and competency as the four foundations of effective adoption.

That combination challenges the idea of a single developing-world trajectory. Large, connected markets with strong technical talent can become important adoption and localization centers even without leading frontier research. Economies with unreliable electricity, expensive internet, limited digital skills or little content in local languages face a much harder route.

What the divide means for creators

For creators, the clearest near-term advantage belongs to people who can combine AI with something the model does not independently possess: a trusted audience, recognizable judgment, subject expertise, original access, local language fluency or direct knowledge of a community. AI can reduce the cost of drafting, translation, editing, research organization and format conversion, but inexpensive production also increases the supply of competing material.

This shifts scarcity away from producing a technically acceptable asset. Attention, verification, distinctiveness and distribution become more valuable when many competitors can generate serviceable text, audio or video. Creators who already control newsletters, communities, storefronts or client relationships are therefore positioned differently from equally talented newcomers who depend entirely on algorithmic discovery.

Regional context matters here. English-language creators currently benefit from broad tool support and large addressable markets, while creators working in underrepresented languages may encounter weaker model performance or less relevant training material. The same gap can become an opportunity for people able to supply trusted local reporting, culturally accurate adaptation, specialized datasets or human review.

Small teams also have a plausible counterweight to corporate scale: they can change workflows quickly. A creator-led studio does not need to own a data center to automate transcription, prepare multiple formats or serve customers across borders. Its vulnerability is dependence on outside platforms whose prices, visibility rules and model access can change.

A plausible regional map for 2028–2029

The evidence supports a conditional map, not a settled ranking. The United States is positioned to capture a large share of producer returns because of investment, models and infrastructure. China remains a major research, industrial and deployment power. Other advanced economies can benefit substantially where firms adopt quickly, although gains may cluster in leading cities, large organizations and knowledge-intensive industries.

Connected middle-income economies can capture meaningful application-layer value through large user markets, technical labor, local-language products and exported digital services. Smaller advanced states can also perform well when they combine dependable infrastructure, skilled populations and access to international compute. By contrast, low-income economies with weak digital foundations risk receiving AI products later, on less favorable terms and with fewer opportunities to build locally owned businesses.

The most defensible forecast is therefore about capabilities rather than flags. By 2028–2029, the largest producer gains are likely to flow toward owners of compute, models, capital and distribution. The broader gains will depend on whether firms, workers and creators elsewhere can obtain affordable access, adapt systems to local needs and retain ownership of the resulting customer relationships and intellectual assets.

The future is not predetermined by today’s headquarters map. But falling model costs alone will not erase disparities in electricity, connectivity, skills, finance or bargaining power. The regions that convert access into locally useful products—and the creators who convert automation into trusted work—have the strongest path from merely using AI to benefiting from it.

Also read:

• Elon Musk's New Government Department Seeking "Super High-IQ" People Who Will Work for Free
• Mac Mini M4 vs M4 Pro: Who Actually Benefits From the More Expensive Chip?
• Peter Thiel and the Project of a New Humanity: Immortality, Techno-States, and the Rule of the Chosen"They built a world for everyone. I'm building one for those who survive."