The key facts

  • AI chips and servers accounted for about one-sixth of global goods trade in 2025 and 42% of the sector's value growth (WTO forecast)
  • Of the ~$2.92 million in tracked value added of the Nvidia GB200 NVL72 server, 68% was profit, mostly going to Nvidia
  • Labor's share of the final value is 14.2%: 4.3% for production workers and 9.9% for engineers and managers

AI Boom Profits Are Concentrated at the Top Stages of the Supply Chain

The production of artificial intelligence spans the entire world, but according to an estimate by the UN Conference on Trade and Development (UNCTAD), most of the value created remains with the companies that control technology design and intellectual property. Profits from the AI boom get 'stuck' at the high-margin stages of the supply chain.

The Sector's Scale in Global Trade

According to a World Trade Organization forecast, AI chips and servers accounted for about one-sixth of global goods trade in 2025, delivering 42% of the sector's value growth. AI server production involves a broad chain — from raw material extraction to semiconductors, power equipment, and cooling systems, with each country receiving different economic benefits depending on the stage available to it.

Division of Roles Between Countries

Chile and Peru supply raw materials for AI equipment, South Korea and Taiwan dominate semiconductor manufacturing, and Malaysia and Mexico lead in final assembly. According to UNCTAD, the combined value of the silicon, copper, and tin used in advanced server systems amounted to only 0.2% of the final sales price.

Breakdown Using the Nvidia GB200 NVL72 Example

UNCTAD analyzed the production of the Nvidia GB200 NVL72 server system, worth about $3.3 million. The study tracked the activities and supplier groups that account for more than 87% of the value added: of the roughly $2.92 million in tracked value, 68% was profit after deducting operating expenses and taxes. The vast majority of the profit went to Nvidia, which designs the chips but outsources physical production.

Labor's Share and the Position of Developing Countries

Only 4.3% of the tracked value went to workers engaged in the physical production of the system, and 9.9% to employees outside production (managers and design engineers); labor's total share was 14.2%. The AI boom has increased mineral exports from developing countries but has not yet raised their share of the wealth created by the chain. Assembly in developing markets creates jobs, but the economic returns lag behind the income from high-tech design.

Recommendations and Barriers

In the 'Trade and Development Report 2026,' published on October 9, UNCTAD called on developing countries to shift their focus from attracting foreign investment to strengthening local suppliers and training skilled personnel. The report notes that it is very difficult for developing countries to penetrate the upper levels of AI component production: the sector requires enormous capital and faces geopolitical barriers.