AI around the world in 2026: what has changed and what to expect

The global artificial intelligence market moved US$ 826 billion in the first half of 2026. This figure is 47% higher than the same period in 2025 (Gartner data, June/2026). But what has truly changed is not in the numbers—it's on the streets, in the courts, and in the laboratories.

AI is no longer an IT department issue. In 2026, it is a topic of foreign policy, labor regulation, and geopolitical disputes. Three economic blocs—the United States, the European Union, and China—have adopted radically different governance models. And the rest of the world tries to balance between them.

This overview analyzes the main transformations. And answers: what to expect going forward?

The new map of global regulation

In May 2026, the European Union completed the full implementation of the AI Act, its artificial intelligence regulation law. The text classifies AI systems by risk level: minimal, limited, high, and unacceptable. Mass social credit systems and real-time facial recognition in public spaces were banned (AI Act, Article 5, 2026).

The US took the opposite path. Instead of a single federal law, the country adopted a mosaic of sectoral regulations. The FDA regulates AI in diagnostics. The FTC oversees commercial use. The Department of Defense controls military applications. The result is fragmented, but agile for innovation (Brookings Institution, May/2026).

China, in turn, launched the third version of its AI Development Plan in February 2026. The focus has shifted: it was once about the quantity of patents. Now it is about technological sovereignty. The country requires that every generative AI model used within Chinese territory be trained on local servers (Ministry of Science and Technology of China, February/2026).

"AI regulation in 2026 is no longer about preventing harm. It's about defining who controls the computational infrastructure of the future." — Dr. Yukari Tanaka, Director of the Tokyo Institute of Digital Governance, in an interview with NeuralPulse (May/2026)

Brazil, India, and South Africa still do not have specific laws. But all three have advanced with bills in progress. The expectation is that at least two of them will be voted on by the end of 2026 (Global South AI Observatory, June/2026).

Corporations redesign value chains with AI

If 2025 was the year of testing, 2026 is the year of integration. Large companies are restructuring entire operations based on AI. The logistics sector is the most advanced.

DHL announced in April that 73% of its delivery routes in Europe are optimized by AI in real-time. The system considers traffic, weather, demand, and even local events. The savings were 18% in operational costs (DHL Annual Report, Q1 2026).

In retail, Amazon expanded the use of predictive models for inventory. The company now maintains an average of only 12 hours of inventory in its US distribution centers. Before AI, it was 72 hours. The reduction in working capital was US$ 4.2 billion (Amazon 10-Q, May/2026).

The financial sector has also changed. Banks like JPMorgan and Itaú use AI for credit analysis, fraud detection, and customer service. JPMorgan reported a 34% drop in undetected fraud in the first quarter of 2026 (JPMorgan Chase, April/2026).

Table: AI impact by sector in the first half of 2026

The data shows a clear pattern: AI is not replacing jobs en masse, as was feared. It is transforming processes. And those who do not adapt lose competitiveness.

The energy bottleneck and new hardware

A growing problem in 2026 is the energy consumption of large AI models. Training a model like GPT-5 (launched by OpenAI in January) consumed the equivalent of 2,300 American homes for a year (Carbon Tracker Initiative, February/2026).

This has forced a race for more efficient hardware. NVIDIA launched the B300 chip in March, promising 40% more performance per watt. AMD responded with the MI400, focused on inference. Both companies have a waiting list of over six months (TechInsights, May/2026).

China, without access to the most advanced chips due to sanctions, has advanced in alternative architectures. Huawei presented the Ascend 920, manufactured with domestic 5-nanometer technology. Performance is inferior to the B300, but energy consumption is 12% lower (Huawei Tech Day, April/2026).

The energy bottleneck has also driven the search for green data centers. Google announced that 100% of its AI data centers have been operating on renewable energy since March 2026. Microsoft promised to reach the same mark by the end of the year (Google Environmental Report, March/2026).

What to expect from the second half of 2026

Three trends will dominate the coming months.

First, regulatory consolidation. The European AI Act will serve as a model for at least five Latin American countries that are expected to present bills by September (UN, May/2026). Brazil could be the first.

Second, the talent war. The global market for AI specialists has an estimated deficit of 1.2 million professionals (LinkedIn Workforce Report, May/2026). Companies are hiring machine learning engineers with salaries 40% above the IT average.

Third, the rise of open AI. Models like Llama 4 (Meta) and Mistral Large (France) are gaining ground against proprietary solutions. The justification is twofold: lower cost and transparency. Medium-sized companies prefer models they can audit (Hugging Face State of AI, May/2026).

Artificial intelligence in 2026 is no longer a promise. It is infrastructure. It is in car engines, medical diagnoses, weather forecasts, and movie recommendations. The question is not whether it will change the world. It already has. The issue now is who will set the rules for this change.

Related Articles

Also check out: South Korea Against AI Chaos: The Law That Could Change the Game in Asia and Affect Brazilian Companies Also check out: The New Chip Cold War: How the US-China Dispute is Redrawing the Global AI Map in 2026 Also check out: The World AI Map in May 2026: EU Retreats, Malta Innovates, Google Creates a New Mouse, and the US-China Race Reaches Boiling Point