Intel targets inference market with specialized agents and cost reductions

Convergencialatina spoke with Joao Bortone, recently appointed by Intel as managing director for Latin America. The executive highlighted the possibility of performing AI inference using traditional architectures such as Xeon, potentially replacing the use of GPUs. He also emphasized the potential of PCs with embedded AI capabilities to shift AI token processing away from large AI labs.

MORE ANALYSIS

Pablo Álvarez - Credit: Antel

América Latina · Economy · Politics20/10/2026

Sovereignty: one word, multiple meanings

At Grupo Convergencia’s recent Carriers Map Day 2026, participants discussed the concept of sovereignty and how the perspective shifts when viewed from a regional standpoint.

El encuentro se realizó en Montevideo - Credit: Asiet

América Latina · Regulation19/10/2026

Regulators discuss governance, resilience and cooperation to improve the digital ecosystem

Held in Montevideo, the 18th Regulatel-ASIET-Comtelca Summit highlighted the need to transition toward "digital regulators" with a broad perspective on the ongoing technological superconvergence. The summit also emphasized the need to strengthen cybersecurity.

América Latina · Data Center15/10/2026

The future of wholesale and the reinvention of connectivity

The wholesale business is entering a more mature phase, with portals and APIs playing a key role in driving growth. Cirion has introduced Network-as-a-Service as an overlay on its underlying infrastructure, enabling it to restore margins and expand its addressable market.

México · Operators14/10/2026

Televisa and Fintech Holdings buy AT&T for US$2 billion

The buyers will each hold a 50% stake in the new entity. Regulatory approval for the acquisition could be completed in the first quarter of 2027. AT&T has 24.7 million mobile connections. In 2025, it reported nearly US$5 billion in revenue and US$81 million in profit.

Globales · Hardware · Software and Applications · Data Center13/10/2026

Shadow financing: Hidden debt behind AI deployment has reached US$1.35 trillion

These liabilities are estimated for five companies: Meta, Alphabet, Amazon, Microsoft and Oracle. As data centers become operational, these liabilities move onto the balance sheet and become visible. The risk is that the spending fails to generate sufficient profits, potentially leading to massive write-downs.

List all analysis