Brazilian infrastructure is not yet ready for large-scale AI, Dell study finds

3–5 minutes

A pilot AI project that performed well in testing was approved by the board and authorized to be expanded across the entire company. It was at this point, when AI moved from the lab to daily operations, that many Brazilian companies discovered their infrastructure couldn’t handle the load.

According to the Modern Enterprise Readiness Study, commissioned by Dell Technologies, 59% of Brazilian executives surveyed say their companies’ data centers and infrastructure are still not fully prepared to run modern AI applications in production.

The study reveals a surprising finding about the source of the problem. For respondents, the bottleneck isn’t just processing capacity, which is the focus of the global race for chips.

For 74% of executives in the country, storage performance and data access speed have already become bigger bottlenecks than computing capacity in many critical applications within their companies.

Why has storage become AI’s main bottleneck?

AI models rely on data at all times. To answer a question, analyze a document, or recommend a product, they need to retrieve information quickly, often spread across different systems. If access to this data is slow, it doesn’t matter how powerful the processors are—they’ll spend part of their time waiting.

“For a long time, discussions about technological capacity focused mainly on processing, but AI applications also depend on fast data access, adequate storage, and integration between different environments,” says Joel Brawerman, director of Dell Technologies’ Infrastructure Solutions Group for Latin America.

According to him, when these components don’t evolve together, limitations arise that make it difficult to expand projects.

A fragmented infrastructure

Another obstacle is fragmentation. In Brazil, 68% of respondents say their companies’ infrastructure is not fully unified.

In practice, many companies operate simultaneously with their own servers, one or more public clouds, and legacy systems that have never been fully integrated. This model allows applications to be distributed as needed but complicates data management—the very resource AI depends on most.

Companies’ response has been to simplify. For 86% of respondents, the goal is to concentrate servers and storage on fewer, more powerful platforms to handle data-intensive applications and reduce operational complexity.

When asked about priorities for modernizing data centers, servers, and storage, executives cited the following:

  • Cybersecurity and resilience: 46%
  • Performance and scalability: 45%
  • Simpler management and automation: 41%

Energy has entered the equation

The research shows that energy efficiency has become as important as performance. For 86% of Brazilian executives, energy efficiency is essential when making infrastructure purchasing decisions.

An even more striking finding is that 67% will only expand AI use if it can be done without significantly increasing energy consumption. They also expect technology suppliers to help meet this goal.

This concern aligns with a global discussion. AI data centers consume increasing amounts of electricity, and Brazil is trying to attract these investments by promising clean and abundant energy through the Redata program, which offers incentives to the sector. The dispute took a new turn this week when the United States began requiring a license to send advanced AI chips to the country.

From pilot to large-scale production

For Dell, the research shows that companies’ challenges have moved to a new phase. Testing AI is no longer the problem; the difficulty lies in maintaining its operation at scale.

“Preparation for artificial intelligence cannot be measured solely by the ability to execute a pilot project. The bigger challenge is to put these applications into production and ensure they continue to function with performance, security, and sustainable costs,” says Brawerman.

The study was conducted by consulting firm Vanson Bourne in June 2026, with 2,950 respondents across 32 countries. In Brazil, 100 decision-makers were surveyed. Therefore, the percentages reflect the perceptions of this group of executives, not a measurement of the infrastructure of all companies in the country.

Frequently Asked Questions

Are Brazilian companies ready for AI?

According to Dell’s research, 59% of executives surveyed in Brazil say their companies’ data centers and infrastructure are still not fully prepared to run AI applications in production.

What is the main bottleneck for AI in companies?

For 74% of Brazilian executives surveyed, storage performance and data access speed are already bigger bottlenecks than processing capacity.

Does energy influence AI use in companies?

Yes. For 86% of executives, energy efficiency is essential when purchasing infrastructure, and 67% say they will only expand AI use if it does not significantly increase energy consumption.

How was the research conducted?

The study was commissioned by Dell and conducted by Vanson Bourne in June 2026, with 2,950 respondents across 32 countries. In Brazil, 100 decision-makers were surveyed.