Owning the Future: How Latin America Can Turn AI Enthusiasm into Real Digital Transformation

Jose Abraham Gomez Flores is the Founder and CEO of Zyrabit LTD, a UK-registered firm on a mission to bring sovereign AI infrastructure to enterprises. After 6+ years leading technical teams in Fintech, Edtech and cloud infrastructure projects across AWS, GCP and Azure, Abraham identified a critical gap: companies in regulated sectors were being forced to send their most sensitive data to third-party clouds just to use AI. That problem became Zyrabit, a private, on-premise AI that gives organizations full data ownership, predictable costs and zero public cloud dependency.

The Gap Between Interest and Investment

Across Latin America, the story of AI is one of contrasts. Surveys show that people in countries like Mexico, Brazil and Chile are more excited about the possibilities of AI than the global average, with enthusiasm levels above 65 to 70 percent in many cases. At the same time, the region accounts for only about 1.12 percent of worldwide AI investment, despite representing more than 6 percent of global GDP and nearly 9 percent of the world’s population. We are highly curious, but structurally underfunded.

Inside enterprises, a similar pattern appears. Many Latin American companies already report experimenting with AI or planning new initiatives in the next one to two years, and the regional AI market is growing at more than 25 percent annually. Yet a large share of these efforts is still pilots, isolated use cases or outsourced solutions rather than deeply integrated capabilities. The risk is that we become heavy users of external AI services without building the internal muscles to design, operate and govern these systems ourselves.

Taking Advantage of Talent and Open Source

The good news is that Latin America has a real strategic asset: its developers. Over the past decade, the region has become a preferred nearshore destination for software engineering, combining strong technical skills with competitive costs. Today, that talent is being amplified by generative AI coding assistants, which are growing rapidly in Latin America with double digit annual rates and an increasing share of global market revenue. For a region where cost is repeatedly cited as a key barrier to AI adoption, tools that make each developer two or three times more productive are not a luxury; they are a structural advantage.

This shift is already visible in everyday workflows. A small team in Mexico City or São Paulo can now bootstrap a full AI enabled application—backend, infrastructure as code, and observability stack—using a combination of open source frameworks, GitHub repositories and AI pair programming tools. What would have taken months of trial and error can often be reduced to weeks] or even days, because the “boilerplate” work is handled by generative assistants while engineers focus on architecture, security and business logic. The net effect is a radical compression of the cost and time required to create serious software.

Hardware is another area where the narrative is changing. For years, the message has been that meaningful AI requires access to large GPU clusters in distant data centers. That model is still important, but it is no longer the whole story. Modern Apple Silicon laptops and workstations, for example, can run surprisingly capable open source language models locally using tools such as Ollama and similar runtimes. This allows engineers in banks, insurers and public institutions to prototype generative AI workflows directly on the machines they already own, without sending any sensitive data to external providers. Local experiments become a safe sandbox where teams can learn, test guardrails and understand how AI behaves before touching production systems.

Open source is the other critical piece. Recent analysis from the Linux Foundation highlights that Latin America’s AI market is growing above 28 percent annually, and that roughly 38 percent of organizations in the region already use open source AI. For small and medium size enterprises, which make up the vast majority of businesses in Latin America, open source is often five to seven times cheaper than proprietary alternatives and far easier to adapt to local regulations, languages and edge cases. It also aligns naturally with the way many Latin American engineers already work: collaborating in public repositories, sharing solutions and building on each other’s code.



From my own work on sovereign AI architectures, including recent research presented on how to mitigate side channel risks in resource constrained environments, I have seen that moving inference closer to the data is not just a security preference; it is increasingly a practical option for our region. When organizations run generative models inside infrastructure, they already know how to monitor and defend; questions about data residency, latency and compliance become much easier to answer.

From Using AI to Owning It

The next phase for Latin America is to move from adoption to co creation. The region’s AI market is already measured in tens of billions of dollars and is projected to grow to several hundred billion by the next decade, yet it still underperforms relative to its economic weight. Closing that gap will not come from copying Silicon Valley’s investment patterns. It will come from leaning into what we already have: a culture that is unusually optimistic about AI, a large base of skilled but cost efficient developers, and a growing ecosystem of open source tools that lower the barrier to building serious, local solutions.

For CIOs and digital leaders in the region, the strategic question is shifting. It is no longer just “How fast can we adopt AI?” but “How much of this intelligence will we truly own?” If Latin America can combine its enthusiasm with deliberate investment in local talent, open infrastructure and sovereign deployments where it matters, it has a real chance not just to consume AI, but to shape it.