HPE’s Director of Private Cloud for AI for Latin America and Southern Europe comes to Cartagena with a thesis that challenges what the industry took for granted just two years ago. 

Until the end of 2024, the answer to where artificial intelligence was being developed seemed clear: the public cloud. That certainty has begun to shift. Today, more companies and public-sector organizations are building their own private AI clouds, driven by considerations that executive leadership teams understand clearly: greater control over data, cost predictability, and the freedom to choose open models. 

That is the conversation Eric Martorell Bouvier brings to ANDICOM 2026. 

Thirty-Four Years, Four Continents, and More Than 1,000 Customers 

Martorell is HPE’s Director of Sales for Private Cloud for AI, with responsibility for Latin America and Southern Europe. He has spent 34 years at HPE and HP, holding roles across sales, marketing, finance, procurement, and technology in Switzerland, the United States, the United Arab Emirates, and France. Throughout his career, he has worked with more than 1,000 HPE customers worldwide. 

His most relevant credential for this conference, however, is more recent: he built HPE’s regional artificial intelligence team from the ground up, and today that region is the company’s most successful globally. He is not speaking about AI adoption in the abstract—he is actively driving it. 

A Spanish and French citizen with a Master’s degree in Business Management from EDHEC Business School and currently based near Geneva, Martorell has spent 25 years presenting to customers, media, and audiences of several hundred people. In a previous role, he was repeatedly recognized among HPE’s top presenters at its global briefing center and represents the company at forums such as IDC events. 

The Problem No One Anticipated: Sovereignty, Confidentiality, and Cost 

Artificial intelligence initiatives built on public cloud infrastructure are facing three challenges that emerged after the initial wave of enthusiasm. 

The first is sovereignty. When data and models reside on third-party infrastructure, potentially in another jurisdiction, organizations may lose control over assets that can be critical or subject to regulation. The second is confidentiality, because every query sent to an external model means information is leaving the organization. The third is cost, which can scale in AI projects in ways that few budgets initially anticipated. 

The conference “AI Adoption Trends and Strategies While Meeting Sovereignty Requirements” starts from an optimistic premise: today, it is already possible to address all three challenges without sacrificing the power of the technology—and with a return on investment that can be demonstrated. 

Martorell explains why artificial intelligence, initially driven by a handful of large technology companies, is taking a more independent path through its adoption in private data centers and the continued growth of the open-source ecosystem. This combination changes the equation: both public-sector organizations and private companies can choose their own path to AI adoption rather than simply accepting the options available to them. 

This is the conversation that closes the ANDICOM 2026 agenda—and potentially one of the conversations with the greatest implications for technology investment and business decisions in the years ahead. 

Friday, September 4 · 11:40 a.m. · Nueva Granada Room, Main Academic Forum 

Explore the agenda and register. 

ANDICOM 2026 · September 1–4 · Cartagena, Colombia 

Register at andicom.co