<strong>-Convergencialatina: How can Latin America’s corporate sector, which often operates with tight budgets, be persuaded to run workloads and inference on traditional architectures such as Xeon rather than relying on GPUs?

-Joao Bertone (JB): AI is a powerful and transformative workload, but it is still a workload. The architecture depends on how a company plans to deploy it. There is the path taken by AI labs, with their LLMs and the race to develop superintelligence. Then there is the other path: inference, which does not necessarily require GPUs. AI agents can be created on an AI PC to learn specific functions and apply them to well-defined tasks.

We see AI becoming more distributed and accessible, rather than necessarily being concentrated in large computing centers that require multimillion-dollar infrastructure.

We are working with the entire ecosystem to showcase inference models. Obviously, we are also involved with large LLMs because a GPU needs a CPU to operate. My role, together with my team, is to work more closely with our customers and showcase this technology.

-Convergencialatina: Which verticals do you see as the most interesting for this model?

-JB: It depends on several factors, but the first one we look at is finance. There are also other industries, such as manufacturing, with a focus on specific areas that can benefit from our technology. A good example is robotics, which is an area where Intel is investing. Government is another area, and we are working with several countries to showcase our platform.

-Convergencialatina: Which Intel products are best positioned for this type of performance?

-JB: Our role is not just to talk about our products, but to showcase the platform. Intel is known for its chips, but we go beyond that. For example, we have a software stack that helps accelerate the capabilities of our processors. ISPs, for example, can use it to accelerate software within their applications using our underlying technology.

We are working across ERP, CRM and virtualization, among other areas, with customers so they can use our technology to further enhance the capabilities of our platform. Another example is our work with hyperscalers and their customers to show them that using their virtual machines on our platform can deliver significant performance gains. Our role is to move the entire industry and ecosystem forward.

-Convergencialatina: Do you have any data on the replacement of traditional PCs by PCs with embedded AI?

-JB: This is a good point. At some point, very soon, all PCs will be AI PCs. They will have the technology built in. Why is that important? Because the cost of using AI through these large hyperscalers is becoming increasingly expensive. Some of that work can start being done on the PC itself, without incurring token costs. Large manufacturers already offer small machines, similar to mini PCs, designed for local AI development. So companies can begin developing and working locally, significantly reducing the cost of using cloud-based tokens.

-Convergencialatina: What plans do you have for partners in the region?

-JB: I used to be on the other side. I worked with manufacturers for the past 15 years. I know how important a channel is to a manufacturer, especially because no manufacturer has enough coverage to reach all of its potential customers. At Intel, we have programs that cover several points across the chain. For example, we work with Global System Integrators. We have specific groups working with the ecosystem at each point in the chain, including wholesalers, distributors, and local and regional channels. We look at the ecosystem as a whole, and that includes all of these areas.

-Convergencialatina: What do you expect in terms of public policy, tax incentives and regulatory developments?

-JB: I see a window of opportunity for our region in data centers. The pace of investment in this area is very strong, and Latin America has a unique opportunity because it has space and energy, is relatively free from natural disasters and is a stable region. Several countries have already begun taking steps to attract investment. Chile is a good example. So are Brazil, with Redata, and Colombia. There are initiatives across the region that suggest we could become a major data center hub for the world. Another thing that, as a citizen, I would love to see is a policy that enables the full use of technology in education, which would require reducing affordability barriers. There could therefore be public policies aimed at achieving this.