Fernando spent three decades in energy, but it was a bold move toward high-performance computing and AI that completely reshaped his career. From bringing the first Bull–NVIDIA supercomputer to Brazil to leading AI-driven breakthroughs like Petrobras’ PetroNeMo, he’s been pushing the industry into its next chapter. In this brief interview, he breaks down what really matters as energy meets AI — and what’s just hype. His message is blunt: innovate with purpose, or prepare to get left beh
The Social Digest: Fernando, your 30-year career in energy spans from law to AI-driven transformation at NVIDIA. What sparked your pivot to emerging tech, and how has this journey shaped your perspective as a leader in a rapidly evolving sector?
The year was 2007, and I was at Bull Latin America, leading the Energy sales team, which covered the Electric Power and Oil & Gas sectors. For strategic and relationship reasons, we opted to place the main Oil & Gas potential customer, Petrobras, under my direct commercial management, as Bull’s portfolio at that time was heavily concentrated in the private sector. Therefore, in an effort to optimize the trajectory of a new technology partner within the account (which was Bull’s situation at that moment), a differentiated approach was necessary for Petrobras, which is a mixed-capital company with strong government participation. This was because the pre-existing relationship was a key differentiator to accelerate a sales cycle that was much longer than those Bull was accustomed to handling with other clients.
I looked inside the company to see what innovative, high-value-added solutions we could bring to Petrobras, to differentiate ourselves from the competition and try to move away from the commodity market that service provision had become—an extremely competitive market where we likely would never be able to compete with already well-positioned companies. In Europe, I discovered that Bull France was starting to create a focused area to sell HPCs (Supercomputers), and the main niche was the Energy sector, more specifically, Oil & Gas, where Geophysics was the primary client. I then decided to bring this business to the region.
Engaged with the person in charge of the area, and in a short time, we were in Brazil, working to deploy the first Supercomputer outside France. This materialized in 2008 when we sold Bull’s first HPC outside France. This HPC was powered by NVIDIA’s then little-known GPUs. That was my first contact with the company, which has been part of my professional journey ever since. Eleven years later, in 2019, I would join the “green team” to lead the Energy, Oil & Gas, and Higher Education and Research verticals—no longer just a GPU company, but above all, a promising AI company that, in 2022, would definitively take the world stage as the locomotive of this new industrial revolution: Artificial Intelligence.
The Social Digest: At NVIDIA, you contributed 70% of regional revenue through the energy vertical. Can you share a pivotal project where AI solutions drove sustainable outcomes in Oil & Gas or Mining?
Each industry has a distinct level of maturity in adopting AI for its operations, and not every company is capital-intensive like the Oil & Gas industry. This industry has increasingly been using the benefits of AI to aid in the analysis of seismic and well data, in order to more accurately identify the presence of hydrocarbon reservoirs, and in production optimization, by adjusting operations in real-time to maximize extraction and improve energy efficiency.
Therefore, what I would highlight is that we are gradually seeing AI-powered HPC infrastructure gain prominence, compared to classic HPCs with the sole function of accelerated computing. This is a paradigm shift that, in E&P (Exploration and Production), will bring more sustainable results in the medium and long term, allowing the classic E&P activity, accelerated by AI, to enable and speed up the energy transition.
In the Mining sector, which you mentioned in the last part of your question, what we are seeing is the pursuit of consolidating several fragmented AI projects, many focused on computer vision, to connect all these initiatives and create a kind of ‘AI of the Production Chain from Mine to Port,’ thus democratizing the use of AI across the entire company. In Latin America, we can highlight Vale as one of the companies pursuing these goals.
The Social Digest: Scaling AI in traditional industries like energy requires bridging legacy systems with new tech. What strategic challenges have you faced, and how do you overcome them to ensure successful adoption?
We can cite a specific project: PetroNeMo. The PetroNeMo is a Generative AI assistant that transforms the maintenance of critical assets at Petrobras, reducing the repair recommendation analysis time from weeks to minutes. The solution was trained with language models and internal data, mastering the technical jargon (‘petrolês’), in addition to sourcing information about these assets directly from SAP.
Essentially, PetroNeMo aims to create an AI agent to assist the engineering team. The beauty of this project’s construction is that it relies heavily on the strong exchange of information with Petrobras’ internal engineering team and the technical teams from NVIDIA and Deloitte. This is a classic example of how successful AI projects should be conducted, by identifying areas or niches where the technology can bring a tangible and measurable economic differential. According to Petrobras, the company expects to save US$4 million by 2029 through efficiency gains with the assistant, even considering conservative metrics. I had the pleasure of commercially leading this initiative, the vision for which arose from a meeting with the company’s Executive Management, who captured the vision and decided to invest in the initiative, and in a few months, we were already developing the project.
All challenges can be overcome when, once the business area that will benefit in the medium and long term is identified, the project proves to be economically viable. That is why I have advocated that AI projects should be niched, initially small, and gradually scaled up, thus avoiding inefficient projects without a clear and defined ROI (Return on Investment).
The Social Digest: Your work with partners like Atos, TCS, and Microsoft has expanded NVIDIA’s ecosystem. How do you evaluate and nurture these collaborations to maximize value in emerging tech?
Curiously, I have worked at all these companies you mention; for me, in a way, it was like seeing friends again. NVIDIA, rightly so, has the premise of operating and commercializing its products through its ecosystem of partners. Therefore, it has programs that support this strategy, which allows it to distribute its technologies across various industries, from Health to Education, from Energy to Retail, and so on.
Companies that want to expand their operations, maintaining their prominence in this context of fierce and increasingly AI-accelerated competition, must above all invest heavily in the partner ecosystem, whether they are startups, service providers, OEMs, and so on. In a world of daily constant transformation, there is no longer room for companies on isolated trajectories. We have seen this in recent announcements of billion-dollar partnerships between OpenAI and various companies in Silicon Valley.
Therefore, my work at NVIDIA was a job of many achievements and innovative projects, mainly built and enabled through constant relationships and interactions with a continuously growing ecosystem.
The Social Digest: AI’s role in energy transition is transformative, but ethical concerns like data privacy and equity persist. How do you address these in your projects, and what critical decisions have shaped your approach?
At this point, we enter a topic that is currently exhaustively discussed: the concept of Sovereign AI. This is a strategic concept that refers to the ability that a nation, a regional bloc, or, I would say, a company, holds to develop, control, and utilize AI technology using its own infrastructure, data, expertise, and regulations.
Within the company, projects like PetroNeMo, which I mentioned in previous questions, can, by analogy, be cited as a classic example: the creation of an in-house model is a sound strategy for the privacy and security of strategic data. In a broader context, looking at a nation or regional bloc, Sovereign AI is crucial. That is why we are witnessing a race to create national Artificial Intelligence infrastructures, some already in full development and others in the planning phase.
Each country, just like each company, has its strategic differential, and to remain geopolitically relevant, it needs to consider Sovereign AI as part of its strategy and planning, otherwise, it risks trailing behind other countries or blocs. Knowledge is the new oil, and the fuel that moves and democratizes this knowledge for the benefit of a people, a nation, or a company is, or will be, Artificial Intelligence from now on. This is inexorable.
The Social Digest: As an emerging tech leader, what global changes in AI—such as edge computing or federated learning—do you see as most disruptive for the energy sector, and how is the industry preparing?
These two approaches help to democratize the use of Artificial Intelligence in the Energy sector, representing the migration from centralized AI to distributed AI.
Edge Computing, for example, solves problems of low latency, reliability, and, above all, infrastructure in high-risk environments. Data processing on an oil platform or at an electrical substation at the edge is absolutely revolutionary, allowing for autonomous and immediate predictive or preventive maintenance, thereby revolutionizing companies’ operational safety. We increasingly see Digital Twins being built from this edge data, and from there, we will see more and more industrial plants being managed remotely without the need for direct human intervention, reducing the risk of loss of life in the field.
Federated Learning, on the other hand, addresses the issue of data sovereignty, allowing a wide range of data from different units (e.g., multiple refineries or partner companies) to be used without the raw data leaving each entity’s local server. Only the model parameters are shared and aggregated.
This allows energy sector companies to train much more accurate and robust AI models (based on a larger volume of information) for tasks such as demand forecasting, trading optimization, or predictive maintenance, while maintaining full regulatory and sovereign control over their proprietary data.
Preparation is seen in Sovereign AI initiatives, where the priority is the development of governance infrastructures and models that enable this collaboration. The focus is on obtaining the benefit of collective intelligence without exposing intellectual property.
The Social Digest: Leading business development in a competitive market requires resilience. Can you share a setback in scaling AI solutions and the key lesson it taught you?
Perhaps by luck, I have not encountered a setback, but rather an extreme anxiety among companies wanting to implement Artificial Intelligence projects at all costs, perhaps due to the hype itself, without the slightest direction on where to start. For these companies, my approach of slowing down and dispelling myths—that AI would solve all problems immediately and bring a quick return on investment—may have generated some frustration, but I do not regret it one bit. This is because the implementation of AI projects, for them to become scalable and justifiable to the high decision-making layer of a company, requires being viewed as a journey, above all, of learning. And, again, it is necessary to identify the niche where (however small) any innovative implementation will bring a productivity gain, mitigate a risk, accelerate an activity, reduce a cost, and so on.
I believe that setbacks often come from the approach of wanting to meet all needs at the same time without an exact vision of the complexity. Every AI project requires investment in infrastructure and services; therefore, planning and strategy are needed to know where this investment will actually be best applied. This avoids future disappointments.
The Social Digest: What emerging trends in AI for energy, like predictive maintenance or carbon capture optimization, excite you most, and how are you positioning clients to capitalize on them?
We can cite the creation of engineering agents for both E&P (Exploration and Production), which is the capital-intensive area of an oil operator, and for those that later expand to other business areas, such as midstream, downstream, and so on. Furthermore, Generative AI is increasingly permeating other solutions, such as Edge Computing, Digital Twins, etc., providing ever-increasing speed in identifying problems and “creating” solutions to mitigate future issues.
Ultimately, clients are migrating from a scheduled maintenance model to a model where AI guarantees them zero unplanned downtime. This translates into a direct ROI (Return on Investment), converting saved production hours into millions of dollars in revenue and significantly increasing the lifespan of expensive assets.
It is increasingly possible to simulate the behavior of a carbon plume or create a Digital Twin of a CCUS (Carbon Capture, Utilization, and Storage) even before the project is created, mitigating or even avoiding future risks, design errors, optimizing costs, and so on.
Thus, CCUS projects become a value business, not just a cost. AI is used to reduce the cost per captured ton and provide the necessary long-term technical validation for regulatory compliance and investor confidence, ensuring the viability of their low-carbon portfolio for the future.
The Social Digest: How do you foster inclusive teams in tech development, ensuring diverse perspectives drive innovation in energy solutions?
Firstly, the approach must be consultative: it is necessary to understand the client’s language. Technology alone does not solve or address problems. It is essential that you deeply understand the moment of that industry, its local, regional, and geopolitical challenges. You must be able to navigate between the technical and business layers with the same ease and work to converge the two areas toward the same purpose. This means not underestimating any involved interlocutor.
At the same time, it is extremely important to know how to listen and curb the tendency to think that your company holds the solution to all problems when, in reality, it is part of a gear composed of several actors. Secondly, and more challenging (I always say this), is selling the project internally. If you have internal support, you will have all the necessary resources for the project to become a success.
At NVIDIA, for example, I was part of a Global Energy team (based in Houston) made up of experts on the topic, which facilitated obtaining resources to promote successful projects, as we spoke the same language.
The Social Digest: In a world of AI hype, how do you critically evaluate which technologies will truly transform energy sectors, and what red flags do you watch for?
Investments in AI by technology companies have reached stratospheric levels, never seen before. Many experts fear that AI is a bubble, just as investors are betting on increasingly exponential gains, which has generated expectations far beyond the reasonable.
My opinion is that AI has moved past the hype to real-world implementation and application. In the medium term, there may be adjustments in expectations, but AI is definitely the revolution of our era, simply because it works based on data created by us over many years. This data is accumulated knowledge, research, industrial data, information—created and recreated every day. AI ingests it, transforms it into tokens that are monetized, and the industry that sidelines itself from this transformation will simply disappear or lose relevance.
Furthermore, another technological leap is beginning to emerge on the horizon and gain prominence: Quantum Computing. It will bring a range of other possibilities and accelerate advances in various areas, and again, AI will have a super important role in this revolution, functioning as the main tool to design quantum software and stabilize quantum hardware.
The Social Digest: For emerging tech leaders from diverse backgrounds, what advice would you offer to navigate challenges and scale innovations in energy?
My advice would be to deeply understand the industry they intend to work in, and to be interested in its pain points and challenges. It is necessary to be aware of geopolitical changes and trends, as in an increasingly multipolar world, new actors are emerging while others are threatened, which generates tension. And this affects the values of commodities, which impacts freight, which ultimately impacts the relevance of one project over another. It is necessary to be attentive to these nuances to define the best approach strategy.
Regarding AI, what calls my attention most today and is sine qua non for its expansion is the issue of energy, which is what will sustain the “manufacturing environment of Als”. Therefore, this is a challenge that is imposed on all regions and directly affects issues related to Data Sovereignty, AI democratization, investments, the creation of new Data Centers, regulations, and energy scarcity. All of this brings to the forefront other important themes: climate change, energy transition, energy matrix, alternative energies, and so on. But that is a topic for another interview…
This interview was conducted by Ansh C Vachhani, The Social Digest on 01/11/2025. If you have any interview recommendations or have a story that you want to share with our readers, get in touch with our editor Vedant Bhrambhatt, at editor@thesocialdigest.com
