Videos
Why AI Adoption in Energy Is a Board Problem
Insights
- Companies no longer control their own adoption timelines, because external pressure now forces the pace of AI far more than it did in earlier technology waves.
- Governance splits into must-do and should-do, where compliance with rules like the EU AI Act is the floor and leadership has to set the direction beyond it.
- The hard constraint on AI has moved to the ecosystem around it, where data-center land and water and the coalition of government, business, and regulators needed to supply compute now decide how far adoption can go.
Recorded at ONS 2026 in Stavanger, Norway, this conversation features Ingrid Solvberg, former Director General of the Norwegian Petroleum Directorate, and Praveen Gonabal, Head of Energy for the Nordics at Infosys. Solvberg draws on three decades in oil and gas to argue that AI differs from past technology waves in its pace and in the accountability questions it raises. She makes the case that AI belongs in business strategy and a board must own it, treating it as more than an IT roadmap, and that industry bodies should set standards and share lessons the way the petroleum industry has. Gonabal contrasts how regions adopt AI, with Europe taking a slower, guardrail-first approach while other markets embrace first and resolve details later. He argues the real bottleneck is the infrastructure value chain, where land, water, energy, and compute now require society and government to act together. Both conclude that the energy industry has met every prior technology shift and will organize itself to meet this one.
Praven Gonabal:
So, it's my pleasure to welcome Ingrid Solberg, Senior Strategic Advisor to Infosys. We're here at the ONS conference, and it's my privilege to talk to Ingrid about the specifics of scaling AI. You have a rich and diverse perspective coming from your experience across multiple decades and multiple business cycles and technology waves in the energy industry. What do you feel are some of the lessons that we can learn from previous technology waves and technology implementations for this new wave of AI?
Ingrid Solvberg:
Mainly in the oil and gas industry for three decades, I've seen a lot of impressive technology development and implementation. We have gone from very manual drilling operations to advanced fields, fully digitalized, very digital drilling, use of 3D seismic for the geologists and geophysicists. So it's been a remarkable journey. Companies don't actually decide themselves anymore almost when the AI technology should be adopted. They have an external pressure because things are evolving so quickly. So I think it's the pace and then not least the accountability of the, if AI gives you a recommendation on what to do and not to do then who is accountable for the ultimate decision? And as I said, the organizational, the effects that the AI technology has on the work processes in an organization, that is a different thing than it has been previously.
Praven Gonabal:
You've been in many diverse roles in your career, from an industry executive and then to a regulator, and then back to industry executive, now as an advisor to many businesses. So, what's in your view the role of governance and how to get it right, especially for something as transformational and complex as AI adoption in the energy industry?
Ingrid Solvberg:
The governance is fundamental, both from, I would say, the regulator side. There has to be a clear governance of what the industries should be doing and not be doing. We have the EU AI Act, of course, which tells you what you must do but not what you should do. And you have the companies themselves that need to decide what they should do and how to do it. So this governance has many layers, I would say. And this about must do and should do, that also makes a gap between compliance and leadership. So you have to comply with what you must do, but then you have to have the leadership that stakes out the direction of what should be done. So the governance is an extremely important part of what needs to come in place in parallel with the adoption of the AI technology.
Praven Gonabal:
What do you feel are the roles of a regulator or industry bodies in general, whether it's a regulator or industry consortia, in helping each player in the ecosystem make the most of a solution technology like AI?
Ingrid Solvberg:
In the petroleum industry, we have been talking about not competing on access to data, but compete on the knowledge and how you analyze the data. And I think maybe we could see some similarities here to the AI technology and the use of data. And the industry associations could take a role in doing what the single companies don't do on their own. They can make industry standards and ensure sharing of information, publish reports so the industry and the companies can learn from mistakes like we have done in the petroleum industry. And then of course, you have the boards of a company. What kind of role does a board have and the leadership have in a company? And this AI isn't a technology project. It's not an IT roadmap. It's a part of a business strategy for a company. And this should be requested by the board, the board should ask the right questions as of to, how do we create value from AI? How does AI underpin our overall business objectives? And the CEO should own this and make an organization that is ready to adopt AI through work processes and also letting the employees build trust in the AI technology.
Ingrid Solvberg:
What kind of experience do you have with working with different regions in the world? Are there differences here in how the companies adopt or develop AI and their fear or their embracement of AI?
Praven Gonabal:
Yeah, that's a very good question, Ingrid. I think there are definitely differences that you see from a geographical perspective. Partly cultural, partly historical, partly regulatory, and increasingly, because of geopolitical reasons as well. So, what we see is in some parts of the world, there is more of a tendency to, and this seems almost like a cultural thing, where new technologies come in, and there's an embrace of the technology and then figure out the details later. Whereas what we see in Europe is a more cautious, much more responsibility-focused approach to AI adoption, where they look at what could go wrong and how do you anticipate some of those things, put in place the guardrails to make sure that those things either don't happen, or if they do happen, then the implications will be that much more manageable. So that has an inherent impact on the pace of AI adoption or technology adoption in general. So, which also means that as a technology or digital services provider for Infosys, we need to work with different approaches to adopting the same technology, whereas it could be a difference in the pace, it could be a difference in the relative importance of technology versus change, technology versus governance. So, it's never a dull moment. I think it's a whole portfolio of approaches that we're seeing across the world now.
Ingrid Solvberg:
How do you view the question around compute capacity and how this should be prioritized, and by whom? When we have all the data centers coming…
Praven Gonabal:
No, absolutely. I think that's such a — all the infrastructure — that's such a fundamental question these days. If you look at the entire value chain for compute capacity or, more broadly, the infrastructure needed for AI, whether it is for training AI, whether it is for fine-tuning AI, whether it is for inference, once AI is in place, the entire value chain, there's so many bottlenecks now. And all the way from physical space for data centers to water for cooling, energy for running those data centers. So it's almost every one of those things has become quite a contentious thing, and for good reason, because the societal impact of each of these things is not small either. So when it comes to making the infrastructure available in the scale for the entire value chain for AI, the role of society and the role of government, the role of regulators, I think it's going to be increasingly important. Otherwise, if you just leave it to the market to sort itself out, which has traditionally been perfectly fine when it comes to any technology adoption, but because of the complexity and the far-reaching implications of every aspect of AI, all the way from, we're talking about availability of land in the first place, right, and planning permission and things like that. Unless you have a much broader coalition of society, business, governments and technologists working together, and working together with a degree of long-term perspectives, it's going to be really hard to really get this to deliver to its potential. So quite a complex question to really address. But it's going to be fundamental to the success of the initiative, success of AI in delivering value that all of these different players come together to answer this in a more sustainable fashion. It's not the technology that has been the constraint previously. It's kind of the ecosystem and the organizations themselves that have been the constraints in succeeding with adopting and...
Ingrid Solvberg:
Absolutely.
Praven Gonabal:
the technology. So it's interesting times for us going forward.
Ingrid Solvberg:
Interesting times indeed.
Praven Gonabal:
Yeah.
Ingrid Solvberg:
Interesting.
Praven Gonabal:
So we will see how this goes, but I'm confident that, as with every technology wave in the past, the energy industry will rise to the occasion, and this entire ecosystem that we spoke about, they will come together in some ways to really figure out what the sustainable path to success from this is. So, and it's going to be interesting few years, and we live in some truly exciting times.
Ingrid Solvberg:
Yeah, we really do. Yes.
Praven Gonabal:
So thank you. Thank you, Ingrid. That's something to look forward to.
Ingrid Solvberg:
Thank you. Yeah.
Praven Gonabal:
So it's been my pleasure to have Ingrid with us addressing some of the most pressing questions of our times in the energy sector and the technology industry. Thank you once again, Ingrid.
Ingrid Solvberg:
Thank you.
