Remarks by Ahmad O. Al-Khowaiter at LEAP 2025

Ahmad O. Al-Khowaiter, Aramco EVP Technology & Innovation

Speech|RIYADH|February 09, 2025

Your Royal Highnesses, Excellencies, distinguished guests, ladies and gentlemen,

Let me begin by thanking His Excellency Minister Abdullah Alswaha, the Minister of Communications and Information Technology. Under his leadership, LEAP has become one of the world’s leading technology conferences and a must visit destination for entrepreneurs, executives and even technology enthusiasts.

This is just the fourth edition of LEAP but every year it gets bigger and better. And as many of you know, we at Aramco are proponents of technology, developing, deploying and driving cutting edge innovation.

Not only because our main products, energy and chemicals are essential for the technology our world depends on, but because we recognize the immense power technology has to transform our own industry and the world.

We know this because technology has driven so much change over the last 100 years. But, the advent of Artificial Intelligence and big data is taking us into a new world with almost limitless opportunity at a pace I don’t believe we have ever seen before.

For our business it is helping our environment, by improving sustainability. It is helping our people, by improving safety. And it is helping our customers, by improving reliability.

Reliability alone, in unplanned downtime and equipment failures cost industries across the world about $1 trillion globally. So that’s about equal to the GDP of the Kingdom of Saudi Arabia, roughly.

So new AI models allow us to predict the future health of equipment and minimize failures before they happen, as one small example. Not only does this mean increased equipment availability, but also reduced impact on the environment. So, our use of AI starts at exploration.

We are using old seismic data, which we utilized in the past and reprocessing it with AI and we are now able to generate new results from old data, which creates great savings.

In production it means we can be more accurate and we are able to drive our drilling operations autonomously to be able to maximize production, reducing costs and reducing emissions.

So those just a few examples of AI applications, but the real challenge is not to actually develop these application, but rather, how to scale it across industrial sized operations, not just to utilize it, but maximize value.

And maximizing its potential requires three main elements, three main enablers:

One is tremendous amounts of real-world data. You need the data first.

Then you need to put in place computing power, computing infrastructure, to be able to do the models, although recently, given Deep Seek’s success, maybe not as much as we thought before.

But finally, and probably the most important element which I think we tend to forget in our excitement around technology, is you need the talent, you need the subject matter experts who can tell you if the model is telling the truth or not.

These three factors are helping Aramco get ahead, and stay ahead.

We have more than 90 years of propriety data from our extensive geological, seismic and process surveys and every day we collect information from 10 billion data points across our facilities, which goes straight to our engineering solutions center.

All of this helps us to optimize our performance. And this was recognized recently, yet again, with an unprecedented Fifth Global Lighthouse Award from the World Economic Forum, this time for our North Ghawar Oil Producing Complex.

We are the only international energy company to have more than three of these awards, which recognize the adoption of Fourth Industrial Revolution technologies. And when it comes to computing power, we have investing heavily recently.

We recently increased our data storage capacity to 1,500 petabytes and we doubled our data center power. We operate a diverse set of supercomputers, including Dammam 7, one of the fastest in the region and a number of NVIDIA Superpods. Last year we announced our partnership with Groq to deliver the first AI inferencing center in the region and I’m glad to say we together we have delivered on time and on budget.

Last year we also announced the launch of aramcoMETABRAIN, the world’s first industrial Large Language Model. Today it is serving our business with a 70 billion parameter model and I’m excited to say we have a one trillion parameter model in development.

And as DeepSeek showed, building capable AI models isn’t limited to global tech companies. It is within reach of enterprises, even start-ups, to design AI suited to their own businesses.

We have believed this from the beginning, developing our own models with our own data.

Which is why it gives me great pleasure to introduce our latest innovation, Aramco MetaBrain for Plants, a time series transformer model, utilizing large time series data sets. Using these data sets we’re able to model the process, the real time processes that underly our operations and we are able to provide actionable insight to operators, engineers and scientists

By working in real time, and with minimal user input, it will also anticipate demand, optimize operations, predict product quality and maximize production.

This frees our experts up to focus on more value-added tasks, rather than being busy with troubleshooting or developing models from the ground up.
There is no point in having a treasure trove of data or the latest, cutting-edge infrastructure, if the talent is not there to take advantage of it.

At Aramco, we believe our AI is only as good as our HI, or Human Intelligence.

This is why we are training more than 6,000 AI developers across the company and are working with world leading institutions like Imperial College, Caltech, and KAUST to put hundreds of our employees through advanced degree courses.

We’re also using the engineers, scientists and operators we already have to work with those AI developers to train new models, making them more robust and more reliable.

Before I conclude, I want to take a step back. All these new models, all these supercomputers, have to be powered from somewhere. Right now, AI and data centers consume about 3% of global electricity demand.

Indeed, with all the developments we are talking about here at LEAP and elsewhere, estimates predict this demand may double by the end of the decade. Notwithstanding the recent models developed, I still believe the energy demand is actually potentially even higher.

Why do I say that?

Training was perceived to be a huge energy consumer, but by bringing down the cost of training in many ways we’ve created a bigger market, because that makes the use and inference of the models much more affordable for much larger consumption around the world.

We can imagine that inference now will go to devices, as we talked about, AI at the edge, even machines, cars and if that happens then we can imagine the tremendous amounts of energy required to serve that tremendous inference demand.

Aramco is well placed to help meet this increased demand for affordable, reliable and sustainable energy.

Our investments in lower carbon technology like Carbon Capture and Storage, hydrogen and renewables can help to lower the carbon footprint of this growth in demand…the development of all of course been enabled and accelerated by AI.

Ladies and gentlemen, to conclude, Aramco is not only using our unique size and scale to maximize the use of AI on an industrial scale, but our decades worth of data and most importantly, our subject matter experts, the talent of people.

That keeps the human experience and human ingenuity at the heart of all we do and will let us leap into a new world of industrial excellence.

Thank you.