Ángel NavarretePhotographs
ADD THE WORLD ON GOOGLEMake our news appear in your searches
Tips of interest Carl Aaby, CEO of Ikea: "It's more than fine that people only come to Ikea to eat, one day they will come back and buy a sofa"
Tips of interest Nicolas Brusson, CEO of BlaBlaCar: "We can add services such as BlaBlaCar hosting in a large trusted community"
IBM celebrates its centenary in Spain this year, a notable figure, but much more typical of an illustrious tavern, a bank or an industrial company than a technological one. Nothing remains of the stenographic machine business to carry out the United States census that catapulted the company in its beginnings and brought it to Spain with Telefónica to extend the telephone network throughout the country. Other emblematic businesses such as personal computers have also left along the way.
Now, the business is marked by others with less visual impact, but perhaps even more relevant, such as artificial intelligence, the cloud and services for companies, waiting for quantum computing to finally abandon its status as a promise and become a commercial reality in 2029, as the company foresees. Horacio Morell, president of IBM in Spain, Portugal, Greece and Israel, assesses how AI has changed the lives of every technology worker and its impact on the future and the brakes it still faces.
Against the global trend of fear of an Apocalypse for humanity caused by an uncontrolled agent, Morell imposes a realistic tone in the interview and, although he makes it clear that there is not room for all actors in this market, he considers that current macroinvestments will be seen as something positive for the economy over the years.
IBM turns 100 years old in Spain. It seems strange that a technology company has been active for so long. How are you living it?
It is true that the first comment they make to me is whether it is 100 years at a global level. We have been in Spain for 116 years and 100 in an industry where there is continuous disruption and it is difficult to survive. It has forced us to reinvent ourselves several times. Spain is one of the first countries where IBM arrived and we have helped its technological transformation.
What IBM does now has nothing to do with before.
Of course, in the end I always say that we have innovation in our DNA. Having one of the most powerful innovation centers in the world, such as IBM Research, guarantees you to continually reinvent yourself. We have more than 4,000 scientists, 16 laboratories and six Nobel Prize winners who have passed through there. Always looking for that innovation applied to the industry and its real problems.
What moments mark IBM's presence in Spain?
Here the first client was Telefónica. At the time we helped it expand throughout the country and began to standardize its processes. We brought the first corporate computers, like the one from Renfe, and then the big projects to computerize the country: the first democratic elections or the referendum on the Constitution, where our technology gave confidence in the results, the ATMs, the first POS in El Corte Inglés and Inditex... There are two events in '92, which was a very special year, the introduction of Spain to the world. We were behind all the technology of the Barcelona Olympic Games and Expo 92. Then all the digital transformation, the creation of the National Supercomputing Center...
Furthermore, the last few years have been intense.
Yes, we created the cloud region here in Madrid, the largest investment in our history in Spain. We bought Bluetab and opened the quantum computing center in San Sebastián, which positions us and gives us an opportunity to develop this disruptive technology. And the latest has been the opening of the digital sales center for all of Europe and the Middle East in Valencia. We have laid the foundation for the next 10 years.
He spoke of a changing world. How do you assess fears of an end to humanity created by out-of-control AI? Are they exaggerated?
We were at the Dartmouth conference 70 years ago where the concept of artificial intelligence originated. Since then we have always been researching and developing it. In 2018 we issued a letter of commitment of intent that says that AI must be used to expand the capabilities of human beings, the data belongs to its author and it must be done in an explainable way, with transparency and traceability.
And regulated?
Now the debate is whether we are going too fast. And it's true. We believe in regulated AI and even more so being positioned in a business environment, where things are done well. Scalability, regulation and privacy must be taken into account. It seems good to me that there is regulation of AI, but we must find a balance with innovation because other countries play with different rules. The ultimate decision has to be human. A pizza recommender is not the same as an air traffic manager. Depending on the risk, a level of regulation must be established.
Change the perspective from a consolidated company of more than 100 years to a startup of less than 10...
Absolutely. We are going out to clients to set up the AI operating system. There are and will be different models - giant or much smaller - that are going to be key to scaling in the business world. Our vision is that the frontier models are going to be a commodity and there will be three or four. That doesn't mean they have no value. Gold is also a commodity and gives value. We mean that the cost of changing from one to another will be very low.
What problem do you see with the great models?
The problem with these models is that they are a universal encyclopedia concept. They are valid for many things. What good is it in a company if AI knows how to make paellas if the cost is 20 times higher than a small model specialized in my domain of knowledge? There is a debate that is about cost and it is one of the inhibitors of this going faster. We have to help choose the best model for each case. And there is another issue. Not everything is going to be AI. It is a probabilistic technology and, as such, it makes mistakes. We can make him do it less and less, but you have to see how far you can tolerate that risk. If you make a mistake on the pizza, nothing happens, but a credit policy must always be issued the same way.
When you have conversations with clients around AI, what are they most concerned about?
Our debate is return on investment. We have passed a stage of experimentation and there are studies that say that companies are not making a return. The technology works and is proven and the recommendation we give now is to take one or two use cases and scale it industrially. There you will capture value and manage change. This is not about doing the same thing automatically, it is about changing the way of doing things and will impact the way of working.
Where does it provide the most value?
Fundamentally there are three areas where we see very great potential productivity and it is worth starting down that path. In customer service and contact center, in an operator, AI already manages 70% of calls and user satisfaction is 35 points higher than with those answered by people. It gets it right 95% of the time compared to 85% for humans. Then there are the developers, who will increase their productivity. We have changed our way of making software. And the third, three years ago we made a project to be the most productive company in the world. We have generated $4.5 billion in increased productivity and this year we have committed another $1 billion to business support functions.
But investors fear the impact of AI on its business and its stock fell sharply after the latest results.
We are seeing a world of change in which we have never seen the speed at which things happen. There are going to be winners and losers here because there is overinvestment, but there is no boom. The net will be positive. Here are companies that are dedicated only to AI that have entered into a race of investing and having data center and supply capabilities that are generating brutal inflation even in hardware. It is a race that if you only dedicate yourself to this you have to be there and win. But then, there are other technology companies that we all know that have more business units than AI. I see it as fiber optics in the 2000s. There were companies that went bankrupt, but we all later took advantage of the networks.
Does the market believe in your strategy?
We are an AI and hybrid cloud company. And it is true that the market is very volatile. The ability to develop software has been substantially improved. We are going to generate more and more products and faster. It's brutal productivity. We trust in the strategy with which we have almost tripled the market value. There has been a quarter in which operations have been delayed, but a large part has been closed now and we continue to trust that we have the appropriate strategy. When you look at the acquisitions we're making, we have the best portfolio in our history right now. We see it as an opportunity.
Do you fear a social backlash against AI if jobs are eliminated?
The focus is to increase the capabilities of the human being. There are going to be certain tasks and certain jobs that are going to be displaced, but I believe that it is going to derobotize us and free us from repetitive and routine tasks. We have seen this in other revolutions such as the industrial revolution. When assembly lines were industrialized, there was a loss of jobs, but then more people were needed to make more cars. We went from thousands of cars to millions of cars. I believe that there is not an employment crisis, but there is a skills crisis, that is, there is this concept of adaptability, of learning and unlearning continuously. People must embrace change to have a better competitive position.
What is the impact within IBM?
At IBM, with our transformation, the employment result is net positive. We have signed people, there are people who have left, yes, but 60% of the total have requalified and continue working. Technology companies have a social responsibility to leave no one behind and we are committed to training 30 million people by 2030. We are applying AI, for example, in Human Resources. There is no more sensitive data than that of people, so you can see how much we trust in the security of our AI. Decisions about people are still made by people, but if someone changes departments, administrative operations previously required different systems and a dedicated worker. Now it is done automatically. 94% of HR queries are managed with AI, 24 hours a day, seven days a week.
Back to your investment in Spain, how is your cloud going?
We are very happy. It serves all of Europe and is a cloud that stands out because it goes to companies in regulated environments and is the most secure and the most open because we do not keep the client prisoner. We started with CaixaBank and several financial entities and many public sector agents and clients from other countries such as Greece, Israel and all of Europe have joined. We are seeing a movement of many small and medium-sized companies that are interested in its reliability. We are not a hyperscalar, we play in cloud very specialized in business environments, but we are very competitive in terms of price and there are many people who see that costs are getting out of control in the world of hyperscalar.
Will these costs moderate?
They will continue to grow. Our vision is that, in three years as a result of what we have seen with AI, the cloud processing unit will double in price. There are clients who say they have committed budget reductions of 30% thanks to AI for productivity; However, this cost will be multiplied by two. Both things happen at the same time. When it comes to AI, costs must be controlled.
How is it controlled then?
Choosing the model well is essential. The other day I read that a user with a flat fee of $200 had used 1.1 billion tokens, which at Anthropic's internal cost was $27,000. Are you going to put your AI strategy in a company that is not passing on the real cost to you? That transfer is already happening. That is why we are committed to transparency and open models that cost 20 times less than a large model and you will be able to incorporate your data and customize it for greater performance. The cost issue is an inhibitor to scaling AI. A client told me that to make summaries with AI at the current cost it was cheaper to do it with people.
The issue of sovereignty is a hot topic. There is a European proposal that I don't know what you think.
The first thing is, what is sovereignty? There are several levels, the data, who operates it and other governance and auditing. We believe that the important thing is to have control of critical elements regardless of where the technology is manufactured. We allow you to work with one technology or another and transfer your data to your servers. So there we are very well positioned. There are others who take their cloud and their data center and you can't get out of here. But sovereignty is not isolation, our vision goes much further and is about joining forces with people who have the same European and Western values and you can trust. The BSC without IBM would not have brought supercomputing to Spain and the same goes for quantum computing. We provide the quantum processor to the Basque Government and everything is open.
And isn't IBM worried about a paradigm that limits critical data to European companies owned by Europeans?
That's why I don't want to be put in the same bag. If you want, you can operate your technology and I will help you operate it and even if an order comes from anywhere in the world to access it, we will not comply. We have a technology in the cloud that gives the client access to the data and they have the key, but I technically cannot. To the point that if the client loses the key, they lose access to the data.
The other big line is quantum. They have declared that they want one of the first commercial computers to be in the Basque Country. Do you see it possible?
Yes, last year we opened the quantum computing center with the computer in San Sebastián. We have trained 300 professionals and there are more than 20 research centers and 30 companies, some like Iberdrola, with which we work on computer maintenance of digital assets, bus route issues, materials, risk optimization. In 2029, we have committed to having a fault-tolerant computer, that is, on a commercial scale and will be able to solve real problems on this scale. Now it's research, but
