Artificial intelligence is rapidly reshaping the global economy, and the insurance, investment management, and software sectors are already feeling the impact. Recent developments, including a partnership between ChatGPT and Spanish firm Tuio, have triggered significant market reactions, as highlighted by Goldman Sachs’ analysis of a recent downturn in insurance stocks. But the massive investments pouring into data processing centers and the soaring valuations of AI-linked companies are raising concerns among investors.
Tony Kim, a fund manager with 30 years of experience in the financial sector and two decades focused on technology investments, offers a perspective on the current landscape. Currently a leader in technology investments at BlackRock – the world’s largest asset manager – Kim oversees a portfolio valued at $39 billion (33 billion euros). He recently visited Milan and Madrid to share his insights with major Southern European banks.
Q: Is there a stock market bubble in AI?
A: I don’t believe so. The universe of companies linked to AI is very broad, ranging from small startups to industry giants. There are companies focused on programming, chip manufacturing, and algorithm development. It’s not accurate to generalize and speak of a bubble.
Q: But there have been excesses…
A: In some areas, yes, we’ve seen companies significantly overvalued, particularly smaller, indebted firms that aren’t publicly traded.
Q: Not in the larger companies?
A: No. When you analyze the multiples of the tech giants, you’ll find many are actually trading at lower valuations than others.
Q: Does this situation remind you of the dot-com bubble burst in 2000?
A: If we appear at the valuations those tech companies reached back then – the relationship between their price and their revenue or profits – they were much higher than they are now.
Q: Are there cheap technology stocks available today?
A: Yes, the programming companies. Some are trading near their all-time lows due to the risk that AI will disrupt their business model. Most depreciated significantly in 2025. Those providing technology services are also relatively inexpensive. However, semiconductor companies are trading at high prices. It’s overly simplistic to speak generally about the technology sector; some parts are relatively cheap, although others are more expensive.
Q: Are you concerned about cross-investments and stakes between companies in the sector?
A: Not particularly. Some of the AI giants have a fabulous capacity to generate cash. They could distribute that money to shareholders in the form of dividends or buy back shares, but they’ve chosen to allocate that capital to support smaller companies in this ecosystem grow. This will facilitate the expansion of the entire sector and ultimately increase demand for the products of those now investing.
Q: Are there risks of conflicts of interest?
A: Of course, problems can arise. A company could go bankrupt if it has a very indebted balance sheet. But when you look at the big picture, you see large, solvent companies with a great capacity to generate income and profits. They earn so much that they have to find somewhere to invest all that money. But I really don’t see a systemic risk.
Q: The amounts being invested in AI are astronomical, reaching trillions of euros annually. Will these investments be able to generate returns?
A: Technology currently represents only a fraction of global GDP – less than 10%. And companies in this sector believe they can gain significant ground in other parts of the economy. Those who invest do so since they don’t want to be left behind. Companies that don’t will end up lagging and disappearing.
Q: But the growth rates are staggering…
A: Orders for semiconductors, chips, and data centers are constantly growing. Suppliers can’t keep up. They are growing at a breakneck pace. The entire sector is growing profitably, earning money. And those who buy from them are already seeing the benefits of these AI investments.
Q: In which sectors will this technological change have the biggest impact?
A: It’s compelling that in 2025, the focus was on the massive investments in data processing centers for AI, the credit risk of some of these companies, and the excess electricity consumption. Now, in 2026, the conversation has shifted to which sectors will be most affected. The industries with the potential for disruption are, potentially, all service industries – tax advice, financial data management, insurance and real estate brokerage.
Q: There will also be an overall improvement in productivity…
A: That’s what economists say. And it seems logical. In banks, in any company, AI will help them be more efficient and productive. It will improve corporate margins. However, it will also require fewer people, which will impact GDP. We’ll have to see the balance and how this transition unfolds.
Q: Is there enough electrical capacity to meet the growing demand from data centers?
A: Electricity demand had been flat in recent years, but now it’s soaring because of this. Each data center needs many gigawatts of power. In the United States, it already represents 10% of electricity demand and could reach 30% in a few years. I believe there will be enough electricity production by 2026 or 2027. Perhaps not by the end of the decade. Increasing capacity is not straightforward. You have to buy land, obtain permits, and secure energy sources. It’s becoming increasingly challenging.
Q: So, will there be a limitation on the development of AI in the United States?
A: Not necessarily. We’re already seeing discussions about alternative solutions, and we’ll even see data processing centers in space, powered by solar energy, generating AI.
Q: How would those develop?
A: Some players in the sector are already exploring these innovative concepts, involving thousands of satellites housing small data processing centers. The idea would be for them to orbit the Earth and send the information processed by AI back to their providers on Earth in a coded form.
Q: Are there companies already working on this technology?
A: Yes. Some of the space giants are already working on it. There are limitations, of course, such as launch and orbital capacity, because that does require a lot of energy. In the future, we’ll see those towers with hundreds of cards from large technology companies that are now the basis of data processing centers, but orbiting the Earth.
Q: Is that similar to Elon Musk’s Starlink communication satellites in orbit?
A: Yes, it’s the same concept. Whereas generating AI requires a higher energy expenditure. They will also require to be larger satellites.
Q: AI is undoubtedly the technological revolution of the decade, but what will be the revolution of the 2030s? Robotics for care, autonomous air taxis…
A: I couldn’t say. But what I do know is that those two examples are just another use of AI. It’s an algorithm performing new tasks, of piloting, of robotics. AI is the brain, which will be used on new bodies. What could really change everything is quantum computing [a way of processing information that, instead of solving problems step by step, analyzes millions of solutions at the same time]. I also reckon we’ll see significant advances in space research and new energy sources.
