Artificial intelligence, bubble aroma

BarcelonaThe world of finance is beginning to dwarf planet Earth. This was stated by Harvard historian Niall Ferguson in his book. The triumph of money (Debate, 2011, updated in 2024). Over the years, this trend has solidified, and the financial economy is outpacing the real world. A prime example is the large US technology companies, especially those most closely related to artificial intelligence (AI), which have surpassed historic price levels, triggering alarms not only about the existence of a bubble, which no one doubts, but also about its potential bursting.

"If we look at it from the perspective of capital accumulation, it's clear we're facing a bubble. Every dollar available goes into the sector, but in any case, it's a technology that's still in its infancy," explains Esteve Almirall, professor in the Operations department at Inn. It's clear, he adds, that the bubble will eventually burst, but we don't know when. Nor do we know if it will have the same effects as the bubble of the early 2000s. telecoms And the internet companies, which starred in the dot-com crisis, in which five trillion dollars vanished, currently equivalent to the value of a single company, Nvidia.

"As an analyst, I observe that the change is very significant. AI is driving investments that reveal it to be a revolution comparable to that of the internet in its day, and perhaps we are not too far from seeing its benefits materialize. As an investor, the current values, with constant increases, of 'Estudis General' (IEF). The investments made by a small group of large US technology companies, between 250 and 300 billion this year, and their stock market valuations, are unprecedented. Everything will depend on market expectations, which tend to be driven by excessive optimism, as Charles Kindleberger (Manias, panics and crashes: a history of financial crises(Ariel, Barcelona, ​​2012) defined it as euphoria, and when there are any rumors or negative information, they overreact with panic. Throughout the centuries and years, crises have occurred in which promising advances were clouded by excessive expectations, from the tulip mania in the 17th century in the Netherlands, the railway crisis in the 19th century, or the mortgage crisis. subprime in 2008.

Record upon record

Nvidia, which produces high-end chips used for this technology, has broken all records when it has surpassed the five trillion –with b– in dollars of stock market valuation –it already made headlines last July when surpassed the four trillion threshold—an amount that more than triples Spain's gross domestic product (GDP), all the wealth generated in the country in a year, and is even equivalent to that of Germany. It leaves the most valuable company on the Spanish Ibex 35—which includes the largest companies by market capitalization—at 149 billion euros by a stratospheric distance. This comes after the reappearance of Michael Burry, one of those who anticipated and profited from the outbreak of the mortgage crisis. subprime, The first profit-taking has occurred, and the five trillion mark has been lost, although prices remain extraordinary.

According to Jordi Valls, fourth deputy mayor of Barcelona and head of the Department of Economy, Finance, Economic Promotion, and Tourism, Nvidia, which has gone from making chips used for video games to those essential for AI, could be considered the third country in the world in terms of GDP generation, based solely on GDP generation. He said this this week at a conference on AI organized by the magazine Forbes in Barcelona.

The movements in the sector don't stop. OpenAI, the creator of ChatGPT, agreed to take 10% of the chip producer AMD, while Nvidia will invest 100 billion in OpenAI, which, after the process of transforming into a company, has Microsoft as one of its main shareholderswith 27% of the capital. And this has allowed the giant founded by Bill Gates to surpass a value of four trillion dollars, an amount similar to the GDP of Japan, India, or the United Kingdom, although it has deflated somewhat. Apple has also joined this select club of trillion-dollar companies. The latest to join these initiatives and alliances to fully unleash the potential of AI has been Amazon, which, through its cloud services subsidiary, Amazon Web Services (AWS) has closed a $38 billion alliance with OpenAI. Its value exceeds two trillion, an amount equivalent to Canada's GDP.

All the figures surrounding artificial intelligence are astronomical. Investments by major tech companies are unstoppable. According to a report by the Swiss bank UBS, AI-related investments will not cease for now and are expected to exceed 500 billion next year. All of this has triggered alarm bells, given the volume of investment that has forced companies to issue debt to finance it. But, experts say, the situation is not yet that of a bubble about to burst, although everything will depend on whether expectations, which are what most inflate bubbles, are too high.

The great battle in this context, in which geotechnology has become one of the protagonists of geopolitics and geoeconomics, is, in any case, between the US and China. The former, the US, through an oligopoly of tech companies such as Nvidia, Apple, Microsoft, Amazon, Tesla, Alphabet –the parent company of companies like Google– or Goal (Facebook, WhatsApp, and Instagram), most of which sided with the current president of the country, Donald Trump, in the elections a year ago. One of them, Nvidia, dominates the segment of chips needed to deploy generative AI, but it is dependent on a supplier in Taiwan and, at the same time, on a firm in the Netherlands that manufactures the machines used to produce the chips, explains Nacho de Pinedo, co-founder and CEO of ISDI. It is a conflictive relationship between countries and blogs, but at the same time one with interdependencies, according to the director of CIDOB, Pol Morillas.

China, for its part, is the other major power, dedicating all its resources to becoming a superpower by controlling the rare earth elements used to produce chips. And there's fierce competition for talent. "There are no more than 150 people capable and prepared for current AI technologies worldwide," Almirall points out. In fact, that's why they are highly sought after and very well paid. In this context, Europe, which is trying to boost its strategic autonomy to depend less on the other two powers, remains a fragmented and consumer-driven bloc. The Elcano Royal Institute cites the telecommunications sector as an example: in Europe, there are 38 companies in this sector with an average investment of €750 million. In the US, there are three companies investing €20 billion, and there are also three in China and India. It's clear that more companies are needed in Europe, experts say.

As with the latest technological disruptions and advances, the pace is accelerating, and investor expectations are skyrocketing. In any case, Nvidia, one of the current stars in the business, is a different kind of company. "They already know the orders for next year. Demand will quadruple what it is now," says Almirall. In this sense, there are no short-term problems in sight regarding demand.

However, dysfunctions could arise in the medium or long term, which are what usually make investors nervous. There are elements we don't know: we don't know if demand will continue to rise like a rocket, as it is currently doing, or if improvements in computing will allow for smarter models, says Almirall. We've gone from chips to quantum computing, that is, from bits to qubits. De Pinedo explains that "we've gone from programming to teaching AI to learn on its own and loading it with data"—this is generative AI, the great leap forward. In any case, as has always happened when companies gain advantages, monopolies eventually become oligopolies, with a few dominant operators, and some companies devour others, merge, cause the disappearance of competitors, or new ones emerge. What this professor and expert from Esade anticipates is that at some point the bubble will burst, as has happened before, for example, with the dot-com bubble crisis. In the end, it's like a selective process that eliminates projects that have no chance of continuing, but some manage to survive, or new ones emerge with greater efficiency, lower costs, and more features. In any case, there are countless possibilities for continued growth. This is the case with self-driving vehicles, which operate using generative AI, and not just private cars, but also public transport, company vehicles, and others, as well as humanoid robots for various functions. All of them use chips that are currently made by Nvidia.