AI: the bubble that threatens the global economy

/ Uncategorized / By /

For years, companies like OpenAI and NVIDIA have been portrayed as the catalysts of the future, the engines of unprecedented transformation.

Today, that narrative is met with a less glamorous reality: dependence on public subsidies to sustain its growth rate.

“Too big to fail”: a phrase we already hear

By: Gabriel E. Levy B.

In 2008, when the global financial system collapsed under the weight of its own greed, the Federal Reserve and governments around the world came to the rescue of banks that they said they could not fail.

They were too big, too connected, too vital to be left to die.

That logic, which the economist Simon Johnson criticized as the beginning of a “financial oligarchy”, returns to the scene with other actors: the titans of artificial intelligence.

OpenAI, according to statements by its CFO Sarah Friar at a Wall Street Journal event, believes that it will not go public until at least 2027.

Instead of profitability, its objective is to continue investing aggressively in R+D. So far, everything is expected in an ambitious start-up.

What was striking was when Friar mentioned that they expect “support from the government” for future infrastructure agreements.

A thinly veiled request for state aid to finance the colossal costs of the data centers that their models require.

Later, Friar tried to qualify his words on LinkedIn.

He said he misused the term “backing,” that the company does not seek direct help from the government. But the message was already out.

Like a stone thrown into the pond, the wave spread quickly in markets and institutions.

“China is going to win the AI race”

NVIDIA CEO Jensen Huang’s words at the Financial Times Summit in London are a warning rather than an analysis.

He said China, thanks to its looser regulation and state energy subsidies, is in an advantageous position to lead the global tech race.

Although his country does not have access to the most advanced chips due to export restrictions imposed by the United States, it has subsidized energy and a regulatory ecosystem adapted to scale quickly.

It’s a message to Western governments: If they don’t subsidize the energy in our data centers, we’ll lose the upper hand.

Thus, the narrative of technological innovation, market freedom and entrepreneurship once again calls for the protective embrace of the State.

Behind this silent plea there is a frightening reality: artificial intelligence is very expensive.

Not only in terms of talent or initial investment, but above all in terms of operational cost.

Powering a model like GPT-4 or training a state-of-the-art machine vision system requires millions of dollars in GPUs, cooling, power, and maintenance.

As specialized journalist Gerrit De Vynck points out in The Washington Post, the depreciation rates of these chips remain uncertain.

A server that today is worth millions can become obsolete in two years. Investing in infrastructure with no guarantee of return is a roulette wheel, and OpenAI wants the State to put the riskiest chips on it.

“Economic growth in 2025 is an illusion”

Economist Jason Furman, a former White House adviser, recently presented a revealing fact: if the construction of artificial intelligence data centers is excluded, US GDP growth in 2025 would have been just 0.1%.

This is also reflected by analyst Tomás Pueyo in his newsletter Uncharted Territories. The economy is growing, yes, but for a very specific reason: the digital gold rush.

The problem is that this growth is not based on concrete profits. According to data from Coatue Management, a benchmark technology fund, 48% of the growth of the S&P 500 index in the last year has come from companies linked to AI.

But many of those companies are still not profitable.

OpenAI, for example, lost $11.5 billion in the last fiscal period, while its competitor Google, with Gemini, achieved record revenue.

Michael Burry, the investor who predicted the 2008 crisis and whose story inspired The Big Short, has bet short against NVIDIA, which was recently valued at $5 trillion.

His move is clear: he believes that we are facing a new bubble. And he is not alone. According to a recent report, 54% of fund managers believe we are experiencing a tech bubble, up from 37% who thought so in July.

The paradox is that no one wants to deflate that bubble.

Neither the investors who still dream of astronomical returns, nor the governments that now fear that a technological collapse could have consequences similar to the banking collapse of 2008.

“Our economy could be reduced to three data centers in trench coats”

The phrase belongs to journalist Rusty Foster, creator of the Today in Tabs newsletter. He said it with irony, but the message is profound: we are reducing our real economy to a virtual infrastructure. We build data centers, we invest in chips, we train models.

But what is the tangible benefit?

What real problem does this explosion of artificial intelligence solve?

OpenAI says it is close to achieving AGI (Artificial General Intelligence), a form of intelligence that would equal or surpass human intelligence.

It’s his founding promise and also the reason why he keeps burning money at a breakneck pace. But, as philosopher Kate Crawford points out in her book Atlas of AI, “artificial intelligence is not a clean technology, nor a neutral one.

It is a network of exploitation of resources, human labor and physical infrastructure.”

A grid that needs energy, subsidies, public support.

And that today it falters if that help does not arrive.

Competition is also tightening. Anthropic, another AI startup funded by Amazon and Google, is gaining traction in the business sector.

Google, for its part, not only attracts more users with its Gemini model, but also obtains concrete revenue, something that OpenAI has not yet achieved.

The rope is taut. If OpenAI fails to justify its losses with tangible results, if AGI fails to show up, or if its models are outperformed by competitors, the shockwave could drag NVIDIA, AMD, Oracle, and other key players with it.

The impact would be profound, because these companies are no longer startups: they are embedded in the global economic system.

“When giants ask for help, it’s time to be suspicious”

Reliance on private capital has defined the age of technology.

But now, amid the rise of artificial intelligence, those same tech giants are calling for public backing.

As if innovation could only be sustained with state money when profitability does not arrive.

What was once presented as disruption is now sold as a strategic necessity.

The logic of “too big to fail” is once again unfolding, but in new forms.

The stakes here are not just the future of technology, but the kind of economy we are building.

One where the risks are private, but the losses, potentially, public.

In conclusion

Artificial intelligence is going through a moment of accelerated expansion and growing dependence. OpenAI and other companies face not only technological challenges, but also financial and political ones.

Its economic model seems unsustainable without state aid, raising pressing questions about the role of the state, the limits of the market, and the real viability of this digital revolution.

Perhaps, as happened in 2008, we are inflating a bubble with promises that the future will not be able to keep.

References:

Related Posts