All blog posts

Unprofitable AI startups accumulate US$ 1 trillion in market value. Analysts compare it to the .com bubble of the 90s. At the same time, a survey reveals that 63% of Brazilians have already used a generative AI platform. How to interpret this paradox?

In the last 24 hours, the debate about a possible “AI bubble” has taken on even more dramatic contours. Julien Garran, of the MacroStrategy Partnership, didn't mince his words: “the current AI bubble is the biggest and most dangerous ever seen”, with a misallocation of capital 17 times greater than the .com bubble.

But are we really facing a catastrophe foretold? Or are we living in a natural moment of technological maturation? The answer may lie in the way we look at data.

The Experts' Warning: Impressive Figures

The data presented by Garran is impressive. According to his analysis, The misallocation of capital in the US is four times greater than that of the 2008 housing crisis. AI startups continue to grow in valuation without showing consistent profits.

The central argument is that language models have reached a limit of scale since the launch of ChatGPT-4 in March 2023, without significant advances in innovation. The dependence on continuous funding, with profits concentrated in companies like Nvidia, fuels this concern.

Even more worrying: a report by TecMundo indicates that profit projections for companies like OpenAI won't occur until 2029. Bank of England, IMF and Yale point out risks of sudden corrections.

The Perspective of Natural Maturation

But there is another side to this coin. Larry Ellison, founder of Oracle, offers a different perspective: “Every major technological revolution was initially called a bubble: the internet, the cloud and now AI”.

StartSe's analysis proposes a more nuanced view, dividing the phenomenon into three dimensions:

• Price bubble: Valuations that project the future as if it were already present
• Capacity bubble: Infrastructure investments operating below potential
• Speech bubble: Transforming each experimental use into a narrative of disruption

This perspective recognizes the excesses, but argues that these layers coexist and paradoxically drive the maturing of the sector. Oversized infrastructure would be “the inevitable price of scale”, comparable to electrification and the internet.

The Brazilian Reality: Mass Adoption

While experts debate bubbles, the reality in Brazil shows impressive adoption. Nexus survey of 2,012 Brazilians reveals figures that contrast with analysts' pessimism:

• 63% have already used a generative AI platform
• 51% believe that AI can make better decisions than humans in specific situations
• 16% use it daily and 20% a few times a week
• 37% consult AI for purchasing decisions, reaching 46% among young people

This data reveals something fundamental: Regardless of valuation issues, AI has already become part of Brazilian daily life. We are not talking about future promises, but about real and present use.

Marcelo Tokarski, CEO of Nexus, points out that “companies need to adapt to this reality of increasing use of AI, with a positive impact on productivity and social behavior”.

Brazil at the Educational Forefront

Another fact that reinforces Brazil's maturity: according to the OECD's Talis 2024 survey, Brazil leads the way in the use of AI among teachers. 56% of Brazilian teachers use AI to prepare lessons, compared to 36% on average for OECD countries.

About that, Piauí became the first territory in the Americas to include AI as a compulsory subject, and was recognized by UNESCO.

The Corporate Sector Also Adapts

In the business world, media agency executives report concrete transformations:

• Democratization of opportunities and tools
• Forecasting trends to guide investments
• Discovering invisible patterns through data analysis
• Automation of planning and activation
• Capacity building with fewer resources

These reports show that, far from market speculation, AI already produces real, measurable value.

The Lesson of Mistakes: 95% Implementations Fail

But not everything is rosy. An MIT report reveals that 95% of AI implementations in companies have failed, mainly due to “lack of learning and poor integration into workflows”.

This doesn't invalidate the technology's potential, but it reinforces something I've been advocating: adopting AI requires strategy, governance and a deep understanding of the technology. It's not enough to follow the hype; you have to implement it with purpose and method.

My Review: Maturity Beyond Hype

In my work with companies and executives, I have observed an interesting pattern. While the financial markets oscillate between euphoria and panic, organizations that have implemented AI strategically reap consistent results.

What separates the wheat from the chaff at the moment?

First: Recognize that financial bubbles and technological revolutions can coexist. The internet also went through a bubble in the 1990s, but that didn't stop it from becoming the infrastructure of the modern world.

Second: Focus on concrete use cases, not futuristic promises. Brazilian data shows this: people use AI to solve real problems, not to chase fantasies of superintelligence.

Third: Invest in training and governance. The fact that 95% implementations fail is not a problem of the technology, but of the way we implement it.

What this means for your organization

If you are leading a company or area, this moment offers a unique opportunity. While speculators debate valuations, you can focus on what really matters: how AI can generate real value for your business.

The companies that will emerge stronger from this phase are those that:

• Implement AI with clear objectives and defined metrics
• Invest in team training, not just technology
• Develop solid governance for ethical and responsible use
• They focus on use cases that solve real problems
• Maintain a long-term vision, beyond market fluctuations

The Cognitive Infrastructure of the Future

Larry Ellison is right when he says “AI is the most valuable technology ever seen”. But its successful application depends on going beyond the hype and focusing on responsible implementation.

The consensus among serious investors is that AI is a cognitive infrastructure that will infiltrate widelyfrom logistics to medicine, from public management to the creative economy. Market corrections will come, but the technological revolution is an established condition.

Brazil, with 63% of the population already using AI and teachers leading the way globally in educational adoption, is well placed to take advantage of this cognitive infrastructure. Our advantage lies not in escaping the “bubble”, but in implementing AI with purpose and strategy.

Conclusion: Maturity Beyond the Market

The debate about the “AI bubble” is relevant for investors and speculators. But also for business leaders and public managers, the central question is not whether there is a bubble, but how to implement AI responsibly and productively.

Brazilian data shows that we are beyond the experimental phase. With 63% of the population already using AI and real use cases multiplying, the time has come for strategic maturity.

In my consulting and mentoring projects, I help executives and companies navigate this very moment: how to turn AI's potential into concrete results, regardless of financial market fluctuations.

Because, after all, the real value of AI is not in Wall Street speculation, but in its ability to solve human problems and generate positive impact. And that, fortunately, doesn't depend on any bubble.