From euphoria to AI phobia in a matter of weeks
Nas bolsas americanas de quinta-feira, a confiança cedeu lugar ao receio: não o medo de que a inteligência artificial falhe, mas o de que triunfe de forma demasiado abrangente. Os três principais índices recuaram entre 1,3% e 2%, enquanto os investidores começaram a questionar se setores inteiros — seguros, gestão de patrimónios, imobiliário, software — sobreviverão a uma tecnologia que promete fazer o mesmo trabalho mais barato e mais depressa. É o momento em que uma narrativa de progresso se dobra sobre si mesma e revela a sua sombra.
- A certeza que alimentou semanas de entusiasmo tecnológico transformou-se subitamente em ansiedade existencial: e se a IA destruir precisamente os negócios em que investimos?
- S&P 500, Nasdaq e Dow Jones caíram em uníssono, com o setor do software e os intermediários humanos — corretores, gestores, agentes imobiliários — a sofrerem as maiores pressões de venda.
- O ouro também recuou, arrastando as mineiras numa liquidação em cascata que se alimentou a si própria, cada transação a desencadear a seguinte.
- A Cisco Systems agravou o clima ao prever margens fracas para o ano, tornando a disrupção menos teórica e mais contabilística.
- Os olhos voltam-se agora para os dados de inflação de sexta-feira: um índice de preços no consumidor benigno poderá dar à Reserva Federal — e aos mercados — espaço para respirar.
A bolsa americana travou bruscamente na quinta-feira, não por falta de fé na inteligência artificial, mas por excesso dela. Durante semanas, o dinheiro tinha fluído para as tecnológicas com a convicção de que a IA remodelaria o mundo. A meio da semana, essa convicção azedou: e se a IA funcionar bem demais, e contra nós?
O S&P 500 perdeu 1,57%, o Nasdaq recuou 2,03% e o Dow Jones cedeu 1,34%, enquanto os investidores abandonavam posições em software, seguros, gestão de patrimónios e imobiliário — setores cujo modelo de negócio assenta em expertise humana, relações e acesso à informação. Tudo aquilo que as máquinas estão a aprender a fazer mais barato e mais depressa. O estratega Ross Mayfield, do Baird, descreveu o mecanismo com frieza: vende primeiro, analisa depois, mas não fiques com as perdas. Steve Sosnick, da Interactive Brokers, foi mais direto: os setores em risco são precisamente os que sempre venderam informação e acesso — vantagens que a tecnologia está a corroer.
O ouro também caiu, arrastando as mineiras numa liquidação em cadeia. A Cisco Systems agravou o ambiente ao anunciar margens fracas para o ano, pressionada pelos custos dos semicondutores de memória — um lembrete de que a disrupção já aparece nos números, não apenas nas previsões. Os analistas da Yardeni Research resumiram a viragem numa frase: da euforia à IA-fobia. A comparação com a bolha dot-com surgiu naturalmente, mas com uma diferença inquietante: a internet criou novas categorias de negócio ao lado das antigas; com a IA, o receio é que ela não acrescente — substitua.
A pergunta imediata é o que mostrarão os dados de inflação de janeiro, divulgados esta sexta-feira. Se o índice de preços no consumidor vier em linha com as expectativas — uma subida de 0,3% — poderá dar à Reserva Federal razões para pausar os cortes nas taxas sem alarmar os mercados. Seria uma trégua, não uma resolução. O mercado não precisa de respostas definitivas sobre a IA; precisa apenas de tempo para as formular.
The American stock market seized up on Thursday, and the reason was simple: investors had stopped believing their own story about artificial intelligence. For weeks, the money had flowed into tech stocks on the assumption that AI would reshape everything. By midweek, that certainty had curdled into something closer to dread. The three major indices all fell—the S&P 500 dropping 1.57 percent to close at 6,832.76 points, the Nasdaq sliding 2.03 percent to 22,597.15, and the Dow Jones retreating 1.34 percent to 49,451.98—as traders began asking a question that had been lurking beneath the enthusiasm all along: what if AI actually destroys the businesses we own?
The anxiety had a specific shape. Investors weren't worried that artificial intelligence wouldn't work. They were worried that it would work too well, and that it would work against them. Software companies, insurance brokers, wealth managers, and real estate agents all found themselves in the crosshairs. These are sectors with business models built on human expertise, human relationships, human gatekeeping. If AI could do those things cheaper and faster, what happened to the people who currently did them for a living? What happened to the companies that employed them? The fear wasn't abstract—it was about whether entire categories of business could survive what was coming.
Ross Mayfield, an investment strategist at Baird, described the dynamic with a certain clarity. What was happening, he said, was mass psychology. Sell first, analyze later, but don't get caught holding the losses. The money leaving software stocks had to go somewhere, and it was looking for exits. Steve Sosnick at Interactive Brokers named the sectors he believed faced potential decimation: the ones whose competitive advantage had always been information, access, or expertise. All things that machines were getting better at, faster than anyone had predicted.
The gold market was collapsing too, which created a secondary pressure. Mining companies were being hit from both sides—the sector rotation away from risk, and the falling price of the commodity itself. Traders were liquidating whatever they could to cover losses elsewhere. It was the kind of cascading selling that feeds on itself, each transaction triggering the next.
Analysts at Yardeni Research captured the mood in a phrase: from euphoria to AI phobia. They drew a parallel to the internet boom, noting that both represented technological ruptures profound enough to alter the behavior of nearly everyone. The difference was that the internet had created new categories of business alongside the old ones. With AI, the fear was that it wouldn't add—it would replace. For those who had lived through the dot-com era, there was something grimly familiar about watching certainty flip to panic in the span of a few trading sessions.
The immediate question now was what Friday's inflation data would show. Economists surveyed by Dow Jones expected the January consumer price index to have risen 0.3 percent. If that number came in positive but not alarming, it might give the Federal Reserve room to pause its interest-rate cuts for a while. One analyst noted that strong labor market data had already shifted the calculus—the Fed didn't need to keep cutting aggressively if employment remained solid. A benign inflation report could buy time, could let the market settle, could give investors a few months of data before anyone had to make a definitive call about where rates were headed.
Cisco Systems fell more than 12 percent after warning that its profit margins would be weak in the year ahead, citing pressure from higher memory semiconductor costs. It was a concrete reminder that the disruption wasn't theoretical. Companies were already feeling it in their numbers, already adjusting their guidance, already bracing for a world that looked different than the one they'd planned for. The market was pricing in not just the possibility of change, but the certainty of it.
Citas Notables
It's mass psychology—sell first and analyze later, but don't get caught with the losses— Ross Mayfield, investment strategist at Baird
From euphoria to AI phobia—a déjà vu for those who lived through the internet boom— Yardeni Research strategists
La Conversación del Hearth Otra perspectiva de la historia
So the market fell because people got scared of AI. But wasn't everyone already scared of AI? What changed on Thursday?
The fear was always there, but it was buried under a different story—the story that AI would make tech companies richer and more powerful. On Thursday, investors started asking what happens to everyone else. What happens to the insurance broker, the real estate agent, the software company that doesn't own the AI. That's when the selling started.
But those sectors have been around for decades. Why would they suddenly become worthless?
They wouldn't become worthless overnight. But their entire value proposition—their reason to exist—is based on expertise and access that machines are now getting better at providing. That's not a small thing. That's existential. And when you're holding stock in a company whose business model might not survive the next five years, you sell.
The article mentions gold falling too. What does that have to do with AI?
Nothing directly. But when traders are panicking and need cash to cover losses, they sell whatever they can. Gold is liquid. Mining companies own gold. So mining gets hit twice—the sector rotation away from risk, and the falling price of the thing they sell.
Is this the beginning of a real crash, or just a correction?
No one knows yet. That's why everyone's watching Friday's inflation number. If it's calm, the Fed can pause, the market can breathe, and maybe people remember that AI is still a tool, not a judgment day. If it's hot, then you've got inflation fears on top of disruption fears, and that's a different story entirely.
One analyst said money leaving software has somewhere to go. Where?
That's the question, isn't it. Defensive sectors, maybe. Bonds. Cash. Places where you're not betting on a business model that might be obsolete in three years. The money doesn't disappear—it just moves to where people think it's safer.