Do the vast sums spent on AI actually produce returns that justify the investment?
Nos mercados financeiros de terça-feira, uma dúvida antiga voltou a pesar sobre Wall Street: a inteligência artificial, por mais promissora que seja, ainda precisa provar que gera retorno real. As quedas do S&P 500, Nasdaq e Dow Jones não foram colapsos, mas sinais de uma paciência que se estreita — investidores que financiaram bilhões em infraestrutura de IA aguardam, com crescente ceticismo, a chegada dos frutos. O anúncio da Anthropic sobre novas capacidades do Claude no setor jurídico funcionou menos como novidade e mais como espelho: a IA pode não criar tanto valor novo quanto redistribuir, de forma imprevisível, o valor que já existe.
- A pergunta que Wall Street não consegue responder — o gasto colossal em IA vai se pagar? — voltou a dominar o pregão com força renovada.
- Nvidia e Microsoft recuaram quase 3% cada, ainda sob o peso da queda de 10% da Microsoft na semana anterior, que destruiu cerca de US$ 360 bilhões em valor de mercado em um único dia.
- O anúncio da Anthropic de que seu chatbot Claude agora executa tarefas jurídicas acendeu um alerta: empresas de software especializadas no setor jurídico podem ver seus modelos de negócio corroídos antes mesmo de a IA provar seu próprio valor.
- Alphabet e Amazon se aproximam de seus relatórios de resultados carregando a expectativa de todo o mercado — suas receitas de nuvem e IA serão lidas como veredicto sobre a viabilidade do investimento tecnológico.
- Em contraste, o Walmart subiu quase 3% e ultrapassou US$ 1 trilhão em valor de mercado, lembrando que nem toda a economia apostou seu futuro na inteligência artificial.
A terça-feira em Wall Street foi marcada por uma ansiedade conhecida, mas cada vez menos tolerada: a dúvida sobre se a inteligência artificial, após bilhões em investimento, vai de fato gerar lucro. O S&P 500 fechou em queda de 0,84%, o Nasdaq recuou 1,43% e o Dow Jones cedeu 0,34% — não um colapso, mas um recuo com recado.
Nvidia e Microsoft lideraram as perdas entre as grandes empresas de tecnologia, cada uma caindo quase 3%. O contexto para a Microsoft era especialmente pesado: na semana anterior, a empresa havia relatado crescimento de receita em nuvem abaixo do esperado enquanto acelerava os gastos com IA. A punição do mercado foi imediata e brutal — uma queda de 10% que apagou cerca de US$ 360 bilhões em valor de mercado. Esse tipo de memória não desaparece rapidamente.
O gatilho imediato do dia foi o anúncio da Anthropic de que seu chatbot Claude havia expandido suas capacidades para o setor jurídico, passando a executar tarefas que empresas de software especializadas vendiam a escritórios de advocacia há anos. O mercado leu o movimento como um aviso: a IA pode não estar criando tanto valor novo quanto redistribuindo, de maneira imprevisível, o valor já existente — corroendo os fossos competitivos de quem apostou nesse nicho.
Nem tudo foi queda. O Walmart subiu quase 3% e ultrapassou pela primeira vez US$ 1 trilhão em valor de mercado, oferecendo um contraponto silencioso à turbulência tecnológica. Ao fundo, relatos de que os Estados Unidos abateram um drone iraniano próximo a um porta-aviões americano adicionaram tensão geopolítica ao dia, embora secundária à preocupação central dos investidores.
Com Alphabet e Amazon prestes a divulgar seus resultados, o mercado aguarda — com paciência cada vez mais curta — alguma evidência de que o maior ciclo de investimento em tecnologia da história moderna vai, afinal, se pagar.
The stock market opened Tuesday with a familiar anxiety: the question of whether artificial intelligence's enormous promise will ever actually make money. By day's end, the numbers told a story of doubt. The S&P 500 fell 0.84% to close at 6,917. The Nasdaq, heavy with technology stocks, dropped 1.43% to 23,255. The Dow Jones slipped 0.34% to 49,240. It was not a collapse, but it was a retreat.
The retreat had a face. Nvidia and Microsoft, the two companies most visibly betting their futures on AI infrastructure and services, each lost nearly 3% of their value. Alphabet, Google's parent company, fell more than 1% ahead of earnings due Wednesday. Amazon shed almost 1.8% before its own results were set to arrive Thursday. These were not small movements for companies of this size and prominence.
The immediate trigger was an announcement from Anthropic, a startup building AI systems, that its Claude chatbot had acquired new capabilities—specifically, the ability to assist with and execute tasks in legal work. The news rippled through the market as a warning. If artificial intelligence could now do the work that specialized software companies had been selling to law firms for years, what happened to those software companies' business models? The fear was not abstract. It was about disruption, about competitive moats eroding, about the possibility that AI would not create new value so much as redistribute existing value in unpredictable ways.
But the deeper anxiety was older and more fundamental. Wall Street has been wrestling for months with a question that no amount of optimistic earnings guidance has fully answered: Do the vast sums that technology companies are spending on AI infrastructure—the chips, the data centers, the research—actually produce returns that justify the investment? Microsoft itself had provided fresh evidence of the doubt just days earlier. On Thursday of the previous week, the company reported cloud revenue growth that fell short of expectations and disclosed that spending on AI had accelerated. The market's response was severe. Microsoft's stock fell 10%, erasing roughly $360 billion in market value in a single day. That kind of punishment does not fade quickly from investor memory.
Not everything moved downward. Walmart's stock climbed nearly 3%, pushing the retailer's market value above $1 trillion for the first time. The company's resilience stood in contrast to the technology sector's uncertainty, a reminder that not all of the economy is betting its future on artificial intelligence.
There was also noise from the geopolitical sphere. Reports that the United States had shot down an Iranian drone approaching an American aircraft carrier added a layer of tension to the day's trading, though it was secondary to the market's primary concern. By the time the closing bell rang, the message was clear: investors wanted proof that AI spending would pay off, and they were not yet convinced.
Citas Notables
Investors questioned whether the enormous spending by technology companies on AI infrastructure would be justified by actual returns— Market sentiment reflected in trading patterns
La Conversación del Hearth Otra perspectiva de la historia
Why did Anthropic's announcement about Claude handling legal work spook the market so badly? It's just one company adding a feature.
Because it made concrete something investors had been trying not to think about. If AI can do legal work, it can do a lot of other specialized work too. That threatens the software companies that have been selling those services for years.
But wouldn't that just mean the market shifts? New winners, new losers?
In theory, yes. But the market hates uncertainty about who the winners will be. And right now, the biggest tech companies have spent enormous amounts on AI without clear evidence it's making them more profitable. So when a startup shows AI can disrupt existing business models, it raises the question: what if the disruption happens to us?
Microsoft lost $360 billion in value last week. That's staggering. Do investors think the company is in real trouble?
Not in real trouble, but in a reckoning. Microsoft is spending heavily on AI, and the market wants to see that spending translate into revenue growth. When cloud growth disappointed, it suggested the spending might not be paying off yet. That's enough to scare people.
So what would restore confidence?
Evidence. Earnings reports that show AI investments are actually driving revenue and profit. Anthropic's announcement did the opposite—it suggested AI might cannibalize existing revenue streams. That's the fear that's driving the selling.
Walmart hit $1 trillion in value while tech was falling. What's that about?
It's a vote for the old economy. Walmart doesn't depend on AI profitability. It sells things people need. When the future feels uncertain, money sometimes flows back to the present.