But during September, investors began to show greater caution.
On 14 September, AI-related stocks fell significantly across global markets. The Philadelphia Semiconductor Index fell 5.2%, while Nvidia, AMD and Micron recorded sharp declines. The Nasdaq 100 fell to its lowest level in six weeks.
This does not mean that investors have given up on AI.
On the contrary, the question becoming increasingly important is:
How much of the expected AI growth is already priced into stocks?
The Problem Is Not Just the Technology. It Is the Price
A company can have very strong prospects and still be an expensive investment.
This is one of the basic principles of financial markets.
If investors expect a company to significantly increase its revenues and profits in the future, they are willing to pay more for its shares today.
But the higher the expectations, the less room there is for disappointment.
This is particularly important for AI-related companies because investments in this sector require extraordinary amounts of capital.
Data centers, chips, energy and cloud infrastructure cost billions. At the same time, investors want to see these expenditures translated into revenues and profits.
The case of Anthropic illustrates this contrast well: the company has rapidly increased its revenues, but at the same time has recorded very large losses and anticipates substantial future infrastructure commitments.
Therefore, the market is moving from the question:
“Will AI change the economy?”
to a much more concrete question:
“How much will this growth cost, and who will actually benefit from it?”
Bonds Are Becoming an Alternative Again
The pressure on technology stocks is not coming only from AI itself.
Another factor is the bond market.
In September, U.S. bond yields rose significantly. The 10-year Treasury yield surpassed 5%, while long-term yields reached multi-year highs.
Why is this important?
When bond yields rise, investors have a more attractive alternative for their capital. Instead of taking high risk by investing in a highly valued stock, they may seek returns through fixed-income instruments.
For technology companies, financing costs also become higher.
And this is particularly important for the AI industry, which requires colossal investments in infrastructure.
According to Reuters, major technology companies and hyperscalers have issued hundreds of billions of dollars in debt to finance investments in the AI ecosystem, while financing costs have increased.
Is This Another “Boom-Bust”?
This is the question markets are trying to answer.
The history of technology has shown that genuine innovations can be accompanied by financial bubbles.
This also happened during the dot-com era: the internet transformed the economy, but that did not mean that every internet company had a justifiable market valuation.
Today, the fact that AI may be transformative does not mean that every company associated with it will be a successful investment.
In fact, one of the most interesting signals is that the market is beginning to become more selective.
At the end of September, an AI infrastructure company, Accelevation, had a weak Nasdaq debut, with its shares falling 2.5% on the first day of trading after the IPO was priced below the targeted range. Reuters linked this to greater scrutiny of valuations, future demand and the cost of capital across the AI ecosystem.
This could be an important sign:
Investors are no longer looking only for growth stories. They are looking for evidence of growth.
What Happens Next?
It is still too early to say whether the correction in AI stocks is the beginning of a longer cycle or simply a normal period of consolidation following an extraordinary rally.
The AI economy itself continues to expand, and infrastructure investment remains very substantial.
But the market is setting a higher standard.
Companies need to demonstrate that large investments in AI can generate real revenues and profits. Meanwhile, investors need to assess not only the technological potential, but also the price they are paying for it.
In this sense, the correction may be less a vote against AI and more a reassessment of expectations.
And if bond yields remain high, markets may continue to be less tolerant of companies with very high valuations and profits that are still far into the future.
Source: Reuters • Federal Reserve • European Central Bank
FINARCH RECOMMENDS
Do not confuse powerful technology with an automatically good investment.
AI can transform the economy and the way businesses operate. But for investors, the question is different: does the price justify the company’s potential?
In a market where expectations are high and the cost of capital is changing, concentrating all capital in one sector or trend increases exposure to volatility.
Diversification does not mean avoiding technology. It means not allowing a single investment story to define the entire portfolio.
Our recommendation:
Do not invest only in what looks like the future. Assess the price, risk, time horizon and alternatives before making a decision.
AI may be one of the greatest economic transformations of our time. But a good investment requires more than a good story.
This material is for informational purposes only and does not constitute individual investment advice.