Google, Amazon and Intel stumble as investors begin to question the price of the AI boom
The latest earnings reports have shifted attention from the promise of artificial intelligence to the enormous costs of building it, sending the Nasdaq toward one of its sharpest monthly declines in recent years.
It started with Google, moved on to Amazon and ended with Intel. Three key players in the “AI revolution” disappointed investors, each in its own way, with financial reports released over the past few days. The fallout sent the Nasdaq into one of its worst trading sessions in recent memory on Friday, capping one of the weakest months for technology stocks in some time.
The tech-heavy index plunged 2.86% on Friday, leaving it nearly 8% below the peak it reached in earOctober, its sharpest decline in a year and a half. The S&P 500 fell 1.8% on the day, although it remains only 3% below its own peak earlier this month.
Since Wall Street's early October highs, which appeared increasingly difficult to justify given elevated bond yields and expectations of further Federal Reserve rate hikes, technology stocks have entered a period of heightened volatility. Investors are increasingly concerned that an “AI bubble” is inflating and could be close to bursting.
A familiar pattern has begun to emerge. Data and figures that would have fueled optimism just a few months ago are now having the opposite effect, triggering anxiety rather than enthusiasm. Amazon and Google have both signaled that they will continue ramping up spending on AI infrastructure as they deepen their involvement in the sector. What was once seen as evidence of confidence in the technology is increasingly being interpreted as a warning about the cost of the race.
The concerns that first emerged around “neo-cloud” companies, a new generation of cloud providers led by firms such as Nebius, which has significant operations in Israel, and CoreWeave, have now spread to the established giants of Big Tech.
At Amazon, the market reaction was relatively contained. The company narrowly missed forecasts, while a second consecutive quarter of negative operating cash flow added to investor concerns. Google's experience was far more dramatic. Its shares plunged 12%, wiping roughly half a trillion dollars off its market capitalization in a single session. The decline continued on Friday, capping the company's worst week since it went public roughly two decades ago.
Google's results painted a troubling picture. Its expansion in AI has been accompanied by a sharp increase in spending and mounting pressure on cash flow. At the same time, growth in its core advertising business has begun to slow. Even Google, the company behind the Gemini AI models, is not immune to changes in user behavior that are affecting other businesses built around digital content and advertising.
Intel added further fuel to the selloff, with its shares plunging 28%. The decline marked the second-worst trading day for the chipmaker since 1974. Although Intel's difficulties are largely company-specific, its weak forecast triggered a broader retreat across the semiconductor sector. Nvidia fell 4%, while the semiconductor index dropped 7%. Nvidia, now valued at less than $5 trillion, also lost its position as the world's most valuable company to Apple.
The cost of the AI revolution
Amid all the market noise, the central question on Wall Street is increasingly about cost. The AI revolution is real and well underway, but how much more capital can the market provide to finance the infrastructure needed to sustain it?
Expectations of further interest rate increases in the United States, combined with the latest results from tech giants such as Amazon and Google, have complicated the picture. Their reports suggest that even the companies with the deepest pockets are beginning to feel the strain of the spending race.
The AI industry has developed a circular financial ecosystem in which its biggest players are increasingly dependent on one another. Nvidia, for example, invests in companies across the sector, helping ensure that they have the resources to continue buying its GPUs, the hardware that powers AI systems developed by companies such as Anthropic and OpenAI.
This ecosystem has been expanding for two years and is now attracting greater scrutiny. At its center are Amazon, Microsoft, Google and SpaceX, which are committing hundreds of billions of dollars to provide cloud computing capacity to AI companies. Capacity that these giants cannot supply is increasingly being provided by “neo-cloud” operators, which rely on debt to finance the construction of new data centers.
The result is a vast network of investments, financing arrangements and commercial relationships. As long as capital remains available and demand continues to grow, the system can expand. But rising borrowing costs, mounting infrastructure expenses and questions about future returns are making investors more cautious.
Anthropic could provide the next test
A clearer picture of the economics of the AI business may emerge from Anthropic's prospectus and its planned public offering. Leaked figures showing substantial losses had raised concerns about the company's ability to build a sustainable business. But the full prospectus, published last week, offers a more nuanced picture and suggests that its underlying business model may be viable.
Unlike the prospectus of a company such as WeWork, Anthropic's financial disclosures suggest that costs and revenues are beginning to converge. The concept Wall Street calls “unit economics” appears to be working: losses need not continue indefinitely if revenue growth eventually outpaces infrastructure costs and the company becomes more efficient as it scales.
The recent decline in AI-related stocks also presents an interesting backdrop for the anticipated offering. It is possible that major institutional investors have been keeping some capital on the sidelines ahead of the deal, driven in part by fear of missing out on what could become one of the largest capital raises in history.
If the offering goes ahead, potentially at a valuation of around $2 trillion toward the end of November, it could become a major catalyst for a year-end rally. That scenario would also depend on broader conditions, including whether the United States avoids becoming deeply involved in a renewed conflict with Iran.
There is another reason investors may begin to look at the market differently. Unlike in early October, valuations across some of the largest AI-related companies have become considerably more attractive following the selloff. Nvidia is trading at roughly 25 times expected earnings, Amazon at 19 times and Google at 14 times. Those multiples are comparable to the historical average for the S&P 500 and below the Nasdaq's average.
The question is whether lower valuations will be enough to bring investors back. The AI boom has not disappeared, and the demand for computing power continues to grow. But after two years of seemingly limitless enthusiasm, Wall Street is beginning to demand a clearer answer to a more difficult question: not how much money can be poured into AI, but how much money can ultimately be made from it.
