Big tech companies lose nearly R$ 4.8 trillion, raising alerts about the AI bubble

The largest global technology corporations recorded a market value loss of nearly US$ 900 billion, equivalent to approximately R$ 4.88 trillion, after releasing financial reports. These results intensified concerns related to the high costs inherent in the competition for artificial intelligence (AI).

Despite the adverse investor reaction, giants such as Amazon, Google, Meta, and Microsoft maintained their plans to increase investments in data centers and chips. This move triggered a new wave of stock trading on Friday (30), with Intel shares falling up to 29.65% at the Nasdaq opening. Amazon, Nvidia, and Alphabet, which controls Google, also saw declines.

On Thursday (29), Alphabet had already suffered a depreciation of over 13%, resulting in the elimination of about US$ 500 billion (R$ 2.71 trillion) from its market value. This scenario fuels fears that the valuation of AI-related companies is based on unsustainable expectations.

Futurism observed that the stock drop signals uncertainties regarding these giants' ability to convert infrastructure investments into adequate revenue. A central factor lies in the rapid increase in expenditures. Amazon, Meta, Microsoft, and Alphabet jointly increased their capital investments from US$ 113 billion to US$ 207 billion (about R$ 1.12 trillion) in one year, dedicating most of these funds to chip acquisition and data center construction.
Even in the face of negative market reaction, the companies forecast spending an additional US$ 232 billion (R$ 1.26 trillion) in the fourth quarter. Consequently, these companies' capital investments could reach US$ 740 billion (R$ 4.01 trillion) by 2026. In addition to spending, the companies face challenges in transforming their new AI resources into advertising revenue comparable to that generated by traditional search methods.
Executives admitted that Google cannot implement new data centers at the necessary speed to meet demand. Meta also reported a negative free cash flow of US$ 2.1 billion (R$ 11.4 billion) in the third quarter, marking its first negative result since its IPO in 2012. Meta invested US$ 35 billion (R$ 189.8 billion) during this period and maintains the projection of reaching US$ 145 billion (R$ 786.4 billion) in capital investments throughout the year.
Microsoft proved to be an exception. Although it raised spending to US$ 45.2 billion (R$ 245.1 billion), the company still managed to generate US$ 18.6 billion (R$ 100.9 billion) in cash in the quarter, and its shares reacted more moderately after the earnings release.

Market apprehension is not focused solely on company growth, but rather on their ability to maintain growing investments without harming the financial stability of the businesses. Brent Thill, an analyst at Jefferies, commented that investors were alarmed by positive results, but that these were not extraordinary given prior expectations. He also pointed to physical restrictions in infrastructure construction as a potential obstacle to corporate expansion.
Brad Reback, an analyst at Stifel, opined that the current market demands greater rigor and requires concrete evidence that AI infrastructure investments are actually generating new revenue. This pressure has forced companies to seek other forms of fundraising. Alphabet has contracted tens of billions of dollars in debt and announced plans to raise about US$ 85 billion (R$ 460.7 billion) in its first stock offering in over two decades.
Meta, meanwhile, presented a share issuance plan worth US$ 95 billion (R$ 514.9 billion), counting on the support of chip manufacturers Nvidia, AMD, and Broadcom. The instability may also affect smaller companies that rely on loans to finance the assembly of data centers and the purchase of equipment, such as so-called neoclouds.
CoreWeave, one of these companies specializing in cloud computing infrastructure for AI, faces project delays and contract cancellations. The company's shares plummeted 47% in the last six months, and part of its debt began trading at levels associated with troubled assets. This scenario raises concerns because these companies depend on financing to continue expanding their infrastructure, and access to new resources may become more difficult with more cautious investors.
Futurism also highlights that the expansion model based on vast data centers, characterized by high energy consumption and billion-dollar investments, is being questioned in light of alternatives that employ more economical and efficient AI models. Although it is not yet possible to confirm the bursting of the AI bubble, the market reaction demonstrates that for tech giants, promising advancements is no longer enough; investors demand financial results that justify the magnitude of the risk taken.

Similar stories

business-standard.comOctober 30, 06:28

Tech Giants' AI Investments Face Hurdles Due to Slowdown in Core Business Growth

The core business segments of American tech giants—advertising, consumer, and retail services, which support total AI spending of $750 billion this year—are beginning to show signs of deceleration. In some cases, the technology itself, on which companies are placing large bets, is starting to displace their traditional business models.

Earlier this week, Alphabet Inc., the owner of Google, reported that advertising sales through search and other channels—a segment accounting for about half of total revenue—grew the slowest in the last five quarters. Nevertheless, the company maintained its capital expenditure forecast at up to $205 billion for the current year.

Amazon.com Inc. stated on Thursday that growth in both its main online store and advertising business slowed in the last quarter, but the company confirmed its plan for capital expenditures of approximately $220 billion for this year. Microsoft Corp. and Meta Platforms Inc. also maintained their spending forecasts, presenting results that ultimately failed to meet investor expectations.

Earnings reports and subsequent stock sell-offs reveal growing tension surrounding the transition of major tech companies to new artificial intelligence-based products and associated costs. Companies are taking on unprecedented levels of debt, selling equity, entering into multi-billion dollar data center deals, and assuming trillions in future obligations to build global AI infrastructure. Meanwhile, new products are reducing profitability and diverting attention from some legacy offerings.

Investor patience is wearing thin. Alphabet shares plummeted by a record amount on Thursday, losing nearly $500 billion in market value. Amazon shares fell by 7% during trading, and Meta dropped 10% after publishing last week's earnings, when the company warned investors that its new AI agent application, Muse, requires enormous computational power despite plans to invest more aggressively in AI.

Dan Morgan, Senior Portfolio Manager at Synovus Trust, commented on the situation following Alphabet's report: 'The market is perplexed: how will all this pay off?'

Even Microsoft, which demonstrated strong growth in its cloud business and controlled costs, did not escape negative consequences. Initially, the company's stock rose after news that sales in Azure, its closely watched cloud computing division, increased by 48%, reaching $33.2 billion in the quarter ending in September. However, by the close of trading on Thursday, the stock erased all gains and fell by 0.2%.

Amazon's results were more mixed. The company postponed its annual Prime Day from the third quarter of last year to the second quarter of this year, which accounted for some of the slowdown. But analysts had already adjusted their estimates to account for this, and the results were still below expectations. The e-commerce giant's advertising services division showed an 11% growth, the slowest pace since Amazon first separated this category in 2022. Online store business grew by only 2%, a level not seen in over three years as people resumed pre-pandemic consumption habits and online commerce slowed down.

These results overshadowed the growth of Amazon's cloud computing segment and intensified investor doubts regarding its aggressive AI investments. Cloud computing was also a bright spot for Google: sales in this division more than doubled compared to the previous year, and operating income tripled. The division's CEO, Thomas Kurian, joined the earnings call for the first time, emphasizing the importance of this business for Google's future.

However, investors focused more on the problems arising in Google Search, which remains the company's lifeblood. To stay relevant amid the popularity of chatbots like ChatGPT, Google has integrated more AI into its search engine, including a dedicated tab called AI Mode. But the growing popularity of this feature has negatively impacted the profitability of Google Search. Philip Schindler, Chief Business Development Officer, admitted on Wednesday that the firm is seeing a 'decrease in monetization' of queries using AI Mode, leading to further stock decline. Schindler stated that the company is working to attract more advertising in AI Mode and views the product's trajectory optimistically.

This year, Google is struggling in the AI race against startups Anthropic and OpenAI. Bloomberg Intelligence analyst Mandip Singh believes that if the company can regain its footing in AI model development, it will likely resume delivering the expected investor revenues. Singh noted: 'With all these transitions, sometimes we focus too much on the donor business. Given the margin expansion in the cloud business, I believe the combination of cloud and search, once they reach parity, will return their profitability to previous levels.'

Jean Munster, Managing Partner at Deepwater Asset Management, commented on the decrease in AI Mode monetization at Google, noting that launching new ad formats often occurs with low margins that gradually increase over time. A more complex challenge for the firm is creating agents capable of making purchases on behalf of users and earning a commission from these transactions—a business model many in Silicon Valley consider achievable following the success of Meta's personal assistant Muse. Munster concluded: 'The old way of doing business was using blue links or display ads to influence consumer behavior. The new way is how to influence a bot.'

iol.co.zaOctober 23, 12:14

Anthropic's $80 billion loss reflects high computing power costs

According to the IPO prospectus draft, Anthropic's net loss for the first half of the year reached $80.7 billion, higher than the $13 billion recorded the previous year. However, most of this amount does not represent cash leaving the company; about $72 billion is an accounting expense resulting from the increased assumed value of convertible bonds that can be converted into Anthropic shares. In fact, this loss is a side effect of the sharp rise in the company's valuation.

A similar situation was observed last year: Anthropic's net loss in 2025 was approximately $42 billion, including a non-cash expense of $34 billion, with an operating loss of roughly $8.1 billion on revenue of $4.6 billion. It is the operating loss that should be tracked, as it is driven by computing power expenses.

Growth rarely achieved by other companies

Revenue growth has been exceptional: Anthropic reportedly generated $16.3 billion in the first half of 2026, compared to $1.1 billion the previous year. Third-quarter revenue increased by 45%, reaching $16.8 billion, up from the second quarter's $11.6 billion, more than doubling the first quarter's figures.

However, this 45% growth is a cause for concern because the consistent increase is slowing down, which investors have noticed. On the investor floor, the company's valuation is being discussed near $1.5 trillion, which is below the expected level of $2 trillion projected in the summer. Nevertheless, compared to the private valuation of $965 billion in May, $1.5 trillion still represents significant growth, albeit less than the market anticipated.

Costs associated with compute shortages

Operating costs explain the caution. Anthropic has committed to spending no less than $518 billion on cloud services, computing power, and infrastructure over ten years. Approximately 80 percent of these commitments cannot be canceled. According to a Wall Street Journal analysis, the prospectus draft indicates larger commitments exceeding $600 billion, with further negotiations ongoing. The difference between these two figures suggests that the total amount is still changing.

Computing and infrastructure costs in the first half of the year approached $9 billion, up from $2.3 billion the previous year. The prospectus also mentions Anthropic's relationships. Amazon, Google, and Microsoft are increasingly viewing Anthropic as a competitor and are reluctant to provide all the necessary power, forcing Anthropic to turn to more expensive and disparate sources. The company integrates dependency on competitors into its cost structure.

What optimists are betting on

The optimistic scenario is based on changes in margins. Bankers and investors are applying revenue multiples to 2028 forecasts, betting that revenue will outpace the costs of supporting it. One prominent sponsor claims that as costs shift from model training to operation, profitability rises sharply, citing an announced inference gross margin of around 85%. If this holds, today's spending looks like an investment in a software-like economy.

Claude Code is acting as the driving force. This coding assistant provides Anthropic with a product that developers pay for regularly, making the company's revenue heavily focused on the corporate sector. Furthermore, Anthropic sells its products to businesses and developers, not end consumers, making it financially more attractive than the model led by OpenAI's ChatGPT.

Risks hidden in the footnotes

When the final prospectus is released, three points should be noted. First, concentration: about one-fifth of the revenue for the first half of the year came from two clients. Second, fragility: a temporary restriction by the U.S. Department of Commerce on Anthropic's top models contributed to a slowdown in revenue growth in June before the business recovered. Third, structure, as non-cancellable contracts mean that a halt in growth cannot be offset by cost reductions.

The accounting for convertible bonds also makes reported losses an unreliable indicator of the underlying condition. Until an audited version appears, revenue rates, reported revenue, and projected profit should not be considered verified results.

Why the entire sector is watching this

Anthropic's debut matters not only for the company itself. A weak reception could negatively impact the broader artificial intelligence market, especially considering that OpenAI is next to go public. OpenAI's CFO stated that the company will become public next year, and its own forecasts show about $278 billion in cash expenditures by 2030. Consultants warn that volatility in tech stocks and SpaceX's results after the IPO may reduce retail investor interest.